Thursday, 2015-06-11

patrickeastand i think its doing the right thing,
patrickeastcloning os-brick from git00:01
*** smoriya has joined #openstack-cinder00:02
*** xiaohui has joined #openstack-cinder00:06
*** lixiaoy1 has joined #openstack-cinder00:07
*** belmoreira has joined #openstack-cinder00:11
*** leeantho has quit IRC00:12
*** annashen has quit IRC00:16
tsekiyamapatrickeast: I'm reading the Cinder internal tenant" spec, and have a question.. Why wouldn't the config be just "cinder_internal_tenant_id", but tenant_user/password?00:16
*** adurbin_ has quit IRC00:17
patrickeasttsekiyama: i’ve been wondering that myself actually00:17
*** setmason has quit IRC00:17
patrickeasttsekiyama: the more i think about it, the more i don’t know why we would actually need the password00:17
patrickeasttsekiyama: just the id00:17
tsekiyamapatrickeast: yeah,, inside the Cinder-volume, user/password wouldn't be used00:18
patrickeasttsekiyama: for now i’ll assume we need both, but i might end up putting a patch in to modify the spec if we end up not needing it00:18
tsekiyamapatrickeast: ok, I understand the current situation.00:18
tsekiyamapatrickeast: I believe the tenant-id is what we need though00:19
patrickeastyea, i agree00:19
openstackgerritAlexey Khodos proposed openstack/cinder: NexentaStor 5 NFS backend driver.
*** heyun has joined #openstack-cinder00:26
*** dannywil_ has joined #openstack-cinder00:30
*** diogogmt has quit IRC00:31
*** dannywilson has quit IRC00:33
openstackgerritAlexey Khodos proposed openstack/cinder: NexentaStor 5 NFS backend driver.
*** dannywil_ has quit IRC00:35
*** alexpilotti has quit IRC00:36
*** tsekiyama has quit IRC00:36
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements
*** mtanino has quit IRC00:44
*** IanGovett has joined #openstack-cinder00:47
*** annashen has joined #openstack-cinder00:47
*** heyun has quit IRC00:48
*** heyun has joined #openstack-cinder00:49
*** annashen has quit IRC00:54
*** akl__ has quit IRC00:56
*** _cjones_ has quit IRC00:57
*** diogogmt has joined #openstack-cinder01:09
*** chlong has quit IRC01:11
*** eantyshev has quit IRC01:15
*** dims_ has joined #openstack-cinder01:18
*** lcurtis has quit IRC01:18
*** dims has quit IRC01:21
*** tobe has joined #openstack-cinder01:24
*** Lee1092 has joined #openstack-cinder01:30
*** IanGovett has quit IRC01:39
*** vilobhmm has quit IRC01:41
*** IlyaG has joined #openstack-cinder01:42
*** IlyaG has quit IRC01:46
*** IlyaG has joined #openstack-cinder01:47
*** belmoreira has quit IRC01:50
*** marcusvrn has joined #openstack-cinder02:01
*** gouthamr has joined #openstack-cinder02:03
*** Apoorva has quit IRC02:06
*** rmstar_ is now known as rmstar02:11
*** lixiaoy1 has quit IRC02:11
openstackgerritBharat Kumar Kobagana (BharatK) proposed openstack/cinder: Using 'fallocate' instead of 'dd' to create raw volumes
*** jungleboyj has joined #openstack-cinder02:11
*** lixiaoy1 has joined #openstack-cinder02:12
*** IlyaG has quit IRC02:32
*** ian_brown has joined #openstack-cinder02:32
*** dims_ has quit IRC02:33
*** nikeshm has quit IRC02:36
*** tobe has quit IRC02:46
*** tobe has joined #openstack-cinder02:47
*** vilobhmm has joined #openstack-cinder02:47
*** BharatK has quit IRC02:48
*** tobe has quit IRC02:50
*** tobe has joined #openstack-cinder02:51
*** setmason has joined #openstack-cinder02:58
*** gouthamr has quit IRC02:59
*** vilobhmm has quit IRC02:59
*** david-lyle has joined #openstack-cinder03:04
*** nikeshm has joined #openstack-cinder03:04
*** sgotliv_ has joined #openstack-cinder03:05
*** nikeshm has quit IRC03:09
*** ronenkat_ has quit IRC03:09
*** nikeshm has joined #openstack-cinder03:12
*** dims has joined #openstack-cinder03:33
openstackgerritVincent Hou proposed openstack/cinder-specs: Volume migration improvement for L
*** fifieldt has joined #openstack-cinder03:58
*** ronenkat has joined #openstack-cinder03:58
*** gouthamr has joined #openstack-cinder04:00
*** tobe has quit IRC04:04
*** yamada-h has joined #openstack-cinder04:05
*** bill_az has quit IRC04:07
jgriffithSwanson: Good question:
*** yamada-h has quit IRC04:08
jgriffithSwanson: check out my answer inline04:08
*** yamada-h has joined #openstack-cinder04:10
*** yamada-h has quit IRC04:10
*** tobe has joined #openstack-cinder04:13
*** annashen has joined #openstack-cinder04:22
*** marcusvrn has quit IRC04:24
*** harlowja_at_home has joined #openstack-cinder04:25
*** markvoelker has quit IRC04:26
*** annashen has quit IRC04:27
*** yamada-h has joined #openstack-cinder04:31
*** gouthamr has quit IRC04:31
jgriffithpatrickeast: what's this: ?04:34
patrickeastjgriffith: uhhh04:34
patrickeastjgriffith: good question04:34
patrickeastjgriffith: i don’t remember setting anything for quota in the configs anywhere04:35
jgriffithhehe.. yeah, I didn't see anything like that either04:35
patrickeastjgriffith: i want to get something like logstash (or maybe even the actual infra logstack) for my ci logs04:36
patrickeastjgriffith: would be cool to search for that kind of stuff in them04:36
patrickeastjgriffith: see when that type of thing started showing up04:36
jgriffithmtreinish: BTW, the culprit in my ostestr/subunit patch is debug loggging from migrate.versioning.script.base04:37
jgriffithmtreinish: but I can't figure out how to *fix* it04:37
mtreinishjgriffith man, even when I'm on the other side of the world your always online when I am04:37
jgriffithpatrickeast: +1 for logstash04:37
jgriffithmtreinish: haha!04:37
jgriffithmtreinish: I've been wondering where you were at!04:37
mtreinishjgriffith: hmm it's probably the sdout pass through. Is that's what's causing the looping though?04:37
asselin_patrickeast, another thing in my backlog....04:37
mtreinishjgriffith: japan until sat. then israel next week04:38
*** diogogmt has quit IRC04:38
jgriffithmtreinish: I'm not sure about the looping, but it seems it's causing the subunit failure for too many chars or whatever it was04:38
patrickeastasselin_: i’ve been meaning to ask in -infra what they thought about letting 3rd party ci’s put logs into the infra logstash04:38
jgriffithmtreinish: Wow!  Racking up the frequent flyer miles04:38
mtreinishjgriffith: heh, yeah something like that :)04:38
asselin_patrickeast, my guess is no-way-jose04:38
jgriffithpatrickeast: that could be tricky... storage and all :)04:38
patrickeastwell yea04:38
mtreinishjgriffith: hmm, I thought that was an issue with the subunit-1to2 calling at the end of the script04:38
patrickeastbut like how much no way04:39
jgriffithmaybe if you send them a Pure :)04:39
patrickeasthard to do and maybe later04:39
mtreinishalthough I pushed a change to which should have fixed that a couple weeks ago04:39
patrickeastor never no way we would die first04:39
jgriffithmtreinish: maybe I should look again...04:39
mtreinishjgriffith: fwiw:
jgriffithmtreinish: hmm... so now I'm confused04:40
jgriffithmtreinish: I'm not seeing the message I was looking for earlier04:40
*** gouthamr has joined #openstack-cinder04:41
jgriffithmtreinish: but I've got several hundred lines of DEBUG output from migrate04:41
mtreinishjgriffith: heh, well my patch should have fixed that, because it was a subunit bug (which I still need to file)04:41
jgriffithmtreinish: and the final "Length too long: 5431171" message04:41
mtreinishIIRC, calling subunit-1to2 is what was generting that message04:41
jgriffithmtreinish: I'll update things and see what happens04:42
jgriffithmtreinish: thanks04:42
*** changbl has joined #openstack-cinder04:42
jgriffithand safe travels :)04:42
jgriffithI'm finding more little "nuggest" running tests this evening than I'd care to04:42
*** gouthamr has quit IRC04:43
mtreinishjgriffith: heh, yeah I can relate to that04:45
jgriffithmtreinish: so quick question :)04:46
jgriffithmtreinish: how can I suck in the newer subunit in my tox run?04:46
asselin_patrickeast, we can look at setting up a separate one for all of 3rd party ci and see....but it would likely be for 'm' at this point04:46
* asselin_ signs off for the night04:46
mtreinishjgriffith: newer subunit? like from pypi?04:46
jgriffithmtreinish: because the tox install of test-reqs is just going to get pypi's version which is "old" right?04:47
jgriffithmtreinish: yeah.. like that04:47
jgriffithmtreinish: ie to get your fix you just pointed me to :)04:47
* jgriffith thinks sometimes automation is *hard*04:47
mtreinishoh, that's in the infra run script it shouldn't need any newer package version04:47
jgriffithOh... crud04:48
mtreinishbut to do that I normally just activate the tox venv and manually run the install commands04:48
jgriffithI wasn't paying attention there04:48
jgriffithmtreinish: OH, no kidding, I didn't know tox let ya do that04:48
jgriffithI learned something valuable/new tonight04:48
mtreinishjgriffith: yeah normally something like 'source .tox/py27/bin/activate'04:49
mtreinishjgriffith: all that change does is switch to using 'testr last --subunit' to just write a subunit v2 stream directly instead of running subunit-1to2 on the stream in .testrepository04:49
jgriffithmtreinish: hey... look at that!  Way cool04:49
*** harlowja_at_home has quit IRC04:51
* jgriffith is apparantly not very sharp04:52
jgriffithls bin04:53
jgriffithI can't quite figure out how we tie all this together on the Cinder side04:54
jgriffiththink I'm going to quit for the night04:54
*** haomaiw__ has joined #openstack-cinder04:55
*** haomaiwang has quit IRC04:56
*** tobe has quit IRC04:56
jgriffithOhh.. DERP04:56
jgriffithyeah, so I don't have any of that because I'm not using ostestr to call everything04:57
jgriffithOhhhh neat!04:59
mtreinishjgriffith: heh, did you figure it out?05:00
*** tobe has joined #openstack-cinder05:01
jgriffithmtreinish: just discovering stuff I can do with ostestr directly in the venv05:01
jgriffithmtreinish: and I did manage to do "something" that makes the problem go away05:01
jgriffithbut I don't know what it was :)05:01
mtreinishheh, ok05:02
jgriffithOk.. I've got to stop, I'm just getting distracted at this pont05:02
jgriffithI'm out of here05:02
jgriffithhave a good night everyone05:02
jgriffithcya tomorrow05:02
*** shausy has joined #openstack-cinder05:09
*** patrickeast has quit IRC05:11
*** deepakcs has joined #openstack-cinder05:12
*** setmason has quit IRC05:14
*** yamada-h has quit IRC05:16
*** davechen has quit IRC05:17
*** yamada-h has joined #openstack-cinder05:18
*** belmoreira has joined #openstack-cinder05:22
*** yamada-h has quit IRC05:22
*** setmason has joined #openstack-cinder05:23
*** markvoelker has joined #openstack-cinder05:27
*** markvoelker has quit IRC05:32
openstackgerritPh. Marek proposed openstack/cinder: Re-add DRBD driver.
*** nkrinner has joined #openstack-cinder05:34
*** aswadr has joined #openstack-cinder05:34
*** belmoreira has quit IRC05:38
*** nihilifer has joined #openstack-cinder05:38
*** setmason has quit IRC05:41
*** rushiagr_away is now known as rushiagr05:41
*** dannywilson has joined #openstack-cinder05:48
*** dannywilson has quit IRC05:48
*** dannywilson has joined #openstack-cinder05:48
*** lpetrut has joined #openstack-cinder05:49
*** nlevinki has joined #openstack-cinder05:49
*** coolsvap|afk is now known as coolsvap05:50
*** haomaiw__ has quit IRC05:51
openstackgerritJessy Lee proposed openstack/cinder: Add volume drivers for Infortrend Storage
*** vilobhmm has joined #openstack-cinder05:57
*** haomaiwang has joined #openstack-cinder05:57
*** BharatK has joined #openstack-cinder06:00
*** mdbooth has quit IRC06:00
*** mdbooth has joined #openstack-cinder06:02
*** BharatK has quit IRC06:05
*** yamada-h has joined #openstack-cinder06:15
*** yamada-h has quit IRC06:15
*** BharatK has joined #openstack-cinder06:17
*** haomaiwang has quit IRC06:18
*** haomaiwa_ has joined #openstack-cinder06:19
*** annashen has joined #openstack-cinder06:24
*** markvoelker has joined #openstack-cinder06:28
*** annashen has quit IRC06:29
*** markvoelker has quit IRC06:32
*** ian_brown has quit IRC06:33
*** aswadr has quit IRC06:35
*** rushiagr is now known as rushiagr_away06:50
*** e0ne has joined #openstack-cinder07:08
*** markus_z has joined #openstack-cinder07:12
*** e0ne has quit IRC07:12
*** deepakcs has quit IRC07:13
*** yamada-h has joined #openstack-cinder07:13
*** anshul has joined #openstack-cinder07:14
*** yamada-h has quit IRC07:19
*** vilobhmm has quit IRC07:21
*** jordanP has joined #openstack-cinder07:22
*** vilobhmm has joined #openstack-cinder07:23
*** ndipanov has quit IRC07:26
*** e0ne has joined #openstack-cinder07:28
openstackgerritAlexey Khodos proposed openstack/cinder: NexentaStor 5 iSCSI backend driver.
*** e0ne has quit IRC07:29
*** yamada-h has joined #openstack-cinder07:32
*** yamada-h has quit IRC07:32
*** jistr has joined #openstack-cinder07:33
*** yamada-h has joined #openstack-cinder07:36
*** yamada-h has quit IRC07:36
*** dannywilson has quit IRC07:38
*** takedakn has joined #openstack-cinder07:46
*** vilobhmm has quit IRC07:46
*** dannywilson has joined #openstack-cinder07:46
*** dannywilson has quit IRC07:51
*** dannywilson has joined #openstack-cinder07:51
*** lpetrut has quit IRC07:52
*** takedakn has quit IRC07:52
*** dannywilson has quit IRC07:56
openstackgerritAlexey Khodos proposed openstack/cinder: NexentaStor 5 NFS backend driver.
*** yamada-h has joined #openstack-cinder08:11
*** yamada-h has quit IRC08:11
*** yamada-h has joined #openstack-cinder08:13
*** yamada-h has quit IRC08:13
*** yamada-h has joined #openstack-cinder08:14
*** yamada-h has quit IRC08:15
*** markvoelker has joined #openstack-cinder08:16
openstackgerritAlexey Khodos proposed openstack/cinder: NexentaStor 5 iSCSI backend driver.
*** yamada-h has joined #openstack-cinder08:18
*** yamada-h has quit IRC08:18
*** markvoelker has quit IRC08:21
*** dulek has joined #openstack-cinder08:24
*** yamada-h has joined #openstack-cinder08:25
*** yamada-h has quit IRC08:25
*** annashen has joined #openstack-cinder08:25
*** jistr has quit IRC08:27
*** yamada-h has joined #openstack-cinder08:29
*** yamada-h has quit IRC08:29
*** annashen has quit IRC08:30
*** ronenkat has quit IRC08:30
*** yamada-h has joined #openstack-cinder08:31
*** yamada-h has quit IRC08:31
*** yamada-h has joined #openstack-cinder08:31
*** yamada-h has quit IRC08:31
*** yamada-h has joined #openstack-cinder08:33
*** yamada-h has quit IRC08:33
*** yamada-h has joined #openstack-cinder08:34
*** yamada-h has quit IRC08:34
openstackgerritwanghao proposed openstack/cinder: Incremental backup improvements for L
*** yamada-h has joined #openstack-cinder08:35
*** yamada-h has quit IRC08:35
*** ronenkat has joined #openstack-cinder08:36
*** yamada-h has joined #openstack-cinder08:36
*** yamada-h has quit IRC08:36
*** yamada-h has joined #openstack-cinder08:38
*** yamada-h has quit IRC08:38
*** yamada-h has joined #openstack-cinder08:39
*** yamada-h has quit IRC08:39
*** yamada-h has joined #openstack-cinder08:41
*** yamada-h has quit IRC08:41
*** afazekas has joined #openstack-cinder08:41
*** yamada-h has joined #openstack-cinder08:43
*** yamada-h has quit IRC08:43
*** jistr has joined #openstack-cinder08:43
*** yamada-h has joined #openstack-cinder08:45
*** yamada-h has quit IRC08:45
*** ndipanov has joined #openstack-cinder08:46
*** yamada-h has joined #openstack-cinder08:47
*** yamada-h has quit IRC08:47
*** dulek has quit IRC08:53
*** Liu has joined #openstack-cinder09:05
*** liuxg has quit IRC09:06
*** dulek has joined #openstack-cinder09:12
*** geui_ has joined #openstack-cinder09:24
geui_cinder-volume not showing in service list  cinder-volume log:
openstackgerritAlexey Khodos proposed openstack/cinder: NexentaStor 5 NFS backend driver.
*** deepakcs has joined #openstack-cinder09:32
*** dims_ has joined #openstack-cinder09:44
*** dims has quit IRC09:47
*** e0ne has joined #openstack-cinder09:47
*** yamada-h has joined #openstack-cinder09:48
*** lixiaoy1 has quit IRC09:54
*** e0ne is now known as e0ne_09:54
*** Daviey has quit IRC09:55
*** Daviey has joined #openstack-cinder09:55
*** jseiler__ has quit IRC09:55
*** jseiler__ has joined #openstack-cinder09:56
*** c0m0 has joined #openstack-cinder09:56
*** geui_ has quit IRC09:57
*** e0ne_ has quit IRC10:00
*** e0ne has joined #openstack-cinder10:01
*** markvoelker has joined #openstack-cinder10:05
*** dulek_ has joined #openstack-cinder10:05
flip214help, please. local tests tell me on snapshot delete "Invalid volume: Volume still has 1 dependent snapshots.", while the CI tests do that successfully and crash later on.10:06
flip214what do I have to look for, to find out where the difference is? Is that a cinder.conf setting?10:06
e0neflip214: hi. what test did fail? do you have any errors in c-vol logs?10:07
*** dulek has quit IRC10:07
flip214e0ne: tempest.scenario.test_volume_boot_pattern.TestVolumeBootPatternV2.test_volume_boot_pattern (and same with V1)10:08
flip214locally I get "Invalid volume: Volume still has 1 dependent snapshots."10:08
flip214but in the -infra CI that seems to work fine, but it crashes later on.10:09
*** dulek_ has quit IRC10:09
flip214I need to reproduce that locally, so how can I enable volume deletion while there're still snapshots?10:09
*** zhenguo has quit IRC10:09
*** markvoelker has quit IRC10:10
*** asselin has quit IRC10:20
*** kmartin has quit IRC10:21
e0neflip214: you can't delete volume with snapshots. cinder doesn't allow it10:23
flip214e0ne: in the -infra CI it did that, if I read the logs correctly.10:23
e0neflip214: i believe, it deletes snapshot first10:24
*** yamada-h has quit IRC10:25
*** annashen has joined #openstack-cinder10:28
*** ociuhandu has joined #openstack-cinder10:30
*** annashen has quit IRC10:32
*** e0ne is now known as e0ne_10:33
*** e0ne_ is now known as e0ne10:34
*** haypo has joined #openstack-cinder10:34
*** ociuhandu has quit IRC10:35
shausyHi During my cinder auto zoning- Brocade my zone manger testing, im getting "No initiators are in the nameserver for SAN FABRIC_1 get_device_mapping_from_network "10:40
shausyDevice map for SAN context: {'BRCD_FAB_A': {'target_port_wwn_list': ['21110002ac0093f4', '23110002ac0093f4', '22110002ac0093f4', '20110002ac0093f4'], 'initiator_port_wwn_list': []}10:41
shausyim getting initator port list as empty during my cinder opertaion10:41
shausyCan some help me on this10:41
*** heyun has quit IRC10:45
*** lpetrut has joined #openstack-cinder10:45
*** timcl has joined #openstack-cinder10:49
openstackgerritKyrylo Romanenko proposed openstack/python-cinderclient: Add test for cinder extra-specs-list command output Add test for cinder list-extension command output Fix pep8 issues "E128 continuation line under-indented for visual indent" in other functions Change-Id: I28266af7a6ff23b83899c50cc230f926a44592dc Partial
*** timcl has quit IRC10:55
*** yamada-h has joined #openstack-cinder10:56
openstackgerritIvan Kolodyazhny proposed openstack/cinder: Ceph driver support retries on rados_connect_timeout
openstackgerritAnton Arefiev proposed openstack/cinder: Fix lvm manage existing volume
*** yamada-h has quit IRC11:01
*** kbyrne has joined #openstack-cinder11:03
*** aix has quit IRC11:07
*** openstackgerrit has quit IRC11:13
*** openstackgerrit has joined #openstack-cinder11:14
*** e0ne is now known as e0ne_11:21
*** markvoelker has joined #openstack-cinder11:21
*** markvoelker has quit IRC11:25
*** e0ne_ has quit IRC11:32
*** e0ne has joined #openstack-cinder11:36
*** eantyshev has joined #openstack-cinder11:41
*** julim has joined #openstack-cinder11:45
*** deepakcs has quit IRC11:55
*** markvoelker has joined #openstack-cinder11:56
*** marcusvrn has joined #openstack-cinder11:57
*** alexpilotti has joined #openstack-cinder12:01
*** akerr has joined #openstack-cinder12:03
*** BharatK has quit IRC12:05
*** aix has joined #openstack-cinder12:09
*** alexpilotti has quit IRC12:09
*** ociuhandu has joined #openstack-cinder12:10
*** dulek has joined #openstack-cinder12:13
aarefievdulek: hi, are you around12:16
*** ociuhandu has quit IRC12:18
aarefievdulek: e0ne asked me to benchmark your patches with persistence, have results12:21
*** timcl has joined #openstack-cinder12:26
*** jungleboyj has quit IRC12:26
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements
*** ociuhandu has joined #openstack-cinder12:33
openstackgerritJessy Lee proposed openstack/cinder: Add volume drivers for Infortrend Storage
*** gouthamr has joined #openstack-cinder12:38
*** shausy has quit IRC12:39
*** shausy has joined #openstack-cinder12:39
*** gouthamr has quit IRC12:42
*** nihilifer has quit IRC12:43
*** smoriya has quit IRC12:45
dulekaarefiev: Cool, I'll take a look in a moment12:46
dulekaarefiev: How big was the setup? What kind of a DB? It was Fuel? DevStack?12:47
aarefievdulek: it was fuel, 3 contr 2 vol12:47
aarefievdulek: now working on profile with devstack12:48
aarefievmysql db12:49
e0neaarefiev: just to convirm: did you setup env with lvm driver or ceph?12:50
aarefieve0ne: lvm12:50
e0neaarefiev: thanks12:50
*** gouthamr has joined #openstack-cinder12:50
*** rakesh_mishra has joined #openstack-cinder12:51
*** fifieldt has quit IRC12:53
*** tobe has quit IRC12:55
*** tobe has joined #openstack-cinder12:57
dulekaarefiev: And it was without any TaskFlow patches for deleting?12:58
aarefievdulek: yep, only your patches with scheduler create12:59
dulekaarefiev: Cool, there's overhead and it's more than 30%...13:01
dulekActually that's quite high number.13:01
aarefievdulek: I'm trying to see it on devstack with osprofiler13:01
*** tobe has quit IRC13:02
*** xyang1 has joined #openstack-cinder13:02
dulekI wouldn't expect dramatically different results.13:02
aarefievdulek: for sure, just want to see bottleneck13:03
*** Yogi1 has joined #openstack-cinder13:03
*** bill_az has joined #openstack-cinder13:04
*** Longgeek_ has quit IRC13:06
*** sgotliv__ has joined #openstack-cinder13:18
*** salv-orlando has quit IRC13:18
*** salv-orlando has joined #openstack-cinder13:19
*** sgotliv_ has quit IRC13:20
*** simondodsley has joined #openstack-cinder13:21
*** salv-orl_ has joined #openstack-cinder13:24
*** salv-orlando has quit IRC13:25
*** annashen has joined #openstack-cinder13:30
*** ociuhandu has quit IRC13:32
haypook, to stop repeating myself, i wrote a blueprint to port Cinder to Python 3:
haypoi hope that it will help ;)13:34
haypojgriffith, ^^13:34
dulekhaypo: You would probably need a spec for that...13:35
haypodulek, jgriffith told me that a spec is not needed13:35
*** annashen has quit IRC13:35
haypodulek, well, the spec is the summary of the blueprint ;)13:35
dulekCool, I think that's a reasonable decision13:35
haypoas i wrote, the nova spec can be reused since i plan to follow exactly the same path for cinder13:36
*** eharney has quit IRC13:38
*** mriedem_away is now known as mriedem13:38
*** dustins has joined #openstack-cinder13:39
*** rushil has joined #openstack-cinder13:42
jgriffithhaypo: dulek That's correct (no spec needed)13:44
haypojgriffith, so, what do you think of the plan in ?13:44
*** setmason has joined #openstack-cinder13:46
*** alexpilotti has joined #openstack-cinder13:46
jgriffithhaypo: looks reasonable to me13:47
jgriffithhaypo: thanks for writing it up13:47
haypoi wrote it because someone else asked me for my plan this morning, the third guy ;)13:48
*** rushil has quit IRC13:48
*** gouthamr has quit IRC13:48
openstackgerritVictor Stinner proposed openstack/cinder: Replace basestring with six.string_types
*** ociuhandu has joined #openstack-cinder13:49
xyang1e0ne: Ping13:50
e0nexyang1: hi13:50
xyang1e0ne: ScaleIO CI is reporting correctly now, can you check13:51
e0nexyang1: thanks for pinging me. checking13:51
xyang1e0ne: Ok, thanks13:51
*** rushil has joined #openstack-cinder13:52
mriedemjbernard: got your ceph debugging hat on?13:53
*** gouthamr has joined #openstack-cinder13:54
mriedemjbernard: race in the gate with ceph:
mriedemdelete snapshot fails since the snapshot is busy13:54
*** asselin_ has quit IRC13:54
*** ociuhandu has quit IRC13:56
*** ociuhandu has joined #openstack-cinder13:57
mriedemjdurgin: ^ since you have a spec up to nova about rbd instance snapshot, maybe you know something about rbd volume snapshots13:58
*** asselin has joined #openstack-cinder13:59
*** rmesta has joined #openstack-cinder13:59
*** jnicklin has joined #openstack-cinder13:59
*** thangp has joined #openstack-cinder14:00
*** jdurgin1 has joined #openstack-cinder14:00
mriedemlooks like cinder's rbd driver just needs a retry loop like was added to nova in
*** zhipeng has joined #openstack-cinder14:00
*** rmesta has left #openstack-cinder14:01
*** breitz has quit IRC14:03
*** breitz has joined #openstack-cinder14:03
*** ganso_ has joined #openstack-cinder14:04
openstackgerritPh. Marek proposed openstack/cinder: Re-add DRBD driver.
openstackgerritVictor Stinner proposed openstack/cinder: Use six to fix imports on Python 3
LiuIs the port of still 29418?14:07
LiuI can not cennect to var port 2941814:08
openstackgerritVictor Stinner proposed openstack/cinder: Use six.reraise() for Python 3 compatibility
openstackgerritVictor Stinner proposed openstack/cinder: Use six.reraise() for Python 3 compatibility
*** eharney has joined #openstack-cinder14:11
openstackgerritVictor Stinner proposed openstack/cinder: Get StringIO from six for Python 3 compatibility
*** ronenkat has quit IRC14:12
xyang1e0ne: Thanks!14:15
e0nexyang1: you're welcome! thank you for addressing my comments14:15
jbernardmriedem: do you know if the volume eventually leaves the busy state?14:15
mriedemjbernard: would that be in the ceph logs? i didn't dig that far14:16
mriedemthe ceph logs are posted though with the job14:16
jbernardmriedem: not sure, ill take a look14:16
*** BharatK has joined #openstack-cinder14:18
*** diogogmt has joined #openstack-cinder14:18
*** takedakn has joined #openstack-cinder14:19
*** setmason has quit IRC14:23
*** gouthamr has quit IRC14:24
*** jaypipes has quit IRC14:26
*** gouthamr has joined #openstack-cinder14:29
openstackgerritVictor Stinner proposed openstack/cinder: Replace xrange() with six.moves.range()
*** eharney_ has joined #openstack-cinder14:30
*** eharney has quit IRC14:32
*** e0ne is now known as e0ne_14:34
*** crose has joined #openstack-cinder14:36
*** eharney_ is now known as eharney14:38
*** zhipeng has quit IRC14:38
*** rushil has quit IRC14:40
*** crose has quit IRC14:41
*** takedakn has quit IRC14:41
*** shausy has quit IRC14:42
*** shausy has joined #openstack-cinder14:43
*** nkrinner has quit IRC14:43
*** rushil has joined #openstack-cinder14:44
*** tsekiyama has joined #openstack-cinder14:45
*** jdurgin1 has quit IRC14:45
*** e0ne_ has quit IRC14:45
*** mtanino has joined #openstack-cinder14:52
*** gouthamr has quit IRC14:58
openstackgerritJohn Griffith proposed openstack/cinder: Update SolidFire to use target driver model
*** Zhongjun has joined #openstack-cinder15:01
*** dims_ has quit IRC15:02
*** dims has joined #openstack-cinder15:02
*** gouthamr has joined #openstack-cinder15:03
*** nlevinki has quit IRC15:04
*** setmason has joined #openstack-cinder15:06
*** timcl has quit IRC15:07
*** rushil has quit IRC15:08
dulekgeguileo: Hi, I've commented on 185545, can you take a look if my concerns aren't invalid?15:10
geguileodulek: Ok, I'll have a look15:10
dulekgeguileo: Thanks! :)15:11
*** rushil has joined #openstack-cinder15:13
*** lcurtis has joined #openstack-cinder15:15
*** kmartin has joined #openstack-cinder15:15
*** jungleboyj has joined #openstack-cinder15:15
*** jistr is now known as jistr|mtg15:17
*** vokt has joined #openstack-cinder15:18
openstackgerritVictor Stinner proposed openstack/cinder: Use six to fix imports on Python 3
*** markus_z has quit IRC15:21
*** juzuluag has joined #openstack-cinder15:23
*** dannywilson has joined #openstack-cinder15:23
*** dustins has quit IRC15:24
*** dannywilson has quit IRC15:24
*** agarciam has joined #openstack-cinder15:24
*** dannywilson has joined #openstack-cinder15:24
*** timcl has joined #openstack-cinder15:25
*** vivek_nandavanam has joined #openstack-cinder15:26
vivek_nandavanam"A pool in Cinder has a one to one mapping with a volume on a StorageController" - Is this statement right?15:28
geguileodulek: In one of the comments (the most important one) I agree with you, so I changed my vote15:28
*** anshul has quit IRC15:29
*** vokt has quit IRC15:30
openstackgerritJoshua Huber proposed openstack/cinder: Add cinder volume driver for Blockbridge EPS
*** coolsvap is now known as coolsvap|afk15:31
*** annashen has joined #openstack-cinder15:32
openstackgerritMichal Dulko proposed openstack/cinder: Refactor API create_volume flow
SwansonAny cores floating around to give an +2/+Amen to this one?  Almost entirely comment changes.  (I get rid of a lot of single character vars, jgriffith!)15:34
*** salv-orlando has joined #openstack-cinder15:34
openstackgerritShay Halsband proposed openstack/cinder: XtremIO Volume driver Consistency Groups support
jgriffithSwanson: :)15:34
*** salv-orl_ has quit IRC15:34
openstackgerritrakesh mishra proposed openstack/cinder: set-size-limit-for-volume
*** jistr|mtg is now known as jistr15:35
dulekgeguileo: Thanks, I need to do reviews in the morning and not in the evening. ;) Now I understand the code.15:35
geguileodulek: I know what you mean  XD15:36
*** annashen has quit IRC15:36
*** vilobhmm has joined #openstack-cinder15:37
SwansonThe newish target driver stuff is still only FC or iSCSI per volume, right?  Can't combine?15:37
*** pots has joined #openstack-cinder15:38
*** shausy has quit IRC15:39
*** hemnafk is now known as hemna15:40
*** vilobhmm1 has joined #openstack-cinder15:41
*** vilobhmm has quit IRC15:42
*** leeantho has joined #openstack-cinder15:42
openstackgerritDanny Wilson proposed openstack/cinder: Volume manager should set filter_function and goodness_function
*** coolsvap|afk is now known as coolsvap15:42
dannywilsonDuncanT: kmartin: ^^15:43
kmartindannywilson, I'll take a look15:44
*** afazekas has quit IRC15:45
dannywilsonkmartin: thanks15:45
Swansonjgriffith: Thanks!15:49
*** potsmaster has joined #openstack-cinder15:51
jdurginmriedem: jbernard: thanks for pointing out that failure. I'd rather not retry blindly, and fix the root cause. the main culprit would be qemu getting shutdown uncleanly (in that case there's a 30s timeout until the volume can be deleted)15:52
*** jordanP has quit IRC15:53
*** hodos has joined #openstack-cinder15:53
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: 3PAR enable multiattach capability reporting
*** agarciam has quit IRC15:57
*** patrickeast has joined #openstack-cinder15:57
openstackgerritWalter A. Boring IV (hemna) proposed openstack/python-cinderclient: Add volume multi attach support
*** coolsvap is now known as coolsvap|afk15:58
*** rushil has quit IRC16:01
openstackgerritVictor Stinner proposed openstack/cinder: Get StringIO from six for Python 3 compatibility
mriedemjdurgin: works for me if you have a fix16:02
*** dulek has quit IRC16:03
mriedemjdurgin: i didn't dig into the qemu logs for the related instance16:03
jdurginmriedem: do you happen to be familiar with the current nova shutdown/hard stop logic as configured in the gate?16:04
mriedemjdurgin: as in we're not doing graceful shutdowns in the gate16:06
mriedemb/c it was causing timeouts16:06
*** asselin_ has joined #openstack-cinder16:09
mriedemjdurgin: actually it looks like the default in nova is to do the clean shutdown for the stop api16:10
jdurginmriedem: thanks, looking further with jbernard16:10
mriedemand i'm not seeing shutdown_timeout changed in nova.conf in devstack, so i'm assuming it's running in the gate16:12
*** _cjones_ has joined #openstack-cinder16:15
*** vilobhmm1 has quit IRC16:15
jdurginmriedem: thanks, that should be ok then. I'm suspecting a racy testcase now since a snapshot delete would only get that error when there's still a clone16:15
*** diemt has joined #openstack-cinder16:16
*** anshul has joined #openstack-cinder16:20
openstackgerritShay Halsband proposed openstack/cinder: XtremIO Volume driver requests and multipath
*** Apoorva has joined #openstack-cinder16:22
*** BharatK has quit IRC16:22
*** krtaylor has quit IRC16:22
*** e0ne has joined #openstack-cinder16:27
*** r-daneel has joined #openstack-cinder16:27
SwansonLots of python 3 things all over openstack.  Is there a plan or timeline for jumping to 3.x?16:29
Swansonhemna: did I just cause a migraine?16:30
openstackgerritVictor Stinner proposed openstack/cinder: Replace with next(it) for py3 compat
haypoSwanson, i just decided to flood openstack with my patches ;)16:30
hemnapy3 leads to pain, pain leads to anger, anger leads to suffering.16:30
haypoSwanson, someone, it was decided to port openstack. the thing is that there is no more major dependency which is incompatible with python 216:30
haypohemna, good news. i like all these things!16:31
angela-shemna: +1, i have a headache just thinking about it16:31
jgriffithIt's really not so bad16:31
jgriffithhaypo: has a good strategy going forward16:32
jgriffithand it needs to be done16:32
Swansonhaypo: so the net of that is it can work under 3 and from other reactions not everyone will move to it.16:32
haypojust kidding16:32
SwansonI'm too stupid to know I'm in for pain before it hits.16:33
haypoSwanson, the plan is to add python 3 support. you're free to keep python 216:33
*** BharatK has joined #openstack-cinder16:34
*** jaypipes has joined #openstack-cinder16:35
SwansonJust trying to schedule some tasks.  One of them requires python 2.7.9 which ubunutu 14.04 LTS doesn't seem to support.  It we were going to py3 I'd ashcan that whole effort.16:35
haypoSwanson, why do you need this specific version of python 2.7? (2.7.9)16:36
SwansonOddly because it has some SSL stuff pulled from 3.x.16:36
SwansonAttack poodles or something.16:37
*** alau2 has joined #openstack-cinder16:38
haypoSwanson, ah yes, the security backports16:39
haypoSwanson, it's the whole ssl module + os.urandom enhancements16:39
openstackgerritJohn Griffith proposed openstack/cinder: Update SolidFire to use target driver model
*** vilobhmm has joined #openstack-cinder16:40
*** vilobhmm1 has joined #openstack-cinder16:41
Swansonhaypo: Yep.  We have a library in no man's land.  Needs those enhancements but doesn't work under 3.x.  I could push to have that moved to 3.x if Openstack is going to run under 3 anytime soon.  (Like Liberty.)16:41
*** c0m0 has quit IRC16:41
*** anshul has quit IRC16:42
haypoSwanson, it's a pain to use a specific version of python, /usr/bin/python is part of Ubuntu!16:42
Swansonhaypo:  That would be my argument.16:43
*** vilobhmm has quit IRC16:44
haypoi remember that ubuntu has/had 3.4.0 whereas this version contains a major regression in the garbage collector :-/16:44
openstackLaunchpad bug 1367907 in python3.4 (Ubuntu Trusty) "Segfault in gc with cyclic trash" [Undecided,Triaged]16:45
haypoah, it looks to be fixed now16:45
*** annashen has joined #openstack-cinder16:45
*** crose has joined #openstack-cinder16:46
e0nejgriffith, dulek: we've got a bad news. profiler results with persistance taskflow:
e0nethanks aarefiev  to make it happen16:46
openstackgerritVictor Stinner proposed openstack/cinder: Drop L suffix from long integers
e0nejgriffith: look on the actions with 'wf_db' type16:47
e0neharlowja: ^^16:47
*** haypo has quit IRC16:49
jgriffithe0ne: so it's kinda cool we have the smoking gun in there :)16:49
*** ndipanov has quit IRC16:50
e0nejgriffith: it's a good feature, but too big overhead, imo16:51
*** salv-orl_ has joined #openstack-cinder16:52
*** rushil has joined #openstack-cinder16:53
*** ronenkat has joined #openstack-cinder16:54
*** salv-orlando has quit IRC16:54
*** rakesh_mishra has quit IRC16:55
*** jungleboyj has quit IRC16:55
*** jungleboyj has joined #openstack-cinder16:56
*** vokt has joined #openstack-cinder16:59
*** mriedem is now known as mriedem_afk17:00
*** gouthamr has quit IRC17:01
*** Yogi11 has joined #openstack-cinder17:02
*** gouthamr has joined #openstack-cinder17:02
*** rushil has quit IRC17:03
*** annegentle has joined #openstack-cinder17:03
*** Yogi1 has quit IRC17:03
*** angela-s is now known as angela-s|afk17:06
*** bnemec has joined #openstack-cinder17:06
*** ronenkat has quit IRC17:07
*** anshul has joined #openstack-cinder17:08
*** e0ne has quit IRC17:09
*** gouthamr has quit IRC17:11
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: LVM add multiattach flag capability
*** breitz has quit IRC17:12
*** gouthamr has joined #openstack-cinder17:12
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: WIP: LVM add multiattach flag capability
*** patrickeast has left #openstack-cinder17:15
*** patrickeast has joined #openstack-cinder17:15
*** ronenkat has joined #openstack-cinder17:15
*** gouthamr has quit IRC17:16
*** marcusvrn1 has joined #openstack-cinder17:16
*** gouthamr has joined #openstack-cinder17:16
*** breitz has joined #openstack-cinder17:18
*** Yogi11 has quit IRC17:22
*** anshul has quit IRC17:24
*** juzuluag has quit IRC17:25
*** gouthamr has quit IRC17:28
marcusvrn1thingee: hi! did you have a chance to take a look on the new CI?17:29
marcusvrn1hmm ok :)17:30
*** lpetrut has quit IRC17:30
harlowjaaarefiev yt, so for any idea how to easily interpret that :-P looks like bunch of queries, to wf_db, perhaps can be optimized more (aka, via some taskflow changes to not store anything at all for tasks that are really 'transient')17:31
harlowjai do see its querying things, which is sorta expected :-P17:31
harlowja*perhaps also need better 'bulk' query apis in taskflow to17:31
thingeemarcusvrn1: please stop asking me too. You will know I had a chance to view it when I either move it in the etherpad or contact you about a problem.17:32
thingeeI realize it might be a common misconception that this is all I do.17:32
hemnayou do things other than look at CI ?17:33
harlowjado u look at IC the rest of the time?17:33
harlowja50% CI, 50% IC17:33
* thingee takes a look at the taskflow ci17:33
thingeetsk tsk17:33
harlowjai only run IC17:34
harlowjanot that CI stuff17:34
thingeeIC = I care17:34
marcusvrn1thingee: sorry, I was trying to know if it was ok, because the deadline is close17:34
harlowjaits the caring that matters the most right17:34
*** BharatK has quit IRC17:35
*** jistr has quit IRC17:35
thingeemarcusvrn1: imagine having 14 different vendors coming to you about the same thing.17:35
marcusvrn1thingee: ok, sorry, I'll not ask again17:35
harlowjathingee there u go, cheer u right up, lol17:35
*** anshul has joined #openstack-cinder17:36
thingeeharlowja: this is the original pitch I gave to bswartz for his project name17:37
harlowjau called it the 'care bear' project?17:37
thingeeI actually had someone at netapp asking me if I would release the idea to them after mentioning it on IRC.17:37
* harlowja thinks that idea might be owned by matell or something :-P17:38
harlowjabut ya +1 to care bear project, lol17:38
thingeeI just meant the sharing is caring part17:38
thingeeunless they own that17:38
thingeewait, there was some gum commercial that used that too17:39
thingeeoh well17:39
harlowjahmmm, only gum commerical i remember is the big red ones, lol17:39
* harlowja hasn't seen those in forever17:39
thingeejuicy fruit. sharing is caring17:39
bswartzI don't see how you can trademark "sharing" and then not share it -- that's just evil17:40
harlowjaamericans be crazy, lol17:40
harlowjathingee damn, i might have nightmares now :-/17:40
bswartzrofl @ 28 seconds17:41
* harlowja goes back to hiding under his desk17:41
bswartzyeah that fish is nightmarish17:41
thingeejgriffith: saw your comment about blockbridge ci, I haven't looked at their other reports, but I did notice their initial report for their driver being added in, the logs seem to work
thingeejgriffith: from what I saw you added here
harlowjato cleanse your pallet
harlowjaahhh, no more nightmares now17:43
harlowjawhatever, i don't do english :-P17:44
thingeeasselin_, jgriffith: I need to add my own under here. don't have the subunit stuff being added in still :(17:44
thingeeI know jgriffith did an update to his CI stuff, I just haven't updated it yet.17:45
*** akerr is now known as akerr_away17:45
*** cbader has joined #openstack-cinder17:51
*** dannywil_ has joined #openstack-cinder17:53
*** IlyaG has joined #openstack-cinder17:54
*** dannywilson has quit IRC17:57
*** gouthamr has joined #openstack-cinder17:57
*** lpetrut has joined #openstack-cinder18:00
hemnajgriffith, posed a question in your FCTarget review.   see what you think....
*** hemna is now known as hemnafk18:01
thingeeasselin_: done18:01
thingeeasselin_, jgriffith is there a plan with these?18:02
*** akerr_away is now known as akerr18:02
*** Yogi11 has joined #openstack-cinder18:02
thingeeasselin_: here's an example of a blockbridge ci
thingeeasselin_: that's one that was posted for their own driver code to be merged18:03
patrickeastthat one looks correct18:04
patrickeastthingee: the one in that etherpad is from may 15th? i think that was before they switched to using asselin’s scripts18:05
thingeepatrickeast: got it18:05
thingeepatrickeast: I'll take a look at the other reports to see if they're consistent18:06
*** ociuhandu has quit IRC18:06
patrickeastthingee: i would hazard a guess that any posted before the driver got merged are potentially bad18:07
thingeepatrickeast: yea looks good from some recent runs18:07
*** rushil has joined #openstack-cinder18:12
patrickeastthingee: also, to answer your question on what the plan was with those systems, this was the converation about it yesterday
patrickeastalthough i’m not sure if anyone has actually signed up to start contacting operators18:12
thingeeI think I can do that.18:12
thingeeas long as no one else has started18:12
*** annegentle has quit IRC18:14
*** jdurgin1 has joined #openstack-cinder18:15
thingeesmcginnisafk: lgtm here
*** jdurgin has quit IRC18:17
*** annegentle has joined #openstack-cinder18:17
*** krtaylor has joined #openstack-cinder18:18
thingeemarcusvrn1: why is the hbsd2 CI reporting FC, if the driver you're proposing ISCSI?
*** salv-orl_ has quit IRC18:20
patrickeastwoa that storpool ci is a beast… 14 min tempest run?!18:20
*** salv-orlando has joined #openstack-cinder18:21
thingeepatrickeast: am I crazy about hbsd2 CI testing FC, but the driver being proposed is iscsi?
patrickeastthingee: uhh that does seem odd18:24
thingeekmartin, hemnafk: are we not going to see a CI for this?
thingeemarcusvrn1: I've updated the etherpad about this. Not exactly sure what's going on there.18:24
patrickeastthingee: do they already have an FC driver?18:25
thingeepatrickeast: well they were telling me in the review this is for the proposed ISCSI driver (the review link I gave you).18:25
*** salv-orlando has quit IRC18:26
patrickeastthingee: maybe i’m missing something, i don’t actually see a new driver class added in that review18:26
patrickeastit does seem to be adding support for iscsi related stuff18:27
*** smcginnisafk is now known as smcginnis18:28
thingeepatrickeast: yeah I don't get it18:28
patrickeastso there is an iscsi and fc driver already in tree, which is what that ci is testing18:28
patrickeastmaybe its just expanding the driver to support more hardware?18:28
jgriffithpatrickeast: thingee it's just adding a product line to the driver; no?18:29
thingeejgriffith: yeah, anyways I was saying the proposed patches are iscsi related, but the new ci they're introducing is FC18:29
jgriffiththingee: oh... I dunno, I've never been able to follow any of their stuff18:30
jgriffiththingee: not going to try and start now18:30
thingeejgriffith, asselin_ I'll start contacting ops on the CI list here, unless that's already been started18:30
thingeejgriffith, asselin_ specifically this list
jgriffiththingee: we proposed yesterday that we should *share* that effort18:31
jgriffiththingee: and divie up the work18:32
thingeeok sure. so do I just mark the ones I'll do?18:32
jgriffiththingee: unless you still want to be the single point of contact18:32
jgriffithI know you wanted it that way before, which is fine18:32
thingeenot opposed to it, as long as I can rely on people18:32
thingeeI can rely on myself, which is why I just took it on18:32
jgriffiththingee: Your call18:32
jgriffiththingee: honestly I think the time for that kind of effort/attention is over18:32
jgriffiththingee: if it were me, send one email out warning.. then things get turned off and removed18:33
thingeejgriffith: understood18:33
thingeejgriffith: I added datera to the list. still need to update sos-ci18:33
thingeefor subunit18:33
jgriffithhaha... that's the least of the problems out there right now IMHO18:33
thingeewell, just trying to not be biased here.18:34
jgriffithI hear ya18:34
thingeeyeah I agree with you on the effort. sharing it seems like a good plan18:34
jgriffiththingee: just seems crazy to have you spending such a large percentage of your time on it18:35
jgriffithor any one person for that matter18:35
jgriffithshould be split up among cores/regular-contributors IMO18:35
jgriffithjust so long as everybody is on the same page ;)18:35
uberjayasselin_: thingee I'm currently looking at our CI -- it appears to be suddenly having issues uploading the logs to our logserver, which is unfortunately failing the build :(18:38
uberjayuh, also... the logs linked from the thirdpartyCI-issues list are not logs from us. (misfiled? they say "openvstorage" in the url...)18:40
*** hemnafk is now known as hemna18:43
hemnathingee, we are working on CI for that one....but it's not ready yet.  hence my -2 on it still.18:43
* asselin_ reads scrollback18:44
*** breitz has quit IRC18:44
*** breitz has joined #openstack-cinder18:44
jgriffithuberjay: that's me :)18:44
jgriffithuberjay: which CI is yours?18:44
uberjayjgriffith: Blockbridge CI18:45
uberjayjgriffith: I think one of my build failures was due to this, which appears to be a Jenkins bug:
*** annegentle has quit IRC18:45
uberjaymaybe just don't have the fix yet, not sure18:45
jgriffithuberjay: cool... removed that entry.  Sorry about any confusion18:46
asselin_thingee, I'm helping the guy set up ci for:  HP XP storage. Posting locally...should have it up later today/tomorrow18:46
*** jdurgin has joined #openstack-cinder18:46
thingeejgriffith, patrickeast, smcginnis, xyang1, winston-d, hemna, eharney, jungleboyj  current driver list ... I will be cutting driver submissions tomorrow18:47
asselin_thingee, we can add sign ups to the etherpad so we don't duplicate effort18:47
hemnathingee, ok18:47
thingeeas a last attempt, I will be contacting maintainers on the urgency18:47
thingeeasselin_: +118:47
*** jdurgin1 has quit IRC18:48
thingeejgriffith: patrickeast smcginnis xyang1 winston-d hemna eharney jungleboyj I'd like the current list that is marked ready to be merged before june 19th. This list might grow after today depending on if people are able to fix there stuff in time...but this might be the final list.18:48
* thingee plays final count down by europe18:48
asselin_uberjay, could you clarify on the etherpad if not done already?18:48
* jungleboyj hear dunah dunah ... dunah dun dun dun ... dunah dunah ... 18:49
thingeemarcusvrn1: ^18:49
xyang1thingee: Sure18:49
jgriffiththingee: so about that, I've been meaning to talk to folks about my driver18:49
*** aix has quit IRC18:49
asselin_uberjay, what's the issue with uploding log files? out of disk space maybe?18:49
jgriffithand FC inparticular18:49
uberjayasselin_: ah, it appeared to just affect one build. one sec.18:50
jgriffithas you can see it's not a new driver, it's target option (can be both)18:50
jgriffiththingee: need to know if it's going to fall under deadline or not18:50
jgriffiththingee: if it is, I'll drop anything else that i'm working on and finish it18:50
thingeetarget drivers are excluded.18:51
*** e0ne has joined #openstack-cinder18:51
*** e0ne is now known as e0ne_18:51
thingeeI think we still need to have a discussion on what is a driver since I know people disagree with my thoughts on that.18:51
jgriffiththingee: good to know, thought that was the case at one point, but figured I should verify so no surprises18:51
patrickeastis it a target driver? or adding FC support to existing volume driver?18:51
patrickeastthey are different things… right? (might be wrong)18:51
hemnathingee, I signed up for HGST, since no one was on that one, and I've been reviewing the os-brick connector for HGST18:51
jgriffithpatrickeast: well, in my case it's adding a target-driver18:51
patrickeastjgriffith: ah ok18:52
jgriffithpatrickeast: but honestly they're kinda the same thing18:52
hemnaearlephilhower, ping18:52
earlephilhowerhemna: howdy18:52
thingeepatrickeast: IMO, if you can use the same volume driver class, and just configure the other target driver, who cares.18:52
patrickeastyea that makes sense18:52
thingeepatrickeast: if it's a whole new volume driver class, you're defining a new driver.18:52
*** jaypipes has quit IRC18:52
patrickeastwhere does something like that fall with ci?18:52
hemnaearlephilhower, heyas.   now that we have 3rd party CI working against os-brick patches, I think you should add testing against the HGST connector18:53
thingeesince you are, well, inheriting from something like cinder.volume.driver.VolumeDriver... that's a new driver.18:53
uberjayasselin: this is what happened:
*** e0ne_ is now known as e0ne18:53
hemnaearlephilhower,  take a look at what's being reported on your patch now.  You can see that HP CI is reporting as is pure storage.18:54
thingeejgriffith: that's pretty neat. If all you have to do is tell solidfire's existing code to just use this new target driver, I don't care. I think you're also fixing FC for everyone as well18:54
earlephilhowerhemna: Let me see how to update.  Is there a new devstack override for OS_BRICK_REPO?  I'm git pulling/installing before runs.  Looking at your link.18:54
hemnaearlephilhower, I would like to see HGST reporting there as well, since your os-brick patch is specifically for adding HGST trasport :)18:54
uberjayasselin: *maybe* some sort of race with the one-shot nodes? i've only seen it once, so maybe not a recurring problem after all18:54
asselin_uberjay, ok18:54
hemnaearlephilhower, yes, asselin_ got it in recently.  you can ping him on how to enable your CI for it.18:54
smcginnisjgriffith: I haven't looked, but what about the fc zone manager decorators? Is that handled in your approach?18:54
thingeejgriffith: did I get that right, that it's just a matter of configuring a different target driver?18:55
jgriffithpatrickeast: just for the sake of argument I'm planning to have CI18:55
jgriffithjust easier that way18:55
jgriffiththingee: and to your questions above "yes"18:55
thingeejgriffith: excellent, yeah no sweat.18:55
jgriffiththingee: it's just adding the target driver, and yes, I've added patches to make it available to everyone18:55
earlephilhowerhemna:  Oh, no problem.  Will do some reading then ping asselin if I'm stuck.  Can it wait until tomorrow or Mon?  Lots of meetings this afternoon for /me.18:56
hemnaearlephilhower, well, our driver deadline for Cinder is what...tomorrow ?18:56
thingeejgriffith: yeah you can even quote my email about this, "This does not include connector drivers in os-brick. This will be a18:56
thingeeseparate discussion."
Swansonsmcginnis: There are no decorators in the actual FC target driver.18:56
thingeeI guess me saying connector code is a bit ambiguous18:57
uberjayasselin: ugh, i take that back, got another failure with the same signature. i'll update the etherpad.18:57
jgriffiththingee: meh.. I knew what you meant ;)18:57
hemnaSwanson, yah there aren't currently, it's something that we've tossed around in the review.18:57
jgriffithhemna: oh.. yeah, BTW did you see my response on that?18:57
thingeejgriffith: but I think that's pretty neat. this is what I wish all drivers would do for support the different fabrics =/18:58
earlephilhowerhemna:  Okay, will try during lunch.   If devstack has new option may be plug and play18:58
jgriffithhemna: and my question WRT "what if I don't want zones"18:58
hemnajgriffith, yah.  I'm still trying to wrap my head around the FC target18:58
jgriffithhemna: k18:58
hemnaand really the purpose of it at this point.   it's confusing.18:58
* winston-1 18:58
jgriffiththingee: me too :)  That's why I started it back in Juno18:58
jgriffithhemna: what's confusing about it?18:58
hemnajgriffith, if you don't want zones, then your driver doesn't return initiator_target_map18:58
jgriffithhemna: maybe I can clarify18:58
hemnaor you don't configure cinder to use the FCZM18:58
jgriffithhemna: right18:59
*** dannywil_ has quit IRC18:59
jgriffithhemna: but it appears that forcing your decorator means you can't do that18:59
hemnabut if you never add the decorators, then the FCZM can never be used with that driver.18:59
hemnaif you add the decorator18:59
jgriffithhemna: sure... but those classes are abc'd18:59
jgriffithhemna: do you understand how the abc stuff works?  I can try and explain if not18:59
hemnayou can simply not configure cinder.conf to enable the zone manager and it all works.18:59
hemnabut at least with the decorator in place, then folks have the option of enabling the fczm19:00
jgriffithhemna: right.. and even if I use your decorator it just becomes a noop right?  IIUC19:00
*** e0ne is now known as e0ne_19:00
jgriffithhemna: yeah.. for sure19:00
hemnaif you want open zoning, then don't configure cinder to use fczm.  done.19:00
jgriffithhemna: totally get it, just wanted to verify I was looking at the code correctly19:00
hemnait's a noop then,.19:00
*** winston-d has quit IRC19:00
*** dannywilson has joined #openstack-cinder19:01
jgriffithSo even if you put a decorator on an abc (and I can do that if it makes people feel better), it will NEVER get called19:01
jgriffithunless there's some magic you can do there that I'm not aware of19:01
hemnathe decorator gets called, but internally it looks to see if the fczm is enabled or not.   if it's not, noop.19:01
jgriffiththe reason being is you've defined that as an abstract base class that MUST be overridden19:01
hemnaalso, if there is no initiator_target_map in the payload it's a noop.19:01
*** winston-1 is now known as winston-d19:02
*** ronenkat has quit IRC19:02
jgriffithhemna: no, I'm saying that in an abc deocrated method it will enever get called if it's not in the child class... unless you're saying you know that works and I'm wrong19:02
jgriffithhemna: oh... sorry... you're back on the other subject :)19:02
jgriffithnever mind :)  Yeah.. we're good there I think19:02
jgriffithI"m trying to clarify the base target stuff that you said you didn't quite get19:03
hemnaso, is the intention with the FCTarget being an FC Base target class that vendors extend ?19:03
hemnaFC3PARTarget ?19:03
hemnaso the FCTarget decorators wouldn't be used19:03
*** winston-d has quit IRC19:03
hemnajust trying to follow the chain of usage for FCTarget19:03
*** winston-d has joined #openstack-cinder19:04
jgriffithhemna: yes19:04
hemnafrom what I can tell the lvm driver can use any target19:04
jgriffithhemna: EXACTLY!!!19:04
jgriffithhemna: that's the entire point19:04
hemnaso 3par creates FC3PARTarget19:04
jgriffithhemna: here's a more concrete example:
hemnathen cinder.conf in my 3par driver section, I specify which target to use.19:04
*** ronenkat has joined #openstack-cinder19:05
jgriffithhemna: just s/SanISCSITarget/FCTarget/19:05
jgriffithhemna: if you want/need.... or like in my case just auto determine, or pick based on a volume-type19:05
hemnaok, so if we put the decorator on the FC3PARTarget, then our 3PAR driver would just call return self.target_driver.initialize_connection(payload)19:06
openstackgerritAlex Meade proposed openstack/cinder: NetApp E-Series driver: Remove caching logic
openstackgerritAlex Meade proposed openstack/cinder: NetApp E-Series: Refactor class structure for FC
openstackgerritAlex Meade proposed openstack/cinder: NetApp E-Series: Add Fibre Channel Support
* hemna is trying to figure out what/if at all/would go in a FC3PARTarget 19:07
*** kjnelson has joined #openstack-cinder19:08
*** annashen has quit IRC19:08
hemnaI was hoping the generic FCTarget's initialize_connection would contain the contents of the fczm_utils.AddFCZone, and then vendor FC drivers would call return self.target_driver.initialize_connection(payload), which would induce the FCZM19:08
*** mriedem_afk is now known as mriedem19:08
jgriffithhemna: like two or three methods and that's all :)19:08
jgriffiththat's the beauty of it19:09
hemnamaybe I'll put a gist together to demostrate what I'm thinking19:09
hemnaand I'm sure I'm all off base here.19:09
hemnajust trying to understand it is all.19:09
hemnasorry for being slow19:09
*** gouthamr has quit IRC19:12
hemnathingee, if the HGST folks don't get the CI for os-brick plugged in by tomorrow, can I request an exception for them now?19:12
hemnasince asselin_ and I didn't get the 3rd party CI capability for os-brick until just recently19:13
asselin_hemna, I think there's a different deadline for os-brick19:13
hemnaasselin_, the problem is, that the HGST Cinder volume driver needs an os-brick Connector to land19:13
winston-dspeaking for CI for os-brick19:13
hemnaasselin_, so the HGST driver shouldn't land until the os-brick patch lands, IMHO19:14
asselin_hemna, interesting...ok...will let thingee decide19:14
*** annashen has joined #openstack-cinder19:14
*** gouthamr has joined #openstack-cinder19:15
asselin_hemna, honestly, it's easy to add ci to os-brick once you have it for cinder19:15
thingeeare we requiring ci for target drivers?19:16
asselin_hemna, also they should add a cross-project depends-on to their patch if they don't already19:16
hemnaasselin_, it should be, if you have the time to work on it.  earlephilhower said he might not have time until tomorrow or Monday.19:16
*** annegentle has joined #openstack-cinder19:16
hemnaasselin_, and I'm just trying to think ahead for him to help make sure he doesn't miss the deadling because of us taking so long to get the os-brick CI stuff in place.  It's not that fair for him and the HGST driver.19:17
hemnathingee, I don't think we figured that out.19:17
asselin_hemna, well the deadline for drivers is next week19:17
asselin_hemna, so he has time to set it up for os-brick19:17
asselin_hemna, he also needs a nova patch to land...19:18
hemnaoh ok L-1 is June 2319:18
*** Zhongjun has quit IRC19:18
hemnaI was worried it was tomorrow :)19:18
hemnaso maybe I'm premature asking for an exception for HGST already.19:19
thingeehemna: the deadline tomorrow just for us to have reasonable amount of time to work with drivers that actually have a chance19:19
hemnanevermind! :)19:19
thingeeto avoid last minute submissions19:19
winston-dspeaking of os-brick CI, the rbd connector is kind of different, 'cos my understanding is Nova won't be using it even when it lands.19:19
hemnathingee, ok my bad.19:19
winston-dthe only usage for rbd connector is for volume migration19:19
hemnawinston-d, so I had that question on the review.19:19
winston-dwhich unfortunately, it's not covered by CI19:19
hemnawinston-d, it's quite a bit different than the libvirt volume driver code.19:19
hemnawinston-d, but I do have a WIP in progress for nova to use os-brick19:20
*** anshul has quit IRC19:20
hemnawinston-d, but not all libvirt volume drivers will be converted in the first pass though.19:20
eharneyyeah i'm not sure i understand the thinking there on the rbd connector19:20
hemnaeharney, +119:20
hemnaI didn't get it either.19:20
eharneyi mean... why the demand for CI now19:20
hemnabut I'm not rbd expert19:20
eharneythe connector is needed for generic volume migration19:21
eharneywhich isn't CI'd...19:21
winston-dunless we have volume migration case covered in CI, having a CI won't help validating RBD connector19:21
hemnaI'd like to make sure that the RBC connector actuall functions the same as the nova libvirt volume driver19:21
hemnaI'm not sure what the heck the current patch in os-brick actually does19:21
winston-dsimply because Nova won't use that connector19:21
eharneyright... what winston-d says19:21
hemnaneed an RBD 'guy' to look at it.19:21
eharneyblocking the RBD connector for CI is holding things up for no benefit at this time19:21
eharneysince it's not being used for anything but migration19:22
hemnawinston-d, copy volume to image and image to volume can exercize os-brick in cinder19:22
winston-dagain: the only usage for RBD connector is for volume migration19:22
hemnadepending on the cinder volume driver of course19:22
winston-dhemna: let me check rbd's copy_image_to_volume19:22
eharneyhemna: it _can_, but since those don't use brick now, i'm not sure why that's a demand19:22
earlephilhowerasselin:  I'm trying the OS_BRICK_REPO / OS_BRICK_BRANCH in my pre_test_hook, but _BRANCH never seems to apply in the scripts19:23
asselin_earlephilhower, you talking about os-brick ci or cinder ci?19:23
winston-dhemna: no, rbd driver won't use connector to do copy_image_to_volume or upload_volume_to_image.19:24
earlephilhowerasselin_:  os-brick CI, but using just plain Cinder CI.  Defined overrides just like for NOVA and CINDER, but logs don't show joy19:24
*** marcusvrn has quit IRC19:24
earlephilhowerWindering if need to apply elsewhere?19:24
asselin_earlephilhower, that won't need to do something differently19:24
hemnawinston-d, yah it looks like the rbd cinder volume driver doesn't currently use os-brick (which makes sense)19:24
asselin_earlephilhower, you need to follow this template
winston-dhemna: right, so that means having a CI won't help unless CI has volume migration test cases19:25
asselin_earlephilhower, this is the main difference:
hemnawinston-d, well specifically to CI rbd, but CI'ing that the RBD connector doesn't break anything else is still ok.19:25
asselin_export DEVSTACK_PROJECT_FROM_GIT=os-brick19:25
asselin_earlephilhower, do this is a new job19:26
hemnaso if others can review the RBD connector that'd be great19:26
marcusvrn1thingee: sry, I was out. I think there's a misunderstanding. We have one driver  HBSD that has CI for FC and iSCSI (Hitachi HBSD CI). That patch ( is not a new driver, it's a patch for extending funcionality for a new hardware. Once you asked us to create a CI, I started to create a CI for each hardware/backend. We didn't create a CI for the iSCSI backend for new hardware yet19:26
*** salv-orlando has joined #openstack-cinder19:26
hemnaI'm RBD challenged19:26
winston-dhemna: yeah, i will review the code19:26
eharneyhemna: does that mean you don't want to block the connector patch for CI then?19:26
hemnaeharney, yah I won't block it19:26
hemnait's not being used anywhere :)19:27
hemnamy -1 was simply to raise the question and start the discussion19:27
earlephilhowerasselin_:  Can do, but what about the main Cinder CI?  Still OK to to do a git fetch from the reviews bug and python before aything?19:27
eharneyhemna: yeah, it's leading up to the generic migration work19:27
marcusvrn1thingee: so, the driver is the same, HBSD driver, what we are doing in this patch is expanding its functionalities for new hardware19:27
hemnaon 1) CI and 2) the difference between the patch and nova's rbd libvirt volume dirver.19:27
hemnabecause if on 2) we could make the patch work like nova's libvirt volume driver and get a release of os-brick out that had the new RBD connector, then I can update my nova WIP patch to change the existing rbd libvirt volume driver to use os-brick.19:28
asselin_earlephilhower, yes...that's why I asked which ci. the job definition is different to test os-brick and cinder19:28
hemnabut I guess we can do that later19:28
thingeemarcusvrn1: ok hang on let me pull the page back up19:28
asselin_earlephilhower, which are yourpatches?19:28
marcusvrn1thingee: which etherpad have you updated? ?19:28
hemnaeharney,  my WIP in nova19:28
hemnaeharney, there is also a glusterfs libvirt volume driver19:29
hemnasmells like we should have a glusterfs os-brick connector as well19:29
thingeemarcusvrn1: so hbsd_horcm is FC only?19:29
eharneyhemna: i thought we had a remotefs connector19:29
hemnaI'd like to eliminate all of the internals of the libvirt volume drivers19:29
earlephilhowerasselin_: Got it.  Let me try it out...19:30
eharneythe glusterfs stuff in nova doesn't do anything other than mount and unmount, i think?19:30
hemnaand then eventually get rid of libvirt volume drivers all together and put a shim to os-brick connectors.19:30
hemnaeharney, I don't know anything about gluster19:30
*** gouthamr has quit IRC19:30
eharneyhemna: what's being done for NFS?19:30
*** ronenkat has quit IRC19:30
hemnaso I was confused about the remotefs vs. gluster vs. nfs, etc. stuffs19:30
marcusvrn1thingee: yes, and we are adding the iSCSI support19:30
eharneyyeah, gluster and nfs should be about the same afaik19:30
thingeemarcusvrn1: where's the horcm iscsi ci?19:31
hemnaeharney, ok maybe you can give me a hand understanding all of that19:31
*** annashen has quit IRC19:31
eharneyhemna: well i think gluster and nfs both just go through the remotefs connector right?19:32
marcusvrn1thingee: we don't have it yet19:32
winston-dhemna: RBD uses LibvirtNetVolumeDriver:
hemnaeharney, yah they both go through the RemoteFsConnector19:32
winston-dhemna: which isn't part of os-brick, at least for now.19:32
thingeemarcusvrn1: ok, that would be where my confusion is. So we would need one for each. For our requirements we require all fabrics to be tested.
hemnawinston-d, yah that's issue #2 I had19:32
winston-dhemna: i mean for Nova19:33
thingeemarcusvrn1: that's why you see some products with _iscsi or _fc19:33
thingeefor their ci's19:33
*** gouthamr has joined #openstack-cinder19:33
hemnawinston-d, I'd like the RBD connector in os-brick function the same as the nova code....but I'm confused by the LibvirtNetVolumeDriver being shared between iscsi, rbd19:33
marcusvrn1thingee: hmm so we need one CI for each backend, right?19:33
thingeemarcusvrn1: yes19:34
hemnawinston-d, eharney
hemnathe source_protocol thing.19:34
eharneyhemna: Nova uses the NetVolumeDriver path because RBD is handled from within qemu19:34
eharneyhemna: so it doesn't mount anything on the compute node itself19:34
winston-deharney: exactly19:34
winston-dhemna: the only thing NetVolumeDriver does in Nova, is to manipulate a few bunch of libvirt config (xml), that's all.19:34
hemnaeharney, so do we need a NetVolumeConnector in os-brick ?19:35
eharneyhemna: i guess that's kinda up to how things are divided between brick and nova... but i'm not sure it would actually do much?19:35
*** e0ne_ is now known as e0ne19:35
hemnathe end goal being nuking virt/libvirt/volume.py19:35
winston-dhemna: personally I don't think so.19:35
hemnaand shimming into os-brick to do all of it.19:35
*** gouthamr has quit IRC19:35
eharneythat means brick has to handle libvirt internals?19:35
winston-dhemna: unless you want to have os-brick/virt/libivrt os-brick/virt/xenapi stuff in os-brick19:36
hemnaeharney, I don't see where it's hacking libvirt internals19:36
marcusvrn1thingee: hmmm now I understood, for each backend, even it's the same driver, we have to setup a CI19:36
hemnaget_config just fetchies things from connection_info, and massages it into something that the caller expects.19:36
eharneyhemna: the conf.source_* stuff is generating a libvirt XML object19:36
eharneyso presumably if brick handled this, it would just read fields and them pass them to Nova's libvirt driver which would still do mostly the same thing it does now?19:37
hemnaeharney, yah19:37
jgriffithhemna: wait... what?  Nuking libvirt/volume?19:37
hemnathe code that's inside of get_config would exist inside of connect_volume for the connector ?19:37
thingeeright... different code paths, etc. Maybe you use zone managers... I don't know. I need to know that product works in that configuration.19:37
thingeemarcusvrn1: ^19:37
jgriffithhemna: that seems out of scope from what I imagined19:38
winston-dstill, even if we do that, it has nothing to do with rbd connector19:38
*** anshul has joined #openstack-cinder19:38
eharneyhemna: i don't think so, because then os-brick has to load libvirt python19:38
hemnaeharney, ick19:38
jgriffithhemna: I thought it was/should be for the host attach pieces only19:38
*** gouthamr has joined #openstack-cinder19:38
eharneywhich seems like not the right place19:38
*** juzuluag has joined #openstack-cinder19:38
hemnajgriffith, the libvirt/ is nothing more than classes for each of the transports that do connect_volume and disconnect_volume19:39
hemnawhich is what os-brick connectors do19:39
hemnathe dupe code19:39
e0nejgriffith: hi. i'm sorry, i'd got internet connection issue yesterday. i'm ready to discuss issues and changes with 'detailed' views it you'll have a time19:39
marcusvrn1thingee: nice! I got it19:39
jgriffithhemna: there's more in there though19:39
winston-dhemna: not exactly, the get_config function here: is specific to Nova.19:39
hemnawinston-d, that's the part I'm trying to understand19:40
hemnawinston-d, and that's specific to libvirt itself no?19:40
winston-dyes, i think so19:40
hemnaso it doesn't make sense to replace that at all with anything in os-brick19:40
hemnaI'm trying to understand it19:41
eharneyit is, because the get_config() called on the super() generates an initial libvirt-specific config obj19:41
eharneyat 25019:41
hemnaso what I'm trying to get to is can we get nova's code to a point where new connectors showing up in os-brick automatically are available to nova without adding a new libvirt volume driver.19:41
hemnaeharney, ah ok.  cool thanks19:41
*** gouthamr has quit IRC19:41
winston-dheading out for quick lunch, brb19:42
*** gouthamr has joined #openstack-cinder19:42
eharneylibvirt/ is the translation layer between those python objects and libvirt XML19:42
jgriffithhemna: yeah, I get where you're going and like it, but I think statements like "replace all of libvirt/ are naive and a bit unrealistic :)19:42
hemnaI'm sure they are.19:43
hemnabut that was my hope19:43
hemnaif it's impossible, that's cool19:43
hemnabut can we get mostly there ?19:43
jgriffithayway... sorry to butt in; just caught that statement and was like whhaaaaa19:43
hemnaand how to get to that point19:43
hemnajgriffith, yah no worries.19:43
jgriffithYou really want to take over libvirt xml :)19:43
hemnaI just want to stop hacking in nova :)19:44
hemnaand how can I get there.19:44
hemnaok I need food....19:44
*** hemna is now known as hemnafood19:44
thingeeme too19:45
e0neharlowja: hi. are you around?19:45
*** gouthamr has quit IRC19:46
jgriffithe0ne: hey19:47
*** gouthamr has joined #openstack-cinder19:47
e0nejgriffith: hi19:48
jgriffithe0ne: so19:49
jgriffithe0ne: where were we19:49
e0nejgriffith: we discussed current implementation of the list view19:50
e0nejgriffith: it handles --detailed param and it doesn't make our REST API easy19:50
xyang1Can anyone help +A this one? .  It has two +2's.  e0ne, eharney, jgriffith ?19:51
thingeeasselin_, jgriffith: for the contacting on correcting third party issues, lets cc the maintainers but keep it to the third-party-announce mailing list19:51
harlowjae0ne i'm around19:51
harlowjaare u around :-P19:51
thingeejust so all the convos are in the open?19:51
harlowjai saw that nice trace, looks like more things to optimize in taskflow19:51
thingeejust asked in infra if this would be fine, and it looks like there were no objections19:51
harlowja*and yes i have idears19:52
harlowjabut none of that CI stuff19:52
e0neharlowja: i'll say only my imo19:52
harlowjameh, its just things that can get optimized19:53
harlowjatis the way of life19:53
jgriffithe0ne: so my feeling was I didn't like that being part of list from teh client inparticular, and my argument was that it's all the same info as "show" so use "show"19:53
e0neharlowja: 20-30 requests to DB is tooo lot for our case19:53
thingeee0ne, harlowja: I'll say only my IMHO19:53
*** rbradbury has joined #openstack-cinder19:53
harlowjae0ne sure, so we can like fix that, in fact will reduce it to 1 call19:53
e0nethingee, harlowja: what CI are you talking about?19:53
harlowjai just need to verify thats used where it needs to be :-P19:53
harlowjae0ne prior joke :)19:53
thingeee0ne: harlowja is just being funny... I'm talking about these
harlowjaCI CI CI l ol19:54
e0nejgriffith: it bacames a holly war19:55
e0nejgriffith: personally, i prefer api that returns the same data for list and show19:55
*** thangp has quit IRC19:55
harlowjaso e0ne otherwise, useful results, good to know, the question still IMHO revolves around what is wanted, if no resumption is wanted, then meh, i could optimize it to 1 call and it won't matter :-P19:55
harlowjai'll optimize it to -1 call!19:55
e0nejgriffith: and detailed shoe could contain more data19:55
jgriffithe0ne: but of course :)19:55
jgriffithe0ne: so I guess get other peoples feedback and do what you like19:56
e0nejgriffith: and of course, your arguments are absolutely valid19:56
jgriffithe0ne: the --detailed option isn't the end of the world for me19:56
e0nejgriffith: :)19:56
jgriffithe0ne: but making the default have everything I do think is awful19:56
e0nejgriffith: if i found a correct english version of proverb, it sould be like: different strokes for different folks19:58
jgriffithe0ne: LOL... seems about right19:58
e0nejgriffith: i'll ask this question in openstack-dev and operators ML's19:59
e0nejgriffith: it started like "adding snapshots field to 'show' command" and now we're trying to change list and show api responces20:00
e0nejgriffith: we could get more opinions in the mailing lists than in the spec's review. let's try it20:01
e0neharlowja: ~30 calls to DB for a pretty simple 2 tasks it's not very good from a performance point of view20:04
*** gouthamr has quit IRC20:05
e0neharlowja: it's about 30% slower than w/o persistance20:05
harlowjae0ne sure, looks like a bunch of selects that just need to be optimized20:05
e0neharlowja: agree20:06
*** juzuluag has quit IRC20:06
harlowjastill goes back to my orginal point though20:06
e0neharlowja: did you see performance result from rally?20:06
*** Yogi11 has quit IRC20:06
harlowja ?20:06
harlowjaor is there something else, idk20:06
e0nethingee, harlowja: oh.. glad to see that we care about quality of CIs20:08
*** Yogi1 has joined #openstack-cinder20:09
harlowjae0ne state persistence (and resumption) isn't magic and/or free ;)20:10
harlowjabut def things can get optimized as we need20:10
e0neharlowja: :)20:10
harlowja*because each task state transtion becomes a mini-checkpoint (that needs to be saved somewhere)20:11
harlowja*unless u reduce your ability to resume from that state (which may be fine for some workflows)20:11
*** amoturi has joined #openstack-cinder20:12
harlowjau guys should try (which uses that capability to distribute work to workers, and resume it using those state checkpoints when u kill -9 or ctrl-c the workers)20:12
harlowjathere's a reason i called the review 'Add indestructible 99 bottles of beer example'20:13
harlowjabut again, adding this capability isn't 'free' or 'magic'20:13
*** vokt has quit IRC20:14
harlowja* (fixes the ability to ctrl-c workers, vs needing to kill -9 them)20:15
*** gouthamr has joined #openstack-cinder20:16
harlowjabasically what happens, is that the song progresses on a worker, u kill it, then the song gets picked up by another worker, resumes from last state, and continues20:16
harlowjaaka, u can't stop the song (the indestructible part, ha)20:16
nikeshmthingee smcginnis hemna DuncanT : CI status for hpmsa and lenovo -CI tests are passed and showing on these patches
*** juzuluag has joined #openstack-cinder20:20
e0neharlowja: thanks. i'll take a look on your patches tomorrow morning20:20
*** kjnelson has quit IRC20:20
e0neharlowja: maybe get profiler results with zookeeper also makes sense20:21
*** annashen has joined #openstack-cinder20:21
harlowjae0ne sure, but again, the question of what is desired from cinder is also useful, if nobody wants resumption, then well the point is mute (although optimizing taskflow would be good no matter)20:22
e0neharlowja: ok. i like this feature, but i'm not not sure that we'll accept it to cinder with ~30% performance impact20:23
harlowjae0ne well help make it better in taskflow?20:23
harlowjai take patches to ;)20:24
e0neharlowja: :)20:24
e0neharlowja: unfortunately, i've got only 24hours per day and sometime i need to sleep:(20:25
harlowjaya, same here20:25
*** vokt has joined #openstack-cinder20:25
harlowjamore important imho is the 'i like this feature' or 'i find this feature useful', optimize things as we go, those later points are IMHO the more important ones20:26
*** timcl has quit IRC20:26
*** gouthamr has quit IRC20:29
e0neharlowja: ok, i like this feature because it's useful but it's pretty slow:(20:30
*** alexpilotti has quit IRC20:31
tbarronakerr: if you would be so kind as to look at for QA+2 I would be, even more, in your debt.20:34
tbarronwrong window :-)20:35
openstackgerritKurt Martin proposed openstack/cinder: Update 3PAR user config help strings
*** rmesta has joined #openstack-cinder20:44
*** Lee1092 has quit IRC20:45
* e0ne bb in ~12 hours20:45
*** annegentle has quit IRC20:45
*** marcusvrn1 has quit IRC20:49
*** e0ne has quit IRC20:54
*** kjnelson has joined #openstack-cinder20:58
*** jdurgin1 has joined #openstack-cinder20:59
*** akerr has quit IRC20:59
*** dannywil_ has joined #openstack-cinder20:59
*** jdurgin has quit IRC21:00
*** patrickeast_ has joined #openstack-cinder21:01
*** jungleboyj has quit IRC21:01
*** dannywilson has quit IRC21:02
*** patrickeast has quit IRC21:02
*** patrickeast_ is now known as patrickeast21:02
*** juzuluag has quit IRC21:04
*** amoturi has quit IRC21:04
openstackgerritSean Chen proposed openstack/cinder: Tintri driver to manage existing backend storage objects
*** setmason has quit IRC21:04
*** simondodsley has quit IRC21:05
*** angela-s|afk is now known as angela-s21:06
*** setmason has joined #openstack-cinder21:06
-openstackstatus- NOTICE: Gerrit has been restarted to terminate a persistent looping third-party CI bot21:06
*** lpetrut has quit IRC21:09
*** annegentle has joined #openstack-cinder21:13
scottdaWhat version of cinderclient was Keystone v3 support introduced? Has it been in there a while?21:14
openstackgerritSean Chen proposed openstack/cinder: Tintri driver to manage existing backend storage objects
*** anshul has quit IRC21:17
*** Yogi1 has quit IRC21:18
scottdaOK, looks like it merged a year ago.21:20
*** kjnelson has quit IRC21:20
*** hemnafood is now known as hemna21:21
*** jungleboyj has joined #openstack-cinder21:24
*** rbradbury has quit IRC21:24
anteayathingee: does cinder have a nova liason?21:24
anteayain the nova meeting today there was mention of some api changes, about not having pass through to neutron and cinder21:25
anteayaI talked to sc68cal the neutron/nova person but thought I would let you know too21:25
anteayaI don't fully know what it means but it might be worth someone from cinder looking at today's nova meeting logs21:25
thingeeanteaya: thanks for letting me know. I'll check it out21:26
earlephilhowerasselin_, hemna: Thx for the os-brick CI help.  The HGST CI system is now reporting on os-brick patches, too.21:27
hemnaearlephilhower, woot!21:27
asselin_earlephilhower, wow that was fast! :)21:28
earlephilhower1-line change, really, w/new devstack-gate scripts21:28
earlephilhowerBut my CI system takes 26 min/run, not 20 like Scality. :(21:28
IlyaGwhat do u run CI on?21:29
anteayathingee: thank you21:30
*** ociuhandu has joined #openstack-cinder21:32
earlephilhowerPackstack on a 2x2.5GHz. 24 core w/64G of memory for now.21:33
hemnaearlephilhower, so your HGST volume driver doesn't use brick to do copy image <--> volume commands21:34
*** eharney has quit IRC21:35
hemnaand os-brick doesn't have the HGST connector yet21:38
hemnaso I'm curious how the tempest volume attaches are working in your CI ?21:38
hemnaand nova doesn't have an HGST libvirt volume driver21:38
*** annegentle has quit IRC21:38
*** ian_brown has joined #openstack-cinder21:39
openstackgerritxing-yang proposed openstack/cinder: Support SMI-S provider v8.0.3 in VMAX driver
earlephilhowerhemna:  Overrides are required for the 3 patches in the CI build  CINDER_REPO, NOVA_REPO, and I do a manual git pull/install for OS-BRICK21:40
earlephilhowerhemna: Should be in localrc dump21:40
mriedemsmcginnis: jgriffith: question in here even though it's merged,
mriedemi'm asking b/c it's proposed to stable/kilo also21:41
*** jungleboyj has quit IRC21:41
hemnaearlephilhower, ok I see the nova patch now from the logs21:42
hemnawe'll have to undo that patch after os-brick lands21:42
smcginnismriedem: Thanks. I'll take a look at the patch you reference. Wasn't aware of that.21:43
earlephilhowerhemna: So does that mean that I don't need to get into Nova L-1, then?  One less thing to juggle would be most welcome.  For a newcomer, getting a driver in is daunting...21:43
hemnaI'm not sure what to do here, as we are in a strange state.21:43
*** dims_ has joined #openstack-cinder21:44
hemnaI think your nova patch is fine for now21:44
hemnaI'm still working on the nova patch to pull in os-brick21:44
hemnabut without your current nova patch, there is no way your CI will work21:45
earlephilhowerWow, that would be cool.  Leaving the Nova patch in but not getting accepted for Nova L1, though, would be doable.  When os-brick patch lands, I can just abandon it.21:45
earlephilhowerI think21:45
*** jwcroppe has joined #openstack-cinder21:45
Swansonhemna: how's that wip coming along?21:46
*** dims has quit IRC21:46
hemnaSwanson, working on it today21:46
hemnatrying to get the unit tests working21:46
*** jdurgin1 has quit IRC21:47
harlowjae0ne guess u aren't around, but (a low hanging fruit to reduce a few persistence interactions...)21:49
harlowjaothers likely easily doable...21:49
*** jdurgin has joined #openstack-cinder21:49
*** annashen has quit IRC21:53
*** jungleboyj has joined #openstack-cinder21:53
*** crose has quit IRC21:54
*** rushil has quit IRC21:54
*** dannywil_ has quit IRC21:54
*** rushil has joined #openstack-cinder21:56
*** dannywilson has joined #openstack-cinder21:56
smcginnismriedem: Yeah, just blame it on jungleboyj. ;)21:56
*** jdandrea_ has joined #openstack-cinder21:56
*** jdandrea has quit IRC21:57
*** jdandrea_ is now known as jdandrea21:57
*** akerr has joined #openstack-cinder21:57
mriedemi don't understand it so i will21:58
mriedemi hate that every project has it's own service and wsgi modules21:58
jungleboyjsmcginnis: Huh?  What did I do now.21:58
smcginnismriedem: Would be great to have consistency. I doubt that's possible now though.21:58
smcginnisjungleboyj: Everything.21:58
mriedemsmcginnis: isn't that what apache is for?21:58
jungleboyjsmcginnis: Thanks man.21:58
smcginnisjungleboyj: Just questions on the db.dispose() thing.21:59
*** diogogmt has quit IRC21:59
*** mriedem is now known as mriedem_away21:59
jungleboyjYeah, my network dropped out and I had to go talk to Matt f2f.21:59
jungleboyjSpare me the network jokes, please.22:00
*** annegentle has joined #openstack-cinder22:00
jgriffithjungleboyj: care to update the review for historical purposes then?22:00
smcginnisjungleboyj: I'm not saying a thing.22:00
jgriffithjungleboyj: because I know that I for one had to do a good deal of reading to get a basic understanding up what you were actually trying to solve with that patch to begin with22:00
jgriffithjungleboyj: Tried to ping you here for more info22:01
jgriffithjungleboyj: but you wre "jungleboyaway" or some nonsense :)22:01
winston-dmriedem_away: service module is mostly just a wrapper for what's in oslo.22:01
winston-dmriedem_away: for deploying API-like service using apache, is there any comprehensive document for that?22:02
jungleboyjSure, I can do that.22:02
hemnasmcginnis, wifi ?22:02
thingeeasselin_: ping22:03
asselin_thingee, give me a few22:03
jgriffithflip214: you online, or just your machine logged in?22:04
jungleboyjjgriffith: Which review do you want that in?>22:05
jungleboyjThe one in master?22:05
jgriffithjungleboyj: yeah22:05
jgriffithjungleboyj: please22:05
jungleboyjOk, late for another meeting but I will do that tonight.22:06
jgriffithjungleboyj: just a respone to mriedem_away 's questions, so folks know you addressed his points22:06
jgriffithjungleboyj: as opposed to "I went and talked to him f2f"22:06
jungleboyjjgriffith: Oh good point.  :-)22:07
uberjayasselin_: good news/bad news -- re-applying the puppet config (i.e., and restarting the jenkins master node appears to have fixed my issues.22:07
uberjaymysterious issues worry me22:08
*** rushil has quit IRC22:08
patrickeastuberjay: that seems about on par for how these ci system work22:10
asselin_thingee,  hi22:12
asselin_uberjay, what's the bad news?22:12
asselin_oh..the mysterious issues....22:12
uberjayasselin_: oh, the bad news is I'm not really sure why it was failing in the first place... but i'm happy for now and will keep an eye on it22:13
uberjaypatrickeast: yeah, but I don't have to like it! :)22:13
*** jungleboyj has quit IRC22:14
thingeeasselin_: I've been looking at CI's like vmware and just trying to understand where communication is for what logs need to be given22:14
thingeeasselin_: example this only appears to mention testr_results
asselin_thingee,  that's the correct link22:15
asselin_it's under requirements22:15
thingeeasselin_: yeah I guess that's the problem is it's not very specific about what are environment details22:16
thingeewhat are "openstack logs"22:16
thingeeI would say by these requirements, ci's like vmware maybe fine? ... just trying to see if we can be more specific so people can find this information, instead of us having to come down on people22:17
asselin_"Include a public link to all test artifacts to make debugging failed tests easier "22:17
asselin_i think that high level statement is accurate22:17
asselin_if a reviewer can't tell what went wrong, or why it worked, then the results are not useful22:18
asselin_so may as well turn it off22:18
thingeeasselin_: well, I'm reviewing this and things like make sure to include localrc local.conf etc is just not clearly explained the list. This list of CI's with problems is more clear on what they should be including than the third party ci documentation itself.22:19
asselin_thingee, that's jgriffith's interpretation of the requirements22:20
asselin_thingee, imho localrc and local.conf are helpful, but the requirements really state the etc/cinder/cinder.conf and related conf files are what is needed22:20
thingeeasselin_: ok... I guess the point of me talking to you about this is seeing if it makes sense to include this information either in the general requirements, or the cinder ci requirements.22:20
jgriffiththingee: asselin_ feel free to delete anything/everything I wrote22:21
jgriffiththingee: asselin_ seemed to be that things like localrc and screen-logs were kinda important22:21
thingeejgriffith: I just want to be inline with the rest of you in my communication to these maintainers22:21
jgriffiththingee: my approach was "if it's in the gate logs, it should be in *your* logs"22:21
asselin_jgriffith, I'm looking here:
asselin_and trying to match up what that says vs. reality22:22
jgriffiththingee: asselin_ to the point of saying:  "here's the template with all the expected files and dir structure"  You can have more, but not less22:22
smcginnisjgriffith: +122:22
*** annegentle has quit IRC22:23
asselin_jgriffith, I've no objection to aligning with gate logs. that's been my personal approach since day 122:23
jgriffithasselin_: indeed, I know it has22:23
jgriffithasselin_: thingee I'm just unclear on the questions/statements then?22:24
jgriffithasselin_: thingee I mean............
jgriffith^^  That's the stuff I expect to see22:24
patrickeast+1 for matching up with whats in the gate logs22:24
*** amoturi has joined #openstack-cinder22:25
jgriffithNow if something like rabbit, apache etc isn't there I don't care as much; but screen-xxx.txt.gz and conf files are kinda critical IMHO22:25
asselin_jgriffith, yes, and that's what the requirements are trying to state22:25
smcginnisI think the problem is they are too general and open to interpretation.22:27
*** akerr has quit IRC22:27
asselin_ if cinder wants to clarify what those mean for cinder, then I don't see an issue.22:29
asselin_That page, however, is trying to apply to all current programs/projects22:29
smcginnisI'd rather not see per-project requirements. Some basic, common, explicit requirements would be ideal, IMO.22:30
asselin_smcginnis, is this good enough?
*** jwcroppe has quit IRC22:31
smcginnisasselin_: I'd like to see local.conf and screen logs called out.22:31
*** jwcroppe has joined #openstack-cinder22:31
smcginnisBut I guess thinking about it, there's no requirement that they are using devstack for CIs22:31
smcginnisSo I guess that isn't necessarily a common thing.22:32
smcginnisWell, common, but not required.22:32
asselin_smcginnis, right22:32
*** raunak has joined #openstack-cinder22:32
smcginnisI guess we just need to get to the point where everyone is running your scripts so they are all consistent. :)22:32
asselin_for the record, I use devstack-gate22:33
hemnasmcginnis, +122:33
asselin_whatever jenkins does, so does our ci22:33
*** bnemec has quit IRC22:34
*** salv-orlando has quit IRC22:34
thingeejgriffith: I'm saying there is a lack of documentation on what we expect from CIs.22:34
thingeethis mentions testr_results.html22:35
thingeeso I was having a discussion with you and asselin_ to see if maybe it makes sense for us to update and be inline with what we're going to communicate to people22:35
thingeeso we can point people to that for reference. and new comers in the future22:35
thingeeto update the documentation*22:36
asselin_thingee, +1 we should stick to this reference22:36
asselin_thingee, if it's helpful, we can include examples to that. I can submit a patch mentioned screen, localrc, etc.22:36
thingeeasselin_: that would be great.22:36
thingeeasselin_: I've already started contact some of the people on this list and realized that I don't think I even know what's correct.22:37
jgriffiththingee: asselin_ what about
thingeejgriffith: so that's the problem. I was saying your list is more descriptive than that.22:39
jgriffiththingee: honestly IMHO my list is "obvious"22:39
jgriffiththingee: most of the items on there are "dead link to results"22:39
jgriffiththingee: and the wiki says: very clearly22:40
thingeejgriffith: I agree it's obvious to "us", but I'm concerned with new comers. I understand there is expectancy for people to learn OpenStack, but there also has to be some explanation somewhere.22:40
jgriffithEnvironment details, Test configuration, OpenStack logs and Tempest logs22:40
thingeeopenstack logs ... can mean a lot of things.22:40
thingeeenvironment details. ... same thing22:40
jgriffiththingee: well then update that doc :)22:40
jgriffiththingee: or like I said before: Open a link to a Gate run and there's the shit you need to have :)22:41
*** amoturi has quit IRC22:41
thingeejgriffith: yeah that would be the plan, again, I just wanted a discussion with you and asselin_ if we felt we should make these ideas more clear in the documentation going out to people22:41
jgriffiththingee: Oh...22:41
jgriffiththingee: well, sure... clarity never hurts22:41
*** raunak has quit IRC22:42
jgriffiththingee: but like I said, I'm of the opinion that the bulk of things are just half-assed, not because they don't know, because it requires effort22:42
*** raunak has joined #openstack-cinder22:42
jgriffiththingee: I'm not here to spoon feed CI to anyone any more... at all22:42
jgriffithI fell into that trap, not doing it22:42
thingeejgriffith: totally understand.22:43
jgriffithasselin_: and others have provided all the tools to duplicate the CI setups22:43
thingeejgriffith: I know some of what you've been through22:43
thingeenot going to claim I know all of it ;)22:43
thingeebut when you get to the point of a docker file and go...22:43
jgriffiththingee: That's what people should use.  Others who are more adventerous and can pull it off can do whatever they want as long as they can "pull it off" ;)22:44
jgriffiththingee: yeah... :)22:44
asselin_jgriffith, yes that link is exactly what we're talking about. I can submit a clarifying patch22:45
thingeeI get it. I just want to make sure our messaging is consistent to maintainers. asselin_ is here saying "that's jgriffith 's interpretation of the requirements". and that worries me.22:46
thingeebecause we plan to share the effort of communication to maintainers.22:46
thingeeasselin_: thank you22:47
thingeeI'll be happy to review too22:47
jgriffiththingee: yeah, pehaps it would be good to just remove my etherpad22:48
rluciothingee: fyi, im putting toes back in the water for VMEM CI, it should be running for existing drivers now22:48
jgriffiththingee: so there's no confusion22:48
thingeejgriffith: No I think it's fine. This is a good starting point.22:48
rluciothingee: im testing the new driver's CI now, should be ready soon22:49
thingeerlucio: cutting it close :)22:49
jgriffiththingee: I think I'd disagreee.  If asselin_ is calling it out as "my interpretation" and doesn't agree then there's certainly a problem there22:49
jgriffithasselin_: is WAY more invested and involved in the CI stuff than I am22:50
thingeejgriffith: ok well if most are just the logs don't show properly at all, i think we can all agree on those.22:50
rluciothingee: srsly, i'll save the story for another day, but progress is progress22:51
asselin_thingee, jgriffith I don't think local.rc or local.conf are requirements per the web-site. I do think the /etc/cinder/cinder.conf and other service configurations are required22:51
*** amoturi has joined #openstack-cinder22:51
asselin_but it's definitely nice to include those localrc and local.conf files22:52
thingeerlucio: if I had a beer for everytime I heard something about a story :) ... well.. I'd probably cut myself from drinking for a bit.22:52
thingeerlucio: I guess what I'm saying is, you're not alone22:52
*** diogogmt has joined #openstack-cinder22:52
jgriffiththingee: asselin_ so I'm confused... "my interpretation" seems to align just fine22:53
*** drjones has joined #openstack-cinder22:53
jgriffiththingee: asselin_ logs must be viewable, logs must be published and publicly accessible22:53
jgriffithThos are the bulk of what i listed22:53
*** annashen has joined #openstack-cinder22:54
jgriffithso I'm unclear on why asselin_ stated "that's jgriffith 's interpretation"22:54
asselin_jgriffith, ok so then we agree22:54
*** _cjones_ has quit IRC22:54
rluciothingee: yep, if there's one thing i've learned its that automation is an "adventure"22:54
thingeeyay! my job here is done22:54
rluciothingee: always nice to know im not the only one :)22:54
*** _cjones_ has joined #openstack-cinder22:54
jgriffithasselin_: so then why the comment and the long drawn out pointless conversation about it?22:54
*** setmason has quit IRC22:55
*** annashen has quit IRC22:55
jgriffithThere, I put a disclaimer at the top of the etherpad.  I'm now officially done with the CI dance22:55
thingeejgriffith: well my point of start discussion was just doc improvement and us being inline with what we want to see in ci reports. Whether we were aligned or not from the start, I just wanted a discussion on what we agree on. It seems we agree on the specific files to make things more clear, so that's good!22:55
thingeejgriffith: aww come on, I thought I was going to have some help on communication22:56
asselin_jgriffith, sorry...I think this isthe line that started confusion "    doesn't include all of the needed log files, no local.rc or local.conf"22:56
asselin_jgriffith, but now we are on the same page..that was just your quick note about missing conf files I suppose22:56
jgriffithasselin_: sure22:57
jgriffithasselin_: so here's the thing....22:57
jgriffithEVERYTHING in OpenStack any more is SOOO bogged down with beuraucracy, red-tape and bike shedding that frankly it's tiring22:57
*** drjones has quit IRC22:57
*** amoturi has quit IRC22:58
thingeeI hear playing bruno mars - "just the way you are" works for some vendors' openstack CI's.22:58
hemnait seems that way, because it's been the wild west for a long time.22:58
jgriffithWhen somebody (including myself) says "hey... let's write up some notes on things that we think need fixed up" and it results in various statements like the ones here and a 30+ minute conversation it just says "ok... I have better things to do with my time"22:58
jgriffithhemna: fair enough22:59
*** amoturi has joined #openstack-cinder22:59
hemnaI think talking about this stuff is a good thing though IMO22:59
jgriffithhemna: ok22:59
thingeejgriffith: if you like, I can just start coming up rules with discussion with you and asselin_ and things will just happen :)23:00
asselin_jgriffith, for what it's worth..I'm trying to get everyone to converge into a working solution that's maintained upstream...and avoid  a lot of this....23:00
thingeewithout discussion*23:00
hemnaGeneral thingee23:00
hemnaall bow23:00
jgriffithalright... so missing the point all the way around23:00
openstackgerritDmitry Guryanov proposed openstack/cinder: Add Virtuozzo Storage Volume Driver
thingeehemna:  I promise there will be a ML post about it from a month ago I will reference ;)23:01
jgriffithdo me a favor and just ignore me at this point23:01
thingeejgriffith: have a beer23:01
thingeejgriffith: or ice cream23:01
thingeeor both23:01
SwansonI've yet to meet the beer that goes with ice cream.23:02
thingeeSwanson: stone smoked porter with vanilla ice cream23:02
hemnathingee, +123:02
patrickeastthingee: +123:03
SwansonI could see a porter going with ice cream.23:03
patrickeasta good porter or oatmeal stout does the trick23:03
thingeepatrickeast: yes sir23:03
* thingee goes back to break ci's with api v1 removal patches23:03
*** krtaylor has quit IRC23:04
*** annashen has joined #openstack-cinder23:11
*** markvoelker has quit IRC23:11
openstackgerritDmitry Guryanov proposed openstack/cinder: Add Virtuozzo Storage Volume Driver
uberjaythingee: oh, that sounds like a great idea. (there is a cafe in Emeryville that will make you a milkshake with Guinness and chocolate ice cream, which is really delicious :)23:15
*** zhenguo has joined #openstack-cinder23:15
*** amoturi has quit IRC23:17
*** ganso_ has quit IRC23:18
*** annashen has quit IRC23:21
*** annegentle has joined #openstack-cinder23:24
openstackgerritEarle F. Philhower, III proposed openstack/cinder: Implement Cinder Volume driver for HGST Solutions
*** annegentle has quit IRC23:29
*** chlong has joined #openstack-cinder23:35
thingeeoh hey my 3d model from the coreos booth is available
thingeejgriffith: ^23:35
thingeeI went respectfully23:36
*** annashen has joined #openstack-cinder23:38
openstackgerritTomoki Sekiyama proposed openstack/os-brick: optimize multipath call to identify IQN
*** hemna is now known as hemnafk23:45
jgriffiththingee: haha... fallen on your sword23:45
asselin_thingee, jgriffith anteaya better?
asselin_patrickeast, smcginnis ^^23:46
anteayaasselin_: here is my problem23:47
anteayaif we make that a requirement then someone needs to enforce it23:48
anteayaI'm not saying it isn't a good structure23:48
anteayathe problem is that not all projects across openstack will agree with and enforce this structure23:48
anteayaso someone setting up a ci for neutron asks neutron about this requirement23:49
*** salv-orlando has joined #openstack-cinder23:49
anteayaneutron says I don't know23:49
anteayaso the ci operator thinks they can ignore all the requirements23:49
anteayasee what I mean?23:49
jgriffithanteaya: bahh... I reviewed before seeing your comments here :)23:50
asselin_anteaya, it's not changing the requriments, just helping people understand what they mean23:50
anishuberjay: wha. what is this place ?23:50
jgriffithanteaya: and here inlies the problem :)23:50
uberjayanish: rudy's can't fail cafe23:50
patrickeastwell wait, this started because we are trying to enforce them and the requirements were apparently not clear enough…23:50
anishuberjay: many thanks23:51
patrickeastso is the suggestion that we don’t update those particular ones, and have cinder specific requirements?23:51
anteayait is an addition to the requirements section23:51
patrickeastanteaya: ^23:51
anteayaor suggestions23:51
jgriffithanteaya: seems to me it's just a "clarification"23:51
anteayasure and for you it is23:51
anteayasince you care23:51
jgriffithanteaya: are you sure about that :)23:51
anteayabut neutron doesn't care anymore23:51
jgriffithanteaya: I see your point23:51
anteayayou are talking about it23:51
anteayaso if we make something manditory23:52
jgriffithanteaya: darn you and your vulcan logic!23:52
anteayaand tell folks to ask the project about it23:52
anteayaand the project doesn't care23:52
anteayait isn't actually mandetory23:52
uberjayanish: np :)23:52
* anteaya polishes her pointy ears23:52
anteayaI so wish the situation were different23:53
anteayaand operators cared and became involved23:53
asselin_anteaya, the point is to help reviewers understand ci results.23:53
anteayalike cinder ops23:53
anteayaoh I understand the point23:53
anteayaand agree with it23:53
anteayathe problem I have is with having it as a requirement23:53
*** salv-orlando has quit IRC23:53
asselin_anteaya, it already is23:54
*** angela-s has quit IRC23:54
anteayahaving logs is a requiremtn23:55
anteayanot having them in a certain structure23:55
asselin_no structure was specified, just an example was given23:56
patrickeastmaybe i’m missing something, that patch just gives examples of what those requirements mean23:56
patrickeastnot new requirements23:56
patrickeastjust clarification of what the existing ones mean23:57
anteayahave you spent any time in nova?23:57
patrickeastnothing significant, no23:57
anteayaor talked to anyone in neutron or nova about their current third party ci situation23:57
anteayawell I really really appreciate what cinder is doing23:58
anteayaas I wish all projects were doing what you are doing23:58
anteayabut the other thing is I can't keep adding things23:58
anteayaas some folks shut down long ago23:58
anteayaand they don't look at what is currently there let alone new things23:58
anteayaI wish it were different23:58
anteayaand I totally support trying to communicate this information23:59

Generated by 2.14.0 by Marius Gedminas - find it at!