| *** haomaiwa_ has joined #openstack-manila | 00:01 | |
| *** yangyape_ has quit IRC | 00:02 | |
| *** chlong_ has quit IRC | 00:03 | |
| *** haomaiwa_ has quit IRC | 00:09 | |
| *** merooney has quit IRC | 00:09 | |
| *** merooney has joined #openstack-manila | 00:11 | |
| *** qeelee has joined #openstack-manila | 00:56 | |
| *** gouthamr has joined #openstack-manila | 00:58 | |
| *** jasonsb has joined #openstack-manila | 01:01 | |
| *** gouthamr_ has joined #openstack-manila | 01:03 | |
| *** gouthamr has quit IRC | 01:03 | |
| *** haomaiwa_ has joined #openstack-manila | 01:06 | |
| *** haomaiwa_ has quit IRC | 01:13 | |
| *** aorourke is now known as aorourke-afk | 01:13 | |
| *** yangyapeng has joined #openstack-manila | 01:13 | |
| *** mtanino has quit IRC | 01:21 | |
| *** cknight has joined #openstack-manila | 01:22 | |
| *** haomaiwang has joined #openstack-manila | 01:25 | |
| *** jasonsb has quit IRC | 01:25 | |
| *** haomaiwang has quit IRC | 01:33 | |
| *** jasonsb has joined #openstack-manila | 01:35 | |
| *** ganso has quit IRC | 01:39 | |
| *** yangyapeng has quit IRC | 01:39 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/manila: Updated from global requirements https://review.openstack.org/282479 | 01:40 |
|---|---|---|
| *** qeelee_ has joined #openstack-manila | 01:42 | |
| *** qeelee has quit IRC | 01:42 | |
| *** qeelee_ is now known as qeelee | 01:42 | |
| *** chlong_ has joined #openstack-manila | 01:49 | |
| *** jasonsb has quit IRC | 02:13 | |
| *** jasonsb has joined #openstack-manila | 02:14 | |
| openstackgerrit | zhaohua proposed openstack/manila: Huawei driver: change CIFS rw to full control https://review.openstack.org/284675 | 02:16 |
| *** haomaiwang has joined #openstack-manila | 02:21 | |
| *** qeelee has quit IRC | 02:29 | |
| *** qeelee has joined #openstack-manila | 02:31 | |
| *** jasonsb has quit IRC | 02:32 | |
| *** a_ta has joined #openstack-manila | 02:34 | |
| *** a_ta has quit IRC | 02:39 | |
| openstackgerrit | xing-yang proposed openstack/manila: Manage and unmanage snapshot https://review.openstack.org/249542 | 03:00 |
| *** haomaiwang has quit IRC | 03:01 | |
| *** haomaiwang has joined #openstack-manila | 03:01 | |
| *** cknight has quit IRC | 03:02 | |
| *** gouthamr_ has quit IRC | 03:26 | |
| bswartz | dustins: ping | 03:29 |
| dustins | bswartz: pong | 03:30 |
| bswartz | I've been experimenting with the LVM driver | 03:31 |
| bswartz | creating 20 threads that spam share create/delete all day long | 03:31 |
| dustins | Anything interesting come out of it? | 03:31 |
| bswartz | I didn't find any problems that way | 03:31 |
| bswartz | but I did find something suspicious | 03:31 |
| dustins | Do share! | 03:32 |
| bswartz | in the logs I linked to you way above | 03:32 |
| bswartz | http://logs.openstack.org/51/245751/18/check/gate-manila-tempest-dsvm-neutron-postgres-lvm-multibackend/4c9df1f/logs/screen-m-shr.txt.gz | 03:32 |
| bswartz | see all of the errors | 03:32 |
| bswartz | Logical volume lvm-shares/share-blah-blah contains a filesystem in use | 03:32 |
| bswartz | we saw the same issue with the ZFS driver | 03:32 |
| bswartz | it's caused by neutron | 03:32 |
| dustins | Well...that can't be good | 03:32 |
| bswartz | if we could turn off neutron I guarantee those errors go away | 03:33 |
| dustins | Chalk up another thing Neutron complicates :/ | 03:33 |
| bswartz | fortunately we have this: https://review.openstack.org/#/c/281477/ | 03:34 |
| bswartz | just merged 4 hours ago | 03:34 |
| dustins | Oh sweet | 03:34 |
| bswartz | I also believe that disabling of quota tests will stop this bug from triggering | 03:34 |
| bswartz | because it's the quota tests that cause creating of neutron network which make everything go downhill | 03:35 |
| dustins | Guess we could kick off a build to find out definitively | 03:35 |
| openstackgerrit | zhaohua proposed openstack/manila: Huawei driver code review https://review.openstack.org/284033 | 03:35 |
| bswartz | well since the failures are random it won't be definitive | 03:35 |
| bswartz | but we can start to build a case | 03:36 |
| dustins | Nothing's random :) | 03:36 |
| bswartz | the goal is to enable LVM tests with concurrency>1 | 03:36 |
| dustins | Exactly, I did a bit of investigation today, trying to find some patches that passed other tests but failed the LVM ones | 03:37 |
| bswartz | dustins: did you file an LP bug | 03:38 |
| openstackgerrit | zhaohua proposed openstack/manila: Huawei driver improve support of StandaloneNetworkPlugin https://review.openstack.org/284033 | 03:38 |
| dustins | Not yet, since I don't know what's causing it yet | 03:38 |
| dustins | I can see the symptoms, but no cause yet | 03:38 |
| dustins | It's on the docket for tomorrow morning | 03:39 |
| dustins | I'll ping the channel with the LP bug when it's filed | 03:40 |
| *** gregsfortytwo has quit IRC | 03:40 | |
| openstackgerrit | zhaohua proposed openstack/manila: Huawei driver code review https://review.openstack.org/285094 | 03:40 |
| openstackgerrit | Merged openstack/manila: Three ways to set Thin/Thick Type in Huawei driver https://review.openstack.org/276977 | 03:41 |
| bswartz | dustins: I've found another bug in the LVM driver which probably isn't related to the gate failures | 03:41 |
| dustins | Oh? | 03:42 |
| bswartz | I'll file this one and push a fix immediately | 03:42 |
| bswartz | it's looks like a simple oversight | 03:42 |
| bswartz | use of sudo instead of rootwrap | 03:42 |
| dustins | Ahh, yeah, simple oversight | 03:42 |
| dustins | I like those bugs, haha | 03:42 |
| bswartz | still it could have consequences that are hard to predict | 03:42 |
| dustins | Yeah, it could | 03:43 |
| bswartz | might cause random failures under some circumstances | 03:43 |
| dustins | I don | 03:43 |
| dustins | I don't know what rootwrap does differently than sudo | 03:43 |
| bswartz | http://logs.openstack.org/39/283939/1/check/gate-manila-tempest-dsvm-neutron-postgres-lvm-multibackend/ea95929/logs/screen-m-shr.txt.gz | 03:43 |
| bswartz | search for "sudo: no tty present and no askpass program specified" | 03:43 |
| dustins | But I'm sure there's enough of a difference to make things strange | 03:43 |
| bswartz | the sudo seems to outright fail | 03:44 |
| *** gregsfortytwo has joined #openstack-manila | 03:44 | |
| dustins | Yeah, looks like it's trying to sudo when passwordless sudo wasn't set up | 03:44 |
| dustins | or wasn't set up correctly | 03:44 |
| dustins | Trying to get password from tty or some other means (like the X password dialog) | 03:44 |
| dustins | Since that isn't there, it just garden-variety fails | 03:45 |
| dustins | The error above that, the failed initialization of the LVM driver, is that related to the sudo failure? | 03:47 |
| *** qeelee has quit IRC | 03:50 | |
| *** erlon has quit IRC | 03:56 | |
| *** haomaiwang has quit IRC | 04:01 | |
| *** haomaiwang has joined #openstack-manila | 04:01 | |
| *** cknight has joined #openstack-manila | 04:05 | |
| *** jasonsb has joined #openstack-manila | 04:08 | |
| bswartz | dustins: sorry was afk | 04:15 |
| bswartz | dustins: it's a failure of the periodic task it seems | 04:15 |
| dustins | hahaha, no worries | 04:15 |
| dustins | Hmm, that's interesting | 04:16 |
| *** rraja has joined #openstack-manila | 04:16 | |
| bswartz | dustins: there also seems to be a setup error | 04:26 |
| dustins | Something that would warrant a recheck at the least? | 04:27 |
| bswartz | honestly I'm not sure why this isn't a more serious problem | 04:28 |
| bswartz | I'm just going to fix the bug and see how many problems that solves | 04:29 |
| dustins | And how many it might expose :) | 04:29 |
| *** a_ta has joined #openstack-manila | 04:36 | |
| openstackgerrit | Merged openstack/manila: Validate qos during share creation https://review.openstack.org/280127 | 04:38 |
| *** a_ta has quit IRC | 04:40 | |
| *** qeelee has joined #openstack-manila | 04:41 | |
| *** dustins has quit IRC | 04:56 | |
| *** qeelee has quit IRC | 04:57 | |
| openstackgerrit | Ben Swartzlander proposed openstack/manila: Change sudo to run_as_root in LVM driver https://review.openstack.org/285107 | 04:58 |
| *** haomaiwang has quit IRC | 05:01 | |
| *** haomaiwang has joined #openstack-manila | 05:01 | |
| openstackgerrit | Ben Swartzlander proposed openstack/manila: Test no quota tests https://review.openstack.org/285109 | 05:03 |
| *** cknight has quit IRC | 05:31 | |
| *** merooney has quit IRC | 05:37 | |
| *** qeelee has joined #openstack-manila | 05:52 | |
| *** haomaiwang has quit IRC | 06:01 | |
| *** haomaiwang has joined #openstack-manila | 06:01 | |
| openstackgerrit | Merged openstack/manila: Check for device node availability before mkfs https://review.openstack.org/281849 | 06:09 |
| *** sgotliv has joined #openstack-manila | 06:14 | |
| *** sgotliv has quit IRC | 06:33 | |
| *** sgotliv has joined #openstack-manila | 06:33 | |
| *** nkrinner has joined #openstack-manila | 06:37 | |
| *** haomaiwang has quit IRC | 06:52 | |
| *** haomaiwang has joined #openstack-manila | 06:53 | |
| *** jasonsb has quit IRC | 07:00 | |
| *** haomaiwang has quit IRC | 07:01 | |
| *** haomaiwang has joined #openstack-manila | 07:01 | |
| *** yangyapeng has joined #openstack-manila | 07:04 | |
| *** sgotliv has quit IRC | 07:10 | |
| *** chlong_ has quit IRC | 07:26 | |
| *** lpetrut has joined #openstack-manila | 07:27 | |
| *** furlongm_ has joined #openstack-manila | 07:52 | |
| *** furlongm has quit IRC | 07:52 | |
| *** haomaiwang has quit IRC | 08:01 | |
| *** haomaiwang has joined #openstack-manila | 08:01 | |
| *** a_ta has joined #openstack-manila | 08:23 | |
| *** adrianofr_ has quit IRC | 08:25 | |
| *** adrianofr_ has joined #openstack-manila | 08:27 | |
| *** a_ta has quit IRC | 08:28 | |
| *** a_ta has joined #openstack-manila | 08:36 | |
| *** a_ta has quit IRC | 08:41 | |
| *** lpetrut has quit IRC | 08:47 | |
| *** haomaiwang has quit IRC | 09:01 | |
| *** haomaiwa_ has joined #openstack-manila | 09:01 | |
| openstackgerrit | liucheng proposed openstack/manila: Update Huawei driver doc for Mitaka https://review.openstack.org/284581 | 09:06 |
| *** Andrew__ has joined #openstack-manila | 09:18 | |
| Andrew__ | Hello folks, I am wondering how a connection is made between a CephFS Manila mount and the file system inside a Nova VM ??? | 09:19 |
| *** dgonzalez has quit IRC | 09:35 | |
| *** dgonzalez has joined #openstack-manila | 09:35 | |
| mkoderer__ | Andrew__: did you check https://review.openstack.org/#/c/270211/ ? | 09:41 |
| mkoderer__ | Andrew__: maybe this answers your question https://review.openstack.org/#/c/270211/34/doc/source/devref/cephfs_native_driver.rst | 09:41 |
| *** yangyape_ has joined #openstack-manila | 09:42 | |
| *** yangyapeng has quit IRC | 09:45 | |
| *** yangyape_ has quit IRC | 09:47 | |
| *** tpsilva has joined #openstack-manila | 09:50 | |
| mkoderer__ | vponomaryov: ping | 09:51 |
| vponomaryov | mkoderer__: pong | 09:52 |
| mkoderer__ | vponomaryov: I configured devstack + tempest without neutron and nova network | 09:52 |
| vponomaryov | Andrew__: admin responsible for making routing between ceph host and Nova VMs | 09:52 |
| mkoderer__ | vponomaryov: can you just have a look at https://etherpad.openstack.org/p/manila-tempest-without-neutron | 09:52 |
| vponomaryov | Andrew__: plugin enables all manila services | 09:53 |
| vponomaryov | mkoderer__: ^ | 09:53 |
| Andrew__ | mkoderer vponomaryov Ahhh thanks, looking in to the links | 09:53 |
| mkoderer__ | vponomaryov: I will have a closer look in the failing tests but in general I would say it it works | 09:53 |
| vponomaryov | mkoderer__: so, tempest smart enough to not call network stuff when services not installed? | 09:54 |
| mkoderer__ | vponomaryov: ah fine.. it's my old vagrant scripts that do that... never touch a running conf :) | 09:54 |
| mkoderer__ | vponomaryov: you need to add in [service_available] neutron = False | 09:54 |
| mkoderer__ | but that's automatically done in devstack | 09:55 |
| vponomaryov | mkoderer__: so, do you need help with failing tests? | 09:55 |
| vponomaryov | mkoderer__: also, yesterday was merged possibility to disable quota tests | 09:56 |
| vponomaryov | mkoderer__: that use dynamic creds | 09:56 |
| vponomaryov | mkoderer__: with dynamic networks | 09:56 |
| vponomaryov | mkoderer__: also, as I can see, you have Nova enabled | 09:58 |
| vponomaryov | mkoderer__: it is not needed for LVM driver functional testing | 09:58 |
| mkoderer__ | vponomaryov: I can also disable nova | 09:58 |
| vponomaryov | mkoderer__: it is expected as well | 09:59 |
| vponomaryov | mkoderer__:only keystone should left from OpenStack services | 09:59 |
| *** Andrew__ has quit IRC | 09:59 | |
| vponomaryov | except manila | 09:59 |
| mkoderer__ | vponomaryov: ok I think there is a disable_all_services statement in devstack | 09:59 |
| vponomaryov | mkoderer__: so, again, do you need help with failing tests? | 10:00 |
| *** haomaiwa_ has quit IRC | 10:01 | |
| mkoderer__ | vponomaryov: strange that the quota tests seem to work | 10:01 |
| *** haomaiwang has joined #openstack-manila | 10:01 | |
| vponomaryov | mkoderer__: why strange? | 10:01 |
| vponomaryov | mkoderer__: if tempest smart enough and do not call network stuff when it is disabled, then it should work | 10:02 |
| vponomaryov | mkoderer__: quota tests use only fact of separate tenant | 10:02 |
| vponomaryov | mkoderer__: not networks | 10:02 |
| mkoderer__ | vponomaryov: yep tempest is smart - I though you already know :) | 10:02 |
| vponomaryov | mkoderer__: oh, come on ))) | 10:03 |
| vponomaryov | mkoderer__: coincedence )) | 10:03 |
| openstackgerrit | Merged openstack/manila: Use assertIsNone instead of assertEqual(None, ***) https://review.openstack.org/284124 | 10:03 |
| openstackgerrit | Csaba Henk proposed openstack/manila: ganesha: implement update_access https://review.openstack.org/282602 | 10:07 |
| *** qeelee has quit IRC | 10:09 | |
| *** Andrew__ has joined #openstack-manila | 10:10 | |
| *** jcsp has quit IRC | 10:16 | |
| Andrew__ | mkoderer__ vponomaryov : The code for cephfs native client is still under review, so i am still not sure how a connection is made between a CephFS Manila mount and the file system inside a Nova VM ??? | 10:17 |
| Andrew__ | NFS ?? | 10:18 |
| *** rraja has left #openstack-manila | 10:19 | |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/python-manilaclient: [DEBUG] do not merge https://review.openstack.org/285199 | 10:23 |
| *** haomaiwang has quit IRC | 10:23 | |
| mkoderer__ | Andrew__: "Guests require a Ceph client in order to mount the filesystem." | 10:24 |
| vponomaryov | Andrew__: you were provided with link https://review.openstack.org/#/c/270211/34/doc/source/devref/cephfs_native_driver.rst and said that looking into it, answer is there | 10:25 |
| mkoderer__ | Andrew__: means, you need to have ceph clients installed on the VM and a direct connection to the ceph cluster (mon, osd and mds) | 10:25 |
| *** lpetrut has joined #openstack-manila | 10:26 | |
| mkoderer__ | Andrew__: very problematic from secuirty point of view IMHO | 10:26 |
| *** haomaiwang has joined #openstack-manila | 10:29 | |
| *** jcsp has joined #openstack-manila | 10:29 | |
| Andrew__ | vponomaryov: As i said, Not Yet Merged. My question was how are we doing it now ? | 10:34 |
| mkoderer__ | Andrew__: for now Manila does not support Ceph as a driver | 10:34 |
| mkoderer__ | Andrew__: if you want to use ceph with manila, use the Generic Driver and configure Cinder using Ceph RBD | 10:35 |
| Andrew__ | mkoderer: cool, do you happen to know which protocol is used in this communication between client and ceph-mons ? | 10:35 |
| *** a_ta has joined #openstack-manila | 10:37 | |
| mkoderer__ | Andrew__: so for the ceph native drive will be someting like that http://docs.ceph.com/docs/hammer/start/quick-cephfs/ | 10:38 |
| mkoderer__ | but I guess it should be stated in the review somewhere too | 10:39 |
| *** a_ta has quit IRC | 10:41 | |
| *** erlon has joined #openstack-manila | 10:48 | |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/python-manilaclient: [DEBUG] do not merge https://review.openstack.org/285199 | 10:49 |
| *** sgotliv has joined #openstack-manila | 10:49 | |
| *** ganso has joined #openstack-manila | 10:59 | |
| *** haomaiwang has quit IRC | 11:01 | |
| *** haomaiwang has joined #openstack-manila | 11:01 | |
| *** Andrew__ has quit IRC | 11:07 | |
| *** sgotliv has quit IRC | 11:08 | |
| *** haomaiwang has quit IRC | 11:14 | |
| *** Andrew__ has joined #openstack-manila | 11:16 | |
| *** qeelee has joined #openstack-manila | 11:20 | |
| *** Andrew__ has quit IRC | 11:21 | |
| mkoderer__ | vponomaryov: don't we have to exclude consistency group tests for LVM. It seems not to be implemented? or am I missing something? | 11:23 |
| mkoderer__ | ganso: ^ | 11:24 |
| vponomaryov | mkoderer__: yes | 11:24 |
| vponomaryov | mkoderer__: consistency groups are not supported by LVM driver and should be skipped | 11:24 |
| vponomaryov | mkoderer__: as well as "manage" | 11:25 |
| mkoderer__ | vponomaryov: ahh it seems I copied the wrong tempest.conf from the gate :) | 11:25 |
| *** Andrew__ has joined #openstack-manila | 11:28 | |
| *** bob___ has quit IRC | 11:36 | |
| *** Andrew__ has quit IRC | 11:37 | |
| openstackgerrit | Merged openstack/manila: Updated from global requirements https://review.openstack.org/282479 | 11:40 |
| *** Andrew__ has joined #openstack-manila | 11:45 | |
| *** jcsp has joined #openstack-manila | 11:48 | |
| *** Andrew__ has quit IRC | 11:53 | |
| *** qeelee has quit IRC | 11:58 | |
| *** cdent has joined #openstack-manila | 12:00 | |
| *** Andrew__ has joined #openstack-manila | 12:01 | |
| openstackgerrit | Tiago Pasqualini da Silva proposed openstack/manila: Add separate process for Data Service copy https://review.openstack.org/279164 | 12:02 |
| *** Andrew__ has quit IRC | 12:08 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: [Do not merge] Test LVM mount capability https://review.openstack.org/282507 | 12:11 |
| *** Andrew__ has joined #openstack-manila | 12:12 | |
| *** krotscheck_dcm is now known as krotscheck | 12:13 | |
| *** haomaiwang has joined #openstack-manila | 12:14 | |
| cdent | bswartz: Are you a good person to talk about manila's use of microversion headers? I'm on a mission from the api-wg to see whether the existing microversioned projects will consider approaching the headers differently. | 12:17 |
| *** haomaiwang has quit IRC | 12:19 | |
| cdent | bswartz: if you look at https://review.openstack.org/#/c/280381/1/guidelines/headers.rst you'll see the issue with the existing headers. That example has been stark enough to make people think maybe we should change the guidelines. | 12:19 |
| cdent | However if we do that existing microversioned projects will need to support both their existing header and the new one. | 12:20 |
| *** Andrew__ has quit IRC | 12:21 | |
| *** Andrew__ has joined #openstack-manila | 12:29 | |
| *** zhongjun22 has joined #openstack-manila | 12:29 | |
| *** zhongjun2 has quit IRC | 12:33 | |
| openstackgerrit | John Spray proposed openstack/manila: Add Ceph Native driver https://review.openstack.org/270211 | 12:36 |
| *** dalgaaf has joined #openstack-manila | 12:40 | |
| *** zhongjun has joined #openstack-manila | 13:04 | |
| *** zhongjun22 has quit IRC | 13:07 | |
| openstackgerrit | John Spray proposed openstack/manila: Add Ceph Native driver https://review.openstack.org/270211 | 13:22 |
| *** cknight has joined #openstack-manila | 13:23 | |
| *** cknight has quit IRC | 13:25 | |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila: Add ZFSonLinux share driver https://review.openstack.org/277192 | 13:28 |
| *** merooney has joined #openstack-manila | 13:30 | |
| *** merooney has quit IRC | 13:34 | |
| *** gouthamr has joined #openstack-manila | 13:35 | |
| *** porrua has joined #openstack-manila | 13:35 | |
| *** akerr has joined #openstack-manila | 13:37 | |
| *** haomaiwang has joined #openstack-manila | 13:38 | |
| *** Andrew__ has quit IRC | 13:42 | |
| *** merooney has joined #openstack-manila | 13:43 | |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila: Add ZFSonLinux share driver https://review.openstack.org/277192 | 13:43 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: [Do not merge] Test LVM mount capability https://review.openstack.org/282507 | 13:50 |
| *** Andrew__ has joined #openstack-manila | 13:56 | |
| gouthamr | vponomaryov: ping | 13:58 |
| *** merooney has quit IRC | 13:59 | |
| *** haomaiwang has quit IRC | 14:01 | |
| *** haomaiwang has joined #openstack-manila | 14:01 | |
| Andrew__ | jcsp: Hello | 14:03 |
| Andrew__ | jcsp: I was wondering how Ceph Native driver will handle non-linux guests ? | 14:04 |
| mkoderer__ | ganso: all manila api tests work fine with just keystone, manila and tempest (using lvm driver) | 14:04 |
| ganso | mkoderer__: great! :D | 14:05 |
| mkoderer__ | Andrew__: which OS do you have in mind? Windows? | 14:05 |
| ganso | mkoderer__: this is going to be the fastest job in CI | 14:05 |
| mkoderer__ | ganso: yeahh | 14:05 |
| akerr | any cores able to +A https://review.openstack.org/#/c/284877/? It's passed our CI and the bug is wreaking havoc with our downstream CI system :) | 14:05 |
| mkoderer__ | ganso: I documented it how I configured it here https://etherpad.openstack.org/p/manila-tempest-without-neutron | 14:05 |
| ganso | akerr: ill take al ook | 14:06 |
| Andrew__ | mkoderer: yes please :) | 14:06 |
| akerr | ganso: thanks | 14:06 |
| mkoderer__ | Andrew__: never herd about ceph clients on windows... you might ask in the ceph channel on irc.oftc.net | 14:07 |
| Andrew__ | mkoderer: Here, Sage had a plan for windows guests as well, https://www.youtube.com/watch?v=dNTCBouMaAU | 14:10 |
| mkoderer__ | Andrew__: yep just ask Sage in #ceph channel | 14:11 |
| vponomaryov | gouthamr: pong | 14:12 |
| vponomaryov | akerr: done | 14:12 |
| gouthamr | vponomaryov: hi! i don't know how the zfs-experimental job works. If i turn on replication tests in the tempest tests patch and run-experimental, would it run against the latest ZFS patch? | 14:12 |
| *** alyson_ has joined #openstack-manila | 14:12 | |
| ganso | akerr: hmm my driver also does not check such situation, but I was aware of that... the way I see it, if it tries to update_access and share does not exist, it should throw error. If this happens during deleting, it should also throw error (it is invoked by manager). Share should be unmanaged in this situation. I am not sure but I think unmanage only works | 14:13 |
| ganso | in DHSS=False. | 14:13 |
| vponomaryov | gouthamr: yes | 14:13 |
| vponomaryov | if they are enabled | 14:13 |
| vponomaryov | gouthamr: but I suspect they will be disabled | 14:13 |
| vponomaryov | gouthamr: so, maybe better to set its new config option enabled in ZFS driver | 14:14 |
| vponomaryov | gouthamr: in that case everything should be ok | 14:14 |
| ganso | vponomaryov: you +A'ed https://review.openstack.org/#/c/284877/ | 14:14 |
| vponomaryov | gouthamr: according to configuration | 14:14 |
| vponomaryov | ganso: I know | 14:14 |
| ganso | vponomaryov: don't you think the update_ccess finishing gracefully and thus setting share_instance_access_status to active when in fact share does not exist in backend is ok? | 14:15 |
| ganso | vponomaryov: s/don't you/do you | 14:15 |
| *** Yogi1 has joined #openstack-manila | 14:15 | |
| Andrew__ | mkoderer: https://github.com/ketor/ceph-dokan | 14:16 |
| vponomaryov | ganso: with deletion of each share "update_access" is called | 14:16 |
| gouthamr | vponomaryov: can you set it then? in the ZFS patch, if driver=='zfsonlinux'? the option is 'RUN_REPLICATION_TESTS' | 14:16 |
| vponomaryov | ganso: and if share creation was failed, in that case we should not raise exceptions | 14:16 |
| gouthamr | vponomaryov: that way, i can check-experimental on the test patch.. | 14:16 |
| vponomaryov | ganso: else we won't be able to delete errored shares at all | 14:16 |
| ganso | vponomaryov: this is not for scenario of share creation failed. The share was deleted in the backend | 14:17 |
| *** xyang1 has joined #openstack-manila | 14:17 | |
| vponomaryov | gouthamr: ok | 14:17 |
| ganso | vponomaryov: I think the correct fix should be try/catch in manager | 14:17 |
| ganso | vponomaryov: not in drivers | 14:17 |
| akerr | ganso: we have an internal test we run to verify that we handle cleaning up resources in bad states properly. In this case if we are attempting to delete a share and the backing resource on the array is already gone we should not throw an error (as what we wanted has already occurred) | 14:17 |
| vponomaryov | gouthamr: you need to set "depends-On" then in your commit with tests | 14:17 |
| ganso | vponomaryov: I believe this is introducing inconsistent behavior | 14:18 |
| ganso | akerr: I agree that an error should not be thrown, but I think the fix should not be in driver, it should be in manager | 14:18 |
| akerr | i see | 14:19 |
| akerr | then we would need a consistent return code (or exception type) to return to manager to indicate the backing resource is already gone that all drivers can agree on | 14:20 |
| ganso | akerr: a not found exception would handle this | 14:20 |
| gouthamr | vponomaryov: sure thing. thanks! | 14:21 |
| gouthamr | ganso: this is true of any workflow as ckight and I were discussing yesterday | 14:22 |
| * bswartz headdesk | 14:23 | |
| ganso | gouthamr: sorry I missed that discussion. What workflow exactly? | 14:23 |
| gouthamr | ganso: we don't poll for the health of shares (and we shouldn't be).. the share will be set to 'error' on the next user initiated interaction | 14:23 |
| openstackgerrit | Ben Swartzlander proposed openstack/manila: Test no quota tests https://review.openstack.org/285109 | 14:24 |
| akerr | i agree I think this could be handled in the manager, but its a little late in the release for a core change that all drivers would need to implement. I would suggest it gets changed in early newton and we can change the netapp driver to the new error handling flow then | 14:24 |
| *** xyang_ has joined #openstack-manila | 14:24 | |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila: Add ZFSonLinux share driver https://review.openstack.org/277192 | 14:25 |
| openstackgerrit | Ben Swartzlander proposed openstack/manila: Change sudo to run_as_root in LVM driver https://review.openstack.org/285107 | 14:25 |
| vponomaryov | gouthamr: now, ZFS driver with PS40 will run replication tests | 14:25 |
| *** yangyapeng has joined #openstack-manila | 14:26 | |
| *** nkrinner has quit IRC | 14:26 | |
| bswartz | vponomaryov: do we know that they will pass? | 14:26 |
| gouthamr | vponomaryov: thanks.. | 14:26 |
| vponomaryov | bswartz: depends on tests | 14:26 |
| bswartz | lol | 14:26 |
| vponomaryov | bswartz: I didn't look at them | 14:26 |
| bswartz | and you didn't try them locally first? | 14:26 |
| bswartz | that means this will be exciting | 14:26 |
| gouthamr | vponomaryov: sure... depends on driver :P | 14:26 |
| *** cknight has joined #openstack-manila | 14:27 | |
| vponomaryov | bswartz: me? replication tests? No | 14:27 |
| vponomaryov | bswartz: only manual testing | 14:28 |
| bswartz | the tempest tests should not vote until all the bugs have been fixed | 14:28 |
| *** Andrew__ has quit IRC | 14:29 | |
| *** merooney has joined #openstack-manila | 14:30 | |
| *** timcl has joined #openstack-manila | 14:32 | |
| *** bogdans has joined #openstack-manila | 14:37 | |
| xyang_ | vponomaryov, ganso, gouthamr, cknight, markstur, bswartz: hi, can you take a look of the manage snapshot patch and see if your concerns are addressed? https://review.openstack.org/#/c/249542/ | 14:37 |
| *** a_ta has joined #openstack-manila | 14:38 | |
| gouthamr | sure thing xyang_ | 14:40 |
| xyang_ | thanks | 14:42 |
| *** eharney has joined #openstack-manila | 14:42 | |
| *** a_ta has quit IRC | 14:43 | |
| *** dustins has joined #openstack-manila | 14:43 | |
| *** bogdans has left #openstack-manila | 14:43 | |
| ganso | xyang_: Hi Xing. I just saw the code. I think it could be done in a different way. Like definig a function that contains the comment and sets the status, instead of repeating the code 3 times. | 14:45 |
| ganso | xyang_: but I am ok with the way it is. I wouldn't -1 because of this | 14:46 |
| ganso | xyang_: I consider the concern addressed | 14:46 |
| xyang_ | ganso: thanks. I'll change it if there are other comments | 14:47 |
| ganso | xyang_: thanks Xing. | 14:47 |
| *** bogdans has joined #openstack-manila | 14:48 | |
| bswartz | vponomaryov: pls send link to failing manilaclient test log | 14:49 |
| bswartz | mkoderer__: you here? | 14:49 |
| vponomaryov | bswartz: http://logs.openstack.org/99/285199/2/check/gate-manilaclient-dsvm-neutron-functional/b121510/logs/devstack-gate-post_test_hook.txt.gz#_2016-02-26_11_11_53_034 | 14:49 |
| vponomaryov | bswartz: https://bugs.launchpad.net/python-manilaclient/+bug/1550166 | 14:49 |
| openstack | Launchpad bug 1550166 in python-manilaclient "Functional tests are broken with impossibility to create share-network" [Critical,New] - Assigned to Valeriy Ponomaryov (vponomaryov) | 14:49 |
| bswartz | mkoderer: ping | 14:50 |
| openstackgerrit | Ben Swartzlander proposed openstack/manila: Add doc for export location metadata https://review.openstack.org/282071 | 14:53 |
| cdent | bswartz: did you see my earlier comments about microversion headers: https://tank.peermore.com/_/06b309f5-d8ba-4b72-a7ab-ae352657aca9 | 14:55 |
| bswartz | cdent: no I haven't seen that, thanks for the pointer | 14:56 |
| bswartz | let me read the doc | 14:56 |
| gouthamr | cdent: question, for multi word service_types, what's ur suggestion for the header? | 14:57 |
| cdent | bswartz: thanks, nova and ironic are happy to head in the new direction and I think manila is the only other deployed microversioned service | 14:57 |
| cdent | gouthamr: from what I can see the precent has been hyphenate | 14:57 |
| gouthamr | cdent: "OpenStack-API-Version: Shared File Systems 2.11" or "OpenStack-API-Version: Shared-File-Systems 2.11" or "OpenStack-API-Version: Shared File Systems 2.11 | 14:57 |
| cdent | sorry precedent | 14:57 |
| *** a_ta has joined #openstack-manila | 14:57 | |
| gouthamr | cdent: ugh, sorry last one was meant to be with underscores | 14:58 |
| *** cknight has quit IRC | 14:58 | |
| gouthamr | cdent: ah thanks. | 14:58 |
| bswartz | cdent: I agree the new form is superior, but it cannot be used without breaking backward compatibility | 14:58 |
| cdent | gouthamr I suspect there will be some debate about multiword stuff, yet to be resolved | 14:58 |
| bswartz | because backward compatibility is the whole point of microversions, I don't think a change makes sense | 14:58 |
| cdent | bswartz: exactly; the existing microversioned services will need to commit to supporting both header names forever | 14:59 |
| gouthamr | cdent: we were concerned about the microversion response header | 14:59 |
| bswartz | cdent: that seems like a needless waste of bandwidth -- why not ask new serivces to use the new format, and grandfather existing services with their existing formats? | 14:59 |
| cdent | gouthamr: probably but we're still getting a feel for the options | 14:59 |
| bswartz | that will avoid the exploision of headers over time, as there will be a fixed number of non-conforming implementaitons | 15:00 |
| cdent | bswartz: because they you end up requiring _new_ clients to support both | 15:00 |
| cdent | which we don't want | 15:00 |
| cdent | the assumption driving this is that there are and should be a multitude of clients | 15:00 |
| bswartz | new clients would only need to support the small number of headers, not an ever-growing number | 15:00 |
| cdent | bswartz: yes, but why make them support a special header for a specific service | 15:00 |
| *** haomaiwang has quit IRC | 15:01 | |
| bswartz | well if the client is the manila client, for example, it only supports manila and doesn't need to worry about other ones | 15:01 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: [Do not merge] Test LVM mount capability https://review.openstack.org/282507 | 15:01 |
| *** haomaiwang has joined #openstack-manila | 15:01 | |
| bswartz | cross project clients would have to special case a few services | 15:01 |
| *** Andrew__ has joined #openstack-manila | 15:01 | |
| bswartz | extra lines of code are better than wasted bytes on the wire forever | 15:02 |
| cdent | bswartz: so I'm not seeing where we would have additional headers. client would send one type of header and receive one type of header (the server would know what to send based on what it received in the first place) | 15:02 |
| bswartz | I would feel differently if there was a deprecation path that could lead to not needing both at some point | 15:03 |
| mkoderer__ | bswartz: yep | 15:03 |
| mkoderer__ | bswartz: right before my weekend but still available :) | 15:03 |
| cdent | I personally feel that a deprecation path would be fine, but there's not total agreement on that | 15:03 |
| bswartz | cdent: I don't see how we can ever stop supporting the existing header | 15:04 |
| bswartz | clients based on liberty need it | 15:04 |
| openstackgerrit | Merged openstack/manila: Fix NetApp cDOT driver update_access negative test https://review.openstack.org/284877 | 15:04 |
| cdent | bswartz; let's table that aspect for a minute, becuse I think it is distracting from the main point. You've asserted that we'd increase bytes on the wire and I've countered that with "no". do you not aggree? | 15:05 |
| bswartz | cdent: (09:59:12 AM) cdent: bswartz: exactly; the existing microversioned services will need to commit to supporting both header names forever | 15:07 |
| bswartz | I'm imagining you want the server to return both headers starting from mitaka | 15:08 |
| cdent | "bswartz: so I'm not seeing where we would have additional headers. client would send one type of header and receive one type of header (the server would know what to send based on what it received in the first place)" | 15:08 |
| cdent | you return the header style that you were sent | 15:08 |
| *** merooney has quit IRC | 15:08 | |
| bswartz | that's proposal is less gross, but I'm not convinced that it's safe | 15:08 |
| bswartz | I need to think through the permutations of possibilities | 15:09 |
| bswartz | mkoderer__: did you find a way to make tempest work without neutron? | 15:10 |
| cdent | bswartz: that's completely fair, let me know if/when you have additional thoughts on this. api-wg has several microversion guidelines in flight that will go one way or another depending on how we resolve this issue | 15:11 |
| mkoderer__ | bswartz: yes | 15:11 |
| mkoderer__ | bswartz: you should read your mails :) | 15:11 |
| bswartz | mkoderer__: I worried that the LVM driver has bad interaction with neutron as well, so I'm very interested in what you've found | 15:12 |
| bswartz | cdent: who should I talk to and/or where can I find them after I've considered it? | 15:12 |
| mkoderer__ | bswartz: so all api test passing with LVM driver and only Keystone, Manila and Tempest activated | 15:12 |
| bswartz | mkoderer__: I never read emails! | 15:12 |
| bswartz | lol | 15:12 |
| * bswartz opens thunderbird | 15:13 | |
| cdent | me, sdague or elmiko in #openstack-sdks or on that linked header proliferation guideline | 15:13 |
| mkoderer__ | bswartz: haha.. ok I documented what it did here: https://etherpad.openstack.org/p/manila-tempest-without-neutron | 15:13 |
| bswartz | mkoderer__: this is super interesting | 15:14 |
| bswartz | so how were the quota tests able to run without neutron? is tempest smart enough to skip over neutron related parts when creating dynamic projects? | 15:15 |
| mkoderer__ | bswartz: tempest has a list of available services.. if you disable one it doesn't try to create resources for it | 15:16 |
| mkoderer__ | bswartz: everything else I would consider as bug | 15:16 |
| bswartz | so are the quota tests able to run successfully? they just do the keystone work to create tenents but do not create assoicated networks? | 15:17 |
| mkoderer__ | bswartz: yes there are successful | 15:17 |
| bswartz | mkoderer__: is it possible to make tempest act as if neutron isn't present even when it is present? | 15:17 |
| *** JoseMello has joined #openstack-manila | 15:17 | |
| mkoderer__ | bswartz: yep. just switch "neutron = False" in [service_available] | 15:17 |
| *** Yogi1 has quit IRC | 15:18 | |
| bswartz | for example, it's harder to make devstack-gate not install neutron at all, but we can easily turn off neutron support in tempest with a manila patch | 15:18 |
| bswartz | do you have a patch upstream that does this? | 15:18 |
| mkoderer__ | bswartz: no but I can do that | 15:18 |
| bswartz | if you have time, I'm very interested | 15:18 |
| mkoderer__ | bswartz: I will do it on monday.. if this works for you | 15:19 |
| bswartz | mkoderer__: this was my attemped to avoid neutron issues: https://review.openstack.org/#/c/285109/ | 15:19 |
| *** cknight has joined #openstack-manila | 15:19 | |
| bswartz | mkoderer__: I will do it if you don't have time today -- assuming that our LVM issues really are neutron releated I'm extremely interested in the ability to turn off neutron where it's not needed | 15:20 |
| mkoderer__ | bswartz: iniset $TEMPEST_CONFIG service_available neutron False should do it | 15:20 |
| bswartz | mkoderer__: thx | 15:20 |
| bswartz | I'll put you on the patch when I push it up | 15:21 |
| mkoderer__ | bswartz: ok thx | 15:21 |
| bswartz | so you can see if it works as planned | 15:21 |
| *** merooney has joined #openstack-manila | 15:21 | |
| dustins | I'm intrigued by the results of this as well | 15:22 |
| bswartz | dustins: good morning! | 15:22 |
| dustins | bswartz: Right back at you! | 15:22 |
| bswartz | dustins: did you see the etherpad | 15:23 |
| dustins | What etherpad? | 15:23 |
| dustins | guess not, haha | 15:23 |
| bswartz | https://etherpad.openstack.org/p/manila-tempest-without-neutron | 15:23 |
| * bswartz is hopeful about reducing gate times by slimming down devstack | 15:24 | |
| *** aorourke-afk is now known as aorourke | 15:24 | |
| dustins | mkoderer__++ | 15:24 |
| dustins | This is pretty sweet | 15:24 |
| dustins | Looks like I could try that myself to see if I can replicate that LVM strangeness without Neutron getting in the way | 15:27 |
| akerr | bswartz: did your patches merge to allow no cinder/nova? | 15:29 |
| *** Yogi1 has joined #openstack-manila | 15:33 | |
| gouthamr | ganso xyang_ toabctl vponomaryov cknight bswartz: reaching out to get +Ws on these two patches that have two +2s each, they'd simplify dependency chains a lot more if they merged right away.. https://review.openstack.org/#/c/277545/ https://review.openstack.org/#/c/282081/ | 15:34 |
| toabctl | gouthamr: done for both | 15:35 |
| gouthamr | toabctl: thanks! :) | 15:35 |
| *** bogdans has quit IRC | 15:38 | |
| *** bogdans has joined #openstack-manila | 15:38 | |
| *** dencaval has joined #openstack-manila | 15:38 | |
| cknight | toabctl: thanks | 15:39 |
| bswartz | cdent: still here? | 15:39 |
| cdent | bswartz: yessir | 15:39 |
| bswartz | cdent: there is one permutation that I think causes problems | 15:39 |
| cknight | gouthamr: we often like to have more than 2 cores +2 major core work | 15:39 |
| cdent | do tell? | 15:39 |
| bswartz | if a new client (say mitaka) talks to an old server (say liberty), then which header will it send? | 15:40 |
| bswartz | if it sends the new header, the server will assume the client doesn't know what its' talking about, and use 2.0 | 15:40 |
| bswartz | if the client uses the old header, then it will work, but you won't have solve the problem you're after solving | 15:41 |
| * cdent thinks | 15:41 | |
| bswartz | if the client sends both headers, it will also work, but you'll be back to wasting bytes | 15:41 |
| cdent | I can think of a few ways around that, none of them entirely satisfactory. If a mitaka client wants to use a liberty server, it is already required to state a lower microversion, internal to itself, so that it sends the right requests. If it has the brains to do that, it has the brains to know that for a server microversion<=X.Y send header "foo". | 15:44 |
| cdent | That's for the clients which care about using up the scarce bytes. for those that don't they can just send both headers. | 15:44 |
| * akerr wonders who has scarce bytes in a cloud | 15:45 | |
| bswartz | well my earlier argument is that you can avoid ever wasting bytes by using a lookup table in the client for servers that need non-standard microversion header keys | 15:45 |
| *** bogdans has quit IRC | 15:46 | |
| * bswartz wonders if akerr secrets loves SOAP+XML | 15:46 | |
| bswartz | s/secrets/secretly/ | 15:46 |
| dustins | akerr <3 SOAP+XML | 15:46 |
| cdent | neener neener | 15:47 |
| * akerr secretly loves jsonx | 15:47 | |
| dustins | akerr: I can dig that | 15:47 |
| cdent | bswartz: so, where does this leave the issue. Is it one? | 15:47 |
| akerr | I'm just saying though that if you can't afford an extra header on your network you're not doing cloud right | 15:47 |
| bswartz | cdent: I think that services that have already shipped microversion support cannot change their headers without either breaking somebody or creating a situation where either the clients or the servers always have to send duplicate headers | 15:48 |
| bswartz | because duplicate headers are super gross, I'm in favor of not changing headers for existing services, and only establishing a standard that applies to new services adopting microversions | 15:49 |
| bswartz | unfortunately clients which talk to multiple services (like osc) will need lookup tables or special case code to deal | 15:49 |
| bswartz | the size of the lookup tables or special case could will be fixed though, as it only needs to cover existing services | 15:50 |
| bswartz | s/could/code/ | 15:50 |
| * bswartz curses autocorrect | 15:50 | |
| bswartz | I wish we could go back in time and do microversions differently | 15:53 |
| bswartz | but we did what we did and now we have to live with it | 15:53 |
| cdent | I don't get it: why doesn't this work: "If a mitaka client wants to use a liberty server, it is already required to state a lower microversion, internal to itself, so that it sends the right requests. If it has the brains to do that, it has the brains to know that for a server microversion<=X.Y send header "foo"" ? | 15:53 |
| *** mtanino has joined #openstack-manila | 15:54 | |
| bswartz | cdent: the problem is that client doesn't know what version the server is until it receives a response to its request -- and by that time it's too late | 15:54 |
| * cdent reviews the microversion flow | 15:54 | |
| bswartz | suppose I send a delete snapshot request from my mitaka client -- I'm going to tag it with the latest microversion I know about (in the my client) which is 2.11 | 15:56 |
| *** yangyapeng has quit IRC | 15:56 | |
| bswartz | if I use the new header, the server will treat the request as it if was 2.0 because it won't see the header at all | 15:56 |
| *** yangyapeng has joined #openstack-manila | 15:56 | |
| *** yangyapeng has quit IRC | 15:56 | |
| bswartz | if I use the old header, the server will correctly throw an error, informing the client that it only speaks 2.6 | 15:56 |
| bswartz | then the client can retry with 2.6, if the client speaks 2.6 | 15:57 |
| *** Yogi1 has quit IRC | 15:58 | |
| cdent | Hmmm, good points. There are some possible solutions: | 15:58 |
| bswartz | the problem occurs if the client passed in some parameter which only makes sense in microversion 2.11, and expected the server to respect it | 15:58 |
| *** bogdans has joined #openstack-manila | 15:58 | |
| cdent | a) clients should do version discovery: https://review.openstack.org/#/c/243429/6/guidelines/microversion_specification.rst | 15:58 |
| cdent | b) just send both headers becuase bytes aren't scarce! | 15:58 |
| bswartz | version discovery requires statefulness and increases latency | 15:59 |
| bswartz | both things are not acceptable to me | 15:59 |
| cdent | hmmm | 16:00 |
| bswartz | the "bytes aren't scarce" argument is a fair one, but one I don't agree with | 16:00 |
| *** martyturner has joined #openstack-manila | 16:00 | |
| bswartz | given the choice between writing more code and wasting some bytes, I'd always choose to write more code | 16:00 |
| *** haomaiwang has quit IRC | 16:01 | |
| cdent | I'd say if you're worried about latency in openstack, http is probably not the place with the highest return on effort spent improving... | 16:01 |
| cdent | But I sympathize | 16:01 |
| bswartz | lazy programmers are already clogging the intertubes too much | 16:01 |
| * cdent doesn't like microversions at all | 16:01 | |
| *** haomaiwang has joined #openstack-manila | 16:01 | |
| cdent | bswartz: I'll see what the rest of the group thinks about how we should go from here. I'll point them to the log of this. You've presented some excellent points. | 16:02 |
| bswartz | I'm not thrilled with microversions, but we needed some kind of versioning mechanism that was better than the old scheme | 16:03 |
| bswartz | and microversions was the first thing that came along so we grabbed it and ran | 16:03 |
| cdent | fair | 16:03 |
| *** Andrew__ has quit IRC | 16:05 | |
| *** Yogi1 has joined #openstack-manila | 16:09 | |
| *** logan- has quit IRC | 16:14 | |
| *** logan- has joined #openstack-manila | 16:14 | |
| *** sgotliv has joined #openstack-manila | 16:17 | |
| *** ayma has joined #openstack-manila | 16:19 | |
| openstackgerrit | Merged openstack/manila: Scheduler enhancements for Share Replication https://review.openstack.org/277545 | 16:20 |
| *** Yogi1 has quit IRC | 16:24 | |
| *** merooney has quit IRC | 16:30 | |
| *** timcl has quit IRC | 16:35 | |
| *** merooney has joined #openstack-manila | 16:39 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: [Do not merge] Test LVM mount capability https://review.openstack.org/282507 | 16:40 |
| *** jasonsb has joined #openstack-manila | 16:44 | |
| openstackgerrit | Scott DAngelo proposed openstack/manila: Allow api_version_request.matches to accept a string or None https://review.openstack.org/285426 | 16:51 |
| openstackgerrit | Ben Swartzlander proposed openstack/manila: Test LVM driver without neutron https://review.openstack.org/285429 | 16:53 |
| *** Yogi1 has joined #openstack-manila | 16:56 | |
| *** merooney has quit IRC | 17:00 | |
| *** haomaiwang has quit IRC | 17:01 | |
| *** haomaiwang has joined #openstack-manila | 17:01 | |
| *** a_ta has quit IRC | 17:10 | |
| *** bogdans has quit IRC | 17:10 | |
| *** a_ta has joined #openstack-manila | 17:10 | |
| *** jasonsb_ has joined #openstack-manila | 17:12 | |
| *** rhagarty__ has joined #openstack-manila | 17:12 | |
| *** bogdans has joined #openstack-manila | 17:14 | |
| *** a_ta has quit IRC | 17:15 | |
| *** gouthamr_ has joined #openstack-manila | 17:19 | |
| *** jasonsb has quit IRC | 17:20 | |
| *** gouthamr has quit IRC | 17:20 | |
| *** sbjorn has quit IRC | 17:20 | |
| *** rhagarty_ has quit IRC | 17:20 | |
| *** ekarlso- has quit IRC | 17:20 | |
| *** krotscheck has quit IRC | 17:20 | |
| *** mkoderer__ has quit IRC | 17:20 | |
| *** bswartz has quit IRC | 17:20 | |
| *** krotscheck has joined #openstack-manila | 17:20 | |
| *** absubram has joined #openstack-manila | 17:20 | |
| *** absubram_ has joined #openstack-manila | 17:21 | |
| *** mkoderer___ has joined #openstack-manila | 17:21 | |
| *** Andrew__ has joined #openstack-manila | 17:24 | |
| *** absubram has quit IRC | 17:25 | |
| *** absubram_ is now known as absubram | 17:25 | |
| *** a_ta has joined #openstack-manila | 17:26 | |
| *** ekarlso- has joined #openstack-manila | 17:27 | |
| *** timcl has joined #openstack-manila | 17:28 | |
| *** sbjorn has joined #openstack-manila | 17:28 | |
| *** Andrew__ has quit IRC | 17:32 | |
| *** lpetrut has quit IRC | 17:32 | |
| openstackgerrit | Clinton Knight proposed openstack/manila: Support export location metadata in NetApp cDOT drivers https://review.openstack.org/280807 | 17:33 |
| *** Andrew__ has joined #openstack-manila | 17:34 | |
| *** lpetrut has joined #openstack-manila | 17:37 | |
| *** porrua has quit IRC | 17:40 | |
| *** a_ta has quit IRC | 17:41 | |
| *** a_ta has joined #openstack-manila | 17:42 | |
| *** bogdans has quit IRC | 17:42 | |
| *** a_ta has quit IRC | 17:46 | |
| *** bogdans has joined #openstack-manila | 17:48 | |
| *** cdent has left #openstack-manila | 17:49 | |
| *** merooney has joined #openstack-manila | 17:50 | |
| *** xyang_ has quit IRC | 17:52 | |
| *** xyang_ has joined #openstack-manila | 17:52 | |
| *** bogdans has quit IRC | 17:53 | |
| *** vbellur has quit IRC | 17:56 | |
| *** lpabon has joined #openstack-manila | 17:59 | |
| openstackgerrit | Merged openstack/manila: Huawei driver: change CIFS rw to full control https://review.openstack.org/284675 | 18:00 |
| *** Andrew__ has quit IRC | 18:00 | |
| *** jasonsb_ has quit IRC | 18:00 | |
| *** haomaiwang has quit IRC | 18:01 | |
| *** lpabon has quit IRC | 18:01 | |
| *** haomaiwang has joined #openstack-manila | 18:01 | |
| *** sgotliv has quit IRC | 18:02 | |
| *** porrua has joined #openstack-manila | 18:06 | |
| *** Andrew__ has joined #openstack-manila | 18:09 | |
| *** a_ta has joined #openstack-manila | 18:10 | |
| *** lpetrut has quit IRC | 18:10 | |
| *** bogdans has joined #openstack-manila | 18:18 | |
| *** bogdans has quit IRC | 18:22 | |
| *** a_ta has quit IRC | 18:25 | |
| *** a_ta has joined #openstack-manila | 18:26 | |
| *** Yogi1 has quit IRC | 18:28 | |
| *** timcl has quit IRC | 18:29 | |
| *** a_ta has quit IRC | 18:30 | |
| *** timcl has joined #openstack-manila | 18:37 | |
| openstackgerrit | Yogesh proposed openstack/manila: Add tempest tests for Share Replication https://review.openstack.org/244843 | 18:39 |
| *** fthiagogv has quit IRC | 18:46 | |
| *** haomaiwang has quit IRC | 18:46 | |
| *** haomaiwang has joined #openstack-manila | 18:48 | |
| *** Andrew__ has quit IRC | 18:50 | |
| *** absubram has quit IRC | 18:50 | |
| openstackgerrit | Ben Swartzlander proposed openstack/manila: Add doc for export location metadata https://review.openstack.org/282071 | 18:52 |
| *** lpetrut has joined #openstack-manila | 18:54 | |
| *** bswartz has joined #openstack-manila | 18:55 | |
| *** haomaiwang has quit IRC | 19:01 | |
| *** 92AAAEL3R has joined #openstack-manila | 19:01 | |
| *** cknight has quit IRC | 19:27 | |
| *** cknight has joined #openstack-manila | 19:30 | |
| *** martyturner has quit IRC | 19:41 | |
| *** a_ta has joined #openstack-manila | 19:45 | |
| *** xyang_ has quit IRC | 19:47 | |
| *** Andrew__ has joined #openstack-manila | 19:48 | |
| *** xyang_ has joined #openstack-manila | 19:50 | |
| *** absubram has joined #openstack-manila | 19:53 | |
| *** sgotliv has joined #openstack-manila | 19:59 | |
| *** 92AAAEL3R has quit IRC | 20:01 | |
| *** haomaiwang has joined #openstack-manila | 20:01 | |
| *** eharney has quit IRC | 20:02 | |
| openstackgerrit | Clinton Knight proposed openstack/manila: Admin networks in NetApp cDOT multi-SVM driver https://review.openstack.org/281950 | 20:08 |
| *** mdenny has quit IRC | 20:10 | |
| *** eharney has joined #openstack-manila | 20:17 | |
| *** bogdans has joined #openstack-manila | 20:18 | |
| *** martyturner has joined #openstack-manila | 20:19 | |
| *** alejandrito has joined #openstack-manila | 20:20 | |
| *** bogdans has quit IRC | 20:23 | |
| *** eharney_ has joined #openstack-manila | 20:32 | |
| *** eharney has quit IRC | 20:33 | |
| *** eharney_ is now known as eharney | 20:33 | |
| *** timcl has left #openstack-manila | 20:33 | |
| *** sgotliv has quit IRC | 20:37 | |
| *** Andrew__ has quit IRC | 20:38 | |
| *** a_ta has quit IRC | 20:39 | |
| *** a_ta has joined #openstack-manila | 20:39 | |
| *** a_ta has quit IRC | 20:44 | |
| *** a_ta has joined #openstack-manila | 20:46 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: [Do not merge] Test LVM mount capability https://review.openstack.org/282507 | 20:49 |
| *** cknight has quit IRC | 20:49 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: Implement update_access() in generic driver + LVM https://review.openstack.org/256281 | 20:49 |
| *** cknight has joined #openstack-manila | 20:53 | |
| *** absubram has quit IRC | 20:54 | |
| *** haomaiwang has quit IRC | 21:01 | |
| *** haomaiwang has joined #openstack-manila | 21:01 | |
| *** akerr has quit IRC | 21:04 | |
| *** xyang_ has quit IRC | 21:05 | |
| *** merooney has quit IRC | 21:05 | |
| *** absubram has joined #openstack-manila | 21:07 | |
| *** pgbridge has quit IRC | 21:08 | |
| openstackgerrit | Merged openstack/manila: Add Replication admin APIs and driver i/f changes https://review.openstack.org/282081 | 21:10 |
| cknight | ganso: ping | 21:13 |
| ganso | cknight: pong | 21:14 |
| cknight | ganso: Can I get you to replace your +2 on a rebased patch, please? It's pure driver code, and CI has voted. https://review.openstack.org/#/c/280807/5 | 21:14 |
| cknight | ganso: Trying to get set up for a fun weekend of rechecks. | 21:15 |
| *** alejandrito has quit IRC | 21:16 | |
| ganso | cknight: hold on a min please | 21:16 |
| cknight | ganso: Actually, that patch is the predecessor for the one you looked at. Unfortunately I missed toabctl before he left. This was the one you approved: https://review.openstack.org/#/c/281950/3 | 21:18 |
| *** JoseMello has quit IRC | 21:18 | |
| ganso | cknight: new gerrit is terrible on cell phone | 21:18 |
| cknight | ganso: It's not great any other way. | 21:19 |
| ganso | cknight: Gonna set up tethering so I can take a look at it | 21:19 |
| ganso | cknight: I am a bus right now | 21:20 |
| cknight | ganso: Thanks. If you could look at the other shorter one for toabctl, that would be terrific. I'm not leaving the office today until I get the migration code going, starting in 10 minutes. | 21:21 |
| ganso | cknight: which one is that? | 21:21 |
| cknight | ganso: I need https://review.openstack.org/#/c/280807/5 and https://review.openstack.org/#/c/281950/3 | 21:22 |
| openstackgerrit | Goutham Pacha Ravi proposed openstack/manila: Add tempest tests for Share Replication https://review.openstack.org/244843 | 21:23 |
| *** alkhodos_ has joined #openstack-manila | 21:25 | |
| *** gouthamr_ has quit IRC | 21:26 | |
| alkhodos_ | Hi guys! This might not be the right channel, but maybe someone faced the same issue: I'm trying to install Liberty devstack with manila, using same local.conf as I do for master, but after the installation my c-api service is down. Here is the output I get on any CLI request: http://paste.openstack.org/show/488429/ | 21:27 |
| ganso | cknight: you probably tested this, but isn't your sort function inverted? | 21:28 |
| *** absubram has quit IRC | 21:28 | |
| bswartz | alkhodos_: it's hard to imagine that anything manila related could cause c-api to be down -- the only thing I can think of is a possible TCP port conflict | 21:29 |
| alkhodos_ | bswartz: oh sry, m-api of course | 21:29 |
| bswartz | the default TCP ports are different, but if you override them you could create a conflict that would prevent c-api from starting | 21:29 |
| bswartz | oh | 21:29 |
| alkhodos_ | too used to cinder | 21:29 |
| bswartz | alkhodos_: yeah it's something related to keystone | 21:30 |
| alkhodos_ | bswartz: what bugs me is that other services, like cinder work fine | 21:30 |
| bswartz | there was a keystone change right before liberty that we had to fix | 21:30 |
| bswartz | and more fixes since then | 21:30 |
| bswartz | it's possible something is broken again | 21:30 |
| bswartz | or maybe it's related to your local conf | 21:31 |
| bswartz | sorry I'm not involved enough with the keystone issues to tell you more | 21:31 |
| alkhodos_ | bswartz: yeh, I thought of keystone or conf too, but still couldn't figure out the issue | 21:31 |
| alkhodos_ | bswartz: thanks anyway :) | 21:32 |
| ganso | cknight: I gotta walk, will finish reviewing as soon as I get home | 21:33 |
| cknight | ganso: Nope, I think it's right. I have a unit test that proves that. | 21:36 |
| cknight | ganso: But sorting is extra credit now anyway, since we have the preferred flag. | 21:36 |
| *** adrianofr_ has quit IRC | 21:40 | |
| *** scottda is now known as scottda_is_outta | 21:56 | |
| *** dustins has quit IRC | 21:59 | |
| *** haomaiwang has quit IRC | 22:01 | |
| *** eharney has quit IRC | 22:01 | |
| *** haomaiwang has joined #openstack-manila | 22:01 | |
| *** martyturner has quit IRC | 22:05 | |
| *** merooney has joined #openstack-manila | 22:06 | |
| *** tpsilva has quit IRC | 22:07 | |
| ganso | cknight: I am home, going to take a look now | 22:09 |
| cknight | ganso: super, thanks! | 22:09 |
| cknight | ganso: it seems I have to restack to pick up your new data service | 22:09 |
| openstackgerrit | xing-yang proposed openstack/manila: Manage and unmanage snapshot https://review.openstack.org/249542 | 22:10 |
| ganso | cknight: you should not | 22:10 |
| ganso | cknight: should not need to | 22:10 |
| ganso | cknight: what error are you getting? | 22:10 |
| cknight | ganso: no error, but I rebooted and used rejoin-stack | 22:10 |
| ganso | cknight: hummm... I have never tried rebooting and using rejoin-stack | 22:11 |
| ganso | cknight: where did patchset 2 go? o_O https://review.openstack.org/#/c/281950 | 22:18 |
| cknight | ganso: no idea. i did have an issue trying to push it up today because of the dependency. | 22:19 |
| cknight | ganso: should just be a simple rebase. | 22:19 |
| *** merooney has quit IRC | 22:23 | |
| ganso | cknight: in https://review.openstack.org/#/c/281950 do you rather have another core looking or prefer a +A? | 22:25 |
| cknight | ganso: I'd rather get it workflowed, please. No telling how many rechecks will be needed. | 22:26 |
| ganso | cknight: ok, but first, let's get a verified from Jenkins | 22:26 |
| cknight | ganso: Sure, that makes sense. I'm glad to workflow it myself once that happens. Having it merged will simplify testing migration. | 22:27 |
| ganso | cknight: I can workflow it, np | 22:27 |
| ganso | cknight: I will keep an eye on it | 22:28 |
| cknight | ganso: thanks. I'll be looking for rechecks over the weekend, too. | 22:28 |
| *** porrua has quit IRC | 22:30 | |
| *** merooney has joined #openstack-manila | 22:30 | |
| ganso | cknight: let me know if you need any help testing migration patch, I'll be around here | 22:32 |
| cknight | ganso: great, thanks. | 22:32 |
| cknight | ganso: is there a wiki or anything that describes how to configure the new service? any config options I need to set? | 22:32 |
| ganso | cknight: as of right now there is no wiki. You only need to set data_node_access_ip in [DEFAULT] in manila.conf | 22:33 |
| cknight | ganso: that's easy | 22:34 |
| *** a_ta has quit IRC | 22:34 | |
| *** a_ta has joined #openstack-manila | 22:35 | |
| ganso | cknight: if you want to test on generic driver, it is a little bit more complicated | 22:36 |
| cknight | ganso: nope, I'll go with what I know to begin with | 22:36 |
| *** a_ta has quit IRC | 22:37 | |
| *** a_ta has joined #openstack-manila | 22:37 | |
| *** merooney has quit IRC | 22:46 | |
| *** marcusvrn_ has quit IRC | 22:47 | |
| *** zhonghua has joined #openstack-manila | 22:48 | |
| *** zhonghua-lee has quit IRC | 22:51 | |
| *** alyson_ has quit IRC | 22:52 | |
| *** a_ta has quit IRC | 22:52 | |
| *** merooney has joined #openstack-manila | 22:57 | |
| *** haomaiwang has quit IRC | 23:01 | |
| *** haomaiwa_ has joined #openstack-manila | 23:01 | |
| openstackgerrit | Merged openstack/manila: Support export location metadata in NetApp cDOT drivers https://review.openstack.org/280807 | 23:08 |
| *** xyang1 has quit IRC | 23:11 | |
| *** merooney has quit IRC | 23:13 | |
| *** lpetrut has quit IRC | 23:31 | |
| *** yangyapeng has joined #openstack-manila | 23:35 | |
| *** erlon has quit IRC | 23:36 | |
| *** ayma has quit IRC | 23:39 | |
| *** aorourke is now known as aorourke-afk | 23:55 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!