Wednesday, 2018-03-14

*** kei-ichi has quit IRC00:02
*** kei-ichi has joined #openstack-swift00:02
*** germs has quit IRC00:15
*** germs has joined #openstack-swift00:17
*** gyee has quit IRC00:20
*** pdardeau has joined #openstack-swift00:29
openstackgerritTim Burke proposed openstack/swift feature/deep: Fix up probetests check on feature/deep  https://review.openstack.org/55029100:30
*** germs has quit IRC00:32
*** tovin07_ has joined #openstack-swift00:36
*** germs has joined #openstack-swift00:42
*** germs has quit IRC00:42
*** germs has joined #openstack-swift00:42
*** germs has quit IRC00:48
*** germs has joined #openstack-swift00:49
*** germs has quit IRC00:53
*** mingyu has quit IRC00:57
*** gerardgine has quit IRC01:06
*** NM has joined #openstack-swift01:12
*** itlinux has quit IRC01:12
*** NM has quit IRC01:18
*** gerardgine has joined #openstack-swift01:25
*** gerardgine has quit IRC02:05
*** gerardgine has joined #openstack-swift02:20
*** germs has joined #openstack-swift02:20
*** germs has quit IRC02:20
*** germs has joined #openstack-swift02:20
*** germs has quit IRC02:43
*** germs has joined #openstack-swift02:43
*** germs has quit IRC02:43
*** germs has joined #openstack-swift02:43
openstackgerritTim Burke proposed openstack/swift feature/deep: Fix up probetests check on feature/deep  https://review.openstack.org/55029102:45
*** gerardgine has quit IRC02:47
*** psachin has joined #openstack-swift02:47
*** germs has quit IRC02:48
*** psachin has quit IRC02:59
*** psachin has joined #openstack-swift03:03
*** psachin has quit IRC03:05
*** gerardgine has joined #openstack-swift03:07
*** psachin has joined #openstack-swift03:08
*** psachin has quit IRC03:17
*** psachin has joined #openstack-swift03:35
*** cshastri has joined #openstack-swift03:36
*** kei_yama has quit IRC03:47
*** gkadam has joined #openstack-swift03:59
*** kei_yama has joined #openstack-swift04:02
*** gerardgine has quit IRC04:03
*** kei_yama has quit IRC04:10
*** kei_yama has joined #openstack-swift04:12
*** ianychoi_ has joined #openstack-swift04:12
*** mlanner has quit IRC04:14
*** kei_yama_ has joined #openstack-swift04:15
*** mlanner has joined #openstack-swift04:15
*** ianychoi has quit IRC04:15
*** kei_yama has quit IRC04:17
*** thurloat has quit IRC04:17
*** kei_yama_ has quit IRC04:21
*** links has joined #openstack-swift04:22
*** kei_yama has joined #openstack-swift04:30
*** sjmurdoch128 has joined #openstack-swift04:44
*** sjmurdoch128 has quit IRC04:46
*** linuxdaemonah has joined #openstack-swift05:02
*** arunman has joined #openstack-swift05:24
*** arunmani_ has joined #openstack-swift05:29
*** arunman has quit IRC05:30
*** thurloat has joined #openstack-swift05:36
*** ianychoi__ has joined #openstack-swift06:43
*** ianychoi_ has quit IRC06:46
kota_hello world06:48
*** gkadam has quit IRC06:50
*** ianychoi__ is now known as ianychoi06:51
*** pcaruana has joined #openstack-swift07:04
*** tovin07 has joined #openstack-swift07:28
*** armaan has quit IRC07:36
*** armaan has joined #openstack-swift07:36
*** gkadam has joined #openstack-swift07:40
*** hseipp has joined #openstack-swift07:43
*** tesseract has joined #openstack-swift08:14
*** geaaru has joined #openstack-swift08:30
*** hoonetorg has quit IRC08:30
*** mingyu has joined #openstack-swift08:30
*** hoonetorg has joined #openstack-swift08:43
*** kei_yama has quit IRC08:47
*** hseipp has quit IRC09:03
*** hseipp has joined #openstack-swift09:14
*** hseipp has quit IRC09:31
*** hseipp has joined #openstack-swift09:31
*** hseipp has quit IRC09:32
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/s3api: Update S3api Docs  https://review.openstack.org/55285309:33
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/s3api: Cleanup swift/common/middleware/s3api/test dir  https://review.openstack.org/55285409:33
openstackgerritAlistair Coles proposed openstack/swift feature/deep: Fix up probetests check on feature/deep  https://review.openstack.org/55029109:37
*** hseipp has joined #openstack-swift09:40
*** tovin07_ has quit IRC10:06
*** rcernin has quit IRC10:13
*** mingyu has quit IRC10:18
openstackgerritAlexandre Lécuyer proposed openstack/swift master: Add optional parameters to common/utils fallocate()  https://review.openstack.org/54803410:25
root_743hello, is there any way to rebuild the rings of my swift single controller cluster? all of the images on swift got moved to a "quarantined" directory in my filesystem because (and I think this was the issue), swift.conf got modified by adding swift_hash_path_preffix (originally we only had swift_hash_path_suffix and now the configuration was reverted back) but the files are still in quarantined10:31
*** mingyu has joined #openstack-swift10:42
*** cshastri has quit IRC10:43
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/s3api: Avoid global CONF instance  https://review.openstack.org/55127310:48
*** ianychoi has quit IRC11:07
*** ianychoi has joined #openstack-swift11:13
kei-ichiI'm now sending this patch and get one comment from infra team. https://review.openstack.org/#/c/552282/11:15
patchbotpatch 552282 - openstack-infra/project-config - Add slogging projects11:15
kei-ichiInfra team propose me how to use swift IRC chanel to inform slogging event.11:15
kei-ichiSo let me confirm one thing. Can I use swift IRC channel from slogging to inform each events ? (like "patchset-created" , "change-merged" , "x-vrif-minus-2")11:15
kei-ichiNow I'm fixing unittest of sloggin so that my patch can pass zuul check.11:16
kei-ichiI made mistake in preparation for this so once remove notmyname mattoliverau kota_ from review of this. (To avoid many number of mail wiil be sent to you ...)11:16
kei-ichiIf I can pass zuul check, let me add everyone as reviewer again!11:16
*** HW-Peter has quit IRC11:22
*** armaan has quit IRC11:25
acoleskei-chi: does #openstack-slogging IRC channel already exist? if it does exist then IMHO that would be the best place for gerrit to announce changes to the slogging project. If it does not exist then the question is perhaps 'does slogging need a separate IRC channel?' and the best way to get consensus on that question would be to add it as a topic for today's swift meeting. You can edit the wiki https://wiki.openstack.org/11:27
acoleswiki/Meetings/Swift11:27
acoleshttps://wiki.openstack.org/wiki/Meetings/Swift11:27
*** arunmani_ has quit IRC11:37
openstackgerritAlistair Coles proposed openstack/swift feature/deep: Fix up probetests check on feature/deep  https://review.openstack.org/55029111:42
kei-ichiacoles Thank you!11:42
kei-ichiI didn't prepare #openstack-slogging yet. But firstly try to create that. After sending patch, infra team propose me about sharing IRC channel so I'd like to confirm to swift team first.11:45
kei-ichiI also think it is better to add this topic at IRC. I'll edit above page o/11:45
*** armaan has joined #openstack-swift11:53
*** armaan has quit IRC12:05
*** armaan has joined #openstack-swift12:06
*** kei-ichi has quit IRC12:08
*** kei-ichi has joined #openstack-swift12:09
*** armaan has quit IRC12:10
*** zhongjun has quit IRC12:14
*** armaan has joined #openstack-swift12:24
*** armaan has quit IRC12:24
*** armaan_ has joined #openstack-swift12:29
*** armaan_ has quit IRC12:53
*** armaan has joined #openstack-swift12:53
openstackgerritAlexandre Lécuyer proposed openstack/swift master: Change object_audit_location_generator() to yield for a single policy.  https://review.openstack.org/54858913:15
*** saint_ has joined #openstack-swift13:28
*** arunman has joined #openstack-swift13:34
*** zhongjun has joined #openstack-swift13:53
*** arunman has quit IRC14:05
acolestimburke: probe tests passed at last https://review.openstack.org/550291 :)14:14
patchbotpatch 550291 - swift (feature/deep) - Fix up probetests check on feature/deep14:14
*** ukaynar has joined #openstack-swift14:26
*** psachin has quit IRC14:30
*** germs has joined #openstack-swift14:35
*** germs has quit IRC14:35
*** germs has joined #openstack-swift14:35
*** wes_dillingham has joined #openstack-swift14:39
*** mvk has quit IRC14:40
*** mrjk__ has quit IRC14:44
openstackgerritAlistair Coles proposed openstack/swift feature/deep: Make concatentaion of shard listings more robust  https://review.openstack.org/55291314:45
*** zaitcev__ has joined #openstack-swift14:45
*** ChanServ sets mode: +v zaitcev__14:45
*** mrjk has joined #openstack-swift14:46
*** zaitcev_ has quit IRC14:48
openstackgerritAlistair Coles proposed openstack/swift feature/deep: Fix up probetests check on feature/deep  https://review.openstack.org/55029114:51
openstackgerritAlistair Coles proposed openstack/swift feature/deep: Pass bind_ip to whataremyips  https://review.openstack.org/55291614:51
*** germs has quit IRC14:55
*** mvk has joined #openstack-swift14:55
*** itlinux has joined #openstack-swift15:12
*** germs has joined #openstack-swift15:13
*** germs has quit IRC15:14
*** germs has joined #openstack-swift15:15
*** germs has quit IRC15:15
*** germs has joined #openstack-swift15:15
*** links has quit IRC15:18
*** wes_dillingham has quit IRC15:22
*** wes_dillingham has joined #openstack-swift15:33
*** ukaynar has quit IRC15:33
*** SkyRocknRoll has joined #openstack-swift15:48
timburkeacoles: w00t! thanks for running that down...15:59
acolestimburke: the logs were quite useful so thanks for those !15:59
*** gerardgine has joined #openstack-swift15:59
timburkeacoles: did we ever get a good state-machine diagram going? between p 551356 and a growing thought that we may want a new state between FOUND and CREATED (so all replicas could look for shards before a leader or external actor decides on a set to use -- COMMITTED, maybe?) the number of states is getting fairly large...16:08
patchbothttps://review.openstack.org/#/c/551356/ - swift (feature/deep) - Use state rather than subtlety to make shards of s...16:08
*** ukaynar has joined #openstack-swift16:10
acolestimburke: re state machine diagram - no, re maybe needing another state - agree16:10
*** ukaynar has quit IRC16:11
acolesalthough IDK whether we'd deliberately allow all replicas to scan for shard ranges - we also had a though for the 'leader'/external actor to prescribe the number of ranges to find16:12
*** ukaynar has joined #openstack-swift16:12
acolesAlthough, I guess all replicas could find potentially differing numbers of ranges and leader selects best set???16:13
timburkethat's along the lines of what i was thinking. and then you can have the find_shard_ranges script be able to serve from cache instead of taking... tens of minutes? hours? to run16:19
acolestimburke: yes. I do like the idea that an external actor just needs to transition states, approve ranges and leave real work to swift processes16:25
acolestimburke: and surely there's a tequila per shard range state once we're done? ;-)16:25
acolestimburke: so were you thinking we'd just have a sharder daemon go ahead and scan for ranges on any sizeable container, then drop some notification and wait for external actor to come along and approve some ranges?16:37
acoleshmmm, the ranges might get stale16:38
timburkemaybe? i hadn't thought it through entirely yet (and might not be in a good state to do so right now)16:38
acoles:)16:38
timburkeyeah, we have to worry about stale ranges. i'm thinking give them a lifetime around one reclaim_age?16:39
timburkeotoh, maybe this is something better kicked out until later... idk16:41
acolesor if actor finds ranges and can infer they may be state, actor deletes them and sharder scans again16:41
acolesstale*16:41
acolesfeature/deep probe tests into gate jobs :D16:44
openstackgerritAlistair Coles proposed openstack/swift feature/deep: Make concatenation of shard listings more robust  https://review.openstack.org/55291316:51
openstackgerritAlexandre Lécuyer proposed openstack/swift master: Change object_audit_location_generator() to yield for a single policy.  https://review.openstack.org/54858916:57
*** armaan has quit IRC16:59
*** SkyRocknRoll has quit IRC17:00
*** gerardgine has quit IRC17:00
*** gkadam has quit IRC17:05
*** armaan has joined #openstack-swift17:13
*** gerardgine has joined #openstack-swift17:14
*** pcaruana has quit IRC17:18
*** hseipp has quit IRC17:25
*** gerardgine has quit IRC17:50
*** itlinux has quit IRC17:50
*** gerardgine has joined #openstack-swift17:50
*** links has joined #openstack-swift17:57
*** itlinux has joined #openstack-swift17:58
*** gyee has joined #openstack-swift18:05
*** mrjk has quit IRC18:08
*** mrjk has joined #openstack-swift18:09
*** geaaru has quit IRC18:14
*** hseipp has joined #openstack-swift18:16
*** hseipp has quit IRC18:36
*** mvk has quit IRC18:41
openstackgerritMerged openstack/swift feature/deep: Pass bind_ip to whataremyips  https://review.openstack.org/55291618:50
*** itlinux has quit IRC18:57
*** itlinux_ has joined #openstack-swift18:57
*** mvk has joined #openstack-swift19:09
*** tesseract has quit IRC19:20
openstackgerritMerged openstack/swift feature/deep: Fix up probetests check on feature/deep  https://review.openstack.org/55029119:31
*** links has quit IRC19:34
*** itlinux_ has quit IRC19:40
timburkeacoles: i wonder if we could get away from having db_state in favor of looking at the state recorded in own_shard_range...19:47
notmynamereminder, especially for those in the US. the meeting is at 2100UTC. that's in an hour. not now19:50
*** itlinux has joined #openstack-swift19:53
tdasilvaoh20:00
tdasilvathat's right...thanks for the reminder20:00
acolestimburke: maybe but there's times when the two are not the same...for example, a shard range moves to sharded once any of the replicas has completed cleaving, whereas other replica's db_state may still be sharding. so there's sometimes a useful distinction. But every now and again I have exactly the thought you had :)20:06
* acoles back later for mtg20:06
spotzAnyone about to answer a quick CLI question?20:11
*** Supun has joined #openstack-swift20:12
*** silor has joined #openstack-swift20:15
*** geaaru has joined #openstack-swift20:20
openstackgerritMerged openstack/swift master: Add handoffs-only mode to DB replicators.  https://review.openstack.org/54549720:36
notmynamespotz: don't ask to ask. just ask :-)20:36
*** silor has quit IRC20:36
spotznotmyname: I tried to be polite:) But trying to expire an object with swift post and it returns the help20:37
spotzso could be a bug, removed functionaility or....20:37
claygtimburke: kota_ : any takers?  http://lists.openstack.org/pipermail/openstack/2018-March/046188.html20:37
*** m_kazuhiro has joined #openstack-swift20:37
spotznotmyname: And actually if I cut and paste the -H content-length from the help it returns the help as well20:38
notmynamespotz: can you pastebin what you're doing (redact any passwords, etc)20:38
spotznotmyname: Sure20:38
spotznotmyname: https://pastebin.com/ZSDsXzqQ20:39
*** Supun has quit IRC20:39
spotznotmyname: packstack stable/pike if that helps any20:40
notmynamespotz: seems to be the --os-project-name that's getting in the way. not sure why, yet20:43
*** wes_dillingham has quit IRC20:44
spotznotmyname: Interesting because it worked fine for both read and write ACLs20:44
spotzas well as creating the container and uploading the object20:44
torgomaticalso, once you get past that, you're going to want "X-Delete-After: 3600", not -At.20:50
timburkeclayg: yeah, seems right. the trick is that you don't want any single host to have more than m frags; ceil((k+m)/m) hosts should ensure that20:50
timburkethe ceil is important, tho -- 8+2 worked out nicely, but 7+3 is a little more awkward; (k+m)/m gets you 3.33 -- if you round nearest, you get three, so a layout like 4-4-320:51
timburkeand a single host failure has a 2/3 chance of causing availability issues20:51
spotztorgomatic: Yeah that was in my original, I've been trying different combos trying to figure out what was going on20:51
timburkeceil forces you to four, so 3-3-2-2 and you can lose one but still reconstruct (assuming your dispersion is good! not necessarily the case, depending on relative drive weights)20:52
torgomaticspotz: just spitballing here since swiftclient is not my area of expertise, but have you tried "swift --os-project-name dev post ..." ?20:54
*** gerardgine has quit IRC20:54
notmynameyes! that's it20:54
spotztorgomatic: let me give it a shot, but not sure why it would only effect headers20:54
*** mathiasb has joined #openstack-swift20:54
torgomaticI couldn't tell you, but if you're getting the help message, something's confused the argument parser20:55
notmynamespotz: turns out the options parsing code isn't that straightforward. and the commands (eg "post") are sortof compartmentalized and almost implemented as plugins (not really, but sortof). beasically, this means that options order is important and "post" doesn't know about that auth thing, but the "swift" command does20:55
spotztorgomatic notmyname - moving it up between swift and posr says no such command20:56
notmynamemine works20:56
notmynameswift --os-project-name dev post resources files/dev.png --header "X-Delete-At:3600"20:56
notmynameObject POST failed: http://192.168.56.2:8080/v1/AUTH_test/resources/files/dev.png 400 Bad Request   X-Delete-At in past20:56
notmynameFailed Transaction ID: txbec23d214a6249c88ff25-005aa98c0dsaio20:56
kota_good morning20:56
notmynamespotz: I'm using swiftclient 3.4.020:57
timburkecrap, i thought when i moved it to argparse i made the arg-ordering problem go away...20:57
spotznotmyname: Ha when I cut and pasted yours in it worked:)20:57
spotzyou used --header I used -H20:57
notmynameah20:58
notmynamenormally I use curl ;-)20:58
spotzshould be the same but yeah20:58
timburkethat just raises more questions :-/ hmm...20:58
spotznotmyname: I'm working on a lesson on ACLs from the CLI. I find lots of issues when I make new lessons:)20:58
kota_clayg: let me check the context of the e-mail thread for the ec question...20:59
notmynameah yes. I've been asked about that before. it's a tricky topic20:59
torgomaticat some point it's just easier to cat > /dev/tcp/192.168.56.2/8080 ;)20:59
notmynameswift team meeting time in #openstack-meeting20:59
*** zaitcev__ is now known as zaitcev20:59
spotznotmyname: Let me know if you want me to bug it. Thanks for the help and happy meeting:)20:59
notmynamespotz: maybe the -H/--header. not the other one21:00
spotznotmyname: will do21:00
notmynamethanks21:01
*** wes_dillingham has joined #openstack-swift21:03
*** rcernin has joined #openstack-swift21:08
*** gerardgine has joined #openstack-swift21:15
notmynamemattoliverau: in the meeting you just got volunteered to rereview https://review.openstack.org/#/c/551956/21:21
patchbotpatch 551956 - python-swiftclient - Add force auth retry mode in swiftclient21:21
clayggreat meeting21:39
*** acoles has left #openstack-swift21:40
*** itlinux has quit IRC21:40
*** acoles has joined #openstack-swift21:40
*** ChanServ sets mode: +v acoles21:40
* kota_ is trying to find my breakfast21:40
acolestimburke: hehe I was about to say thanks again to tdasilva for probe tests in zuul21:40
*** m_kazuhiro has quit IRC21:40
notmynamethose probe tests are great!21:40
acolesnotmyname: ...until they fail ;)21:41
timburkewell that's why we made it non-voting :P21:49
*** dmorita has joined #openstack-swift21:50
*** dmorita has quit IRC21:51
*** dmorita has joined #openstack-swift21:51
*** dmorita has quit IRC21:52
*** geaaru has quit IRC22:16
*** dmorita has joined #openstack-swift22:22
*** dmorita has quit IRC22:24
*** dmorita has joined #openstack-swift22:24
*** ukaynar has quit IRC22:26
*** dmorita has quit IRC22:29
*** wes_dillingham has quit IRC22:38
*** dmorita has joined #openstack-swift22:39
*** dmorita has quit IRC22:40
*** armaan has quit IRC23:11
openstackgerritSamuel Merritt proposed openstack/swift master: Remove object replicator's lockup detector/mitigator.  https://review.openstack.org/55265923:15
*** kei_yama has joined #openstack-swift23:21
*** NM has joined #openstack-swift23:22
*** NM has quit IRC23:29
*** itlinux has joined #openstack-swift23:47
*** NM has joined #openstack-swift23:49
mattoliverauSure, sorry, been hit with a gastro bug, /me feeling much better but now Bec has it. So sitting in a hospital with a drip to rehydrate her (just for safely re: baby)... Fun times :(23:56

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!