| clayg | comparing to the gate to see how it's *supposed* to work was a protip - thanks | 00:01 |
|---|---|---|
| *** links has joined #openstack-swift | 00:06 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Make probe tests work when policy-0 is EC https://review.openstack.org/469688 | 00:09 |
| timburke | notmyname: thoughts on https://review.openstack.org/#/c/464982/ ? given the reactions of clayg and acoles in the meeting, maybe it's best to abandon, given the the alternatives of either (1) majorly reworking the unit test, (2) having the unit test diverge pretty far from what's actually possible in newton, or (3) also backporting https://review.openstack.org/#/c/425493/ | 00:17 |
| patchbot | patch 464982 - swift (stable/newton) - Do not sync suffixes when remote rejects reconstru... | 00:17 |
| patchbot | patch 425493 - swift - Deprecate broken handoffs_first in favor of handof... (MERGED) | 00:17 |
| notmyname | I think you summed it up nicely :-) | 00:17 |
| *** alenavolk has joined #openstack-swift | 00:34 | |
| *** alenavolk has quit IRC | 00:38 | |
| *** tonanhngo has quit IRC | 00:47 | |
| *** zhengyin has joined #openstack-swift | 00:49 | |
| *** alenavolk has joined #openstack-swift | 00:50 | |
| *** tonanhngo has joined #openstack-swift | 00:54 | |
| *** gyee has quit IRC | 00:54 | |
| *** tonanhngo has quit IRC | 00:58 | |
| *** tovin07_ has joined #openstack-swift | 01:02 | |
| *** Sukhdev has quit IRC | 01:04 | |
| *** j_king_ has quit IRC | 01:10 | |
| timburke | thanks kota_! i've been running with ec as default on my vagrant saio, and always scratched my head about why some probetests wouldn't work | 01:12 |
| timburke | now i just need to get it to stop failing with "[Errno 24] Too many open files" when i run all of them :-/ | 01:12 |
| *** klrmn has quit IRC | 01:12 | |
| *** links has quit IRC | 01:14 | |
| zhongjun | notmyname: thanks, yes, the time was really more convenient for me. Nice to see you too :) | 01:36 |
| *** jeffl has joined #openstack-swift | 01:37 | |
| *** lan has joined #openstack-swift | 01:52 | |
| *** lan1 has joined #openstack-swift | 02:00 | |
| *** lan has quit IRC | 02:01 | |
| *** lan2 has joined #openstack-swift | 02:03 | |
| *** lan1 has quit IRC | 02:06 | |
| *** klrmn has joined #openstack-swift | 02:14 | |
| *** alenavolk has quit IRC | 02:32 | |
| *** links has joined #openstack-swift | 02:53 | |
| *** alenavolk has joined #openstack-swift | 02:53 | |
| openstackgerrit | Merged openstack/swift master: Trivial fix typos https://review.openstack.org/469009 | 03:09 |
| *** Sukhdev has joined #openstack-swift | 03:11 | |
| *** Sukhdev has quit IRC | 03:20 | |
| *** Sukhdev has joined #openstack-swift | 03:22 | |
| *** chsc has joined #openstack-swift | 03:25 | |
| *** zhurong has joined #openstack-swift | 03:27 | |
| clarkb | clayg: fyi I pushed https://review.openstack.org/469732 to hopefully make that better for you in the future | 03:30 |
| patchbot | patch 469732 - openstack-dev/devstack - Run devstack CA and cert setup early | 03:30 |
| *** psachin has joined #openstack-swift | 03:31 | |
| *** cshastri has joined #openstack-swift | 03:32 | |
| *** kei_yama has quit IRC | 03:45 | |
| *** chsc has quit IRC | 03:45 | |
| *** lan2 has quit IRC | 03:49 | |
| *** kei_yama has joined #openstack-swift | 04:00 | |
| *** klrmn has quit IRC | 04:06 | |
| *** Sukhdev has quit IRC | 04:08 | |
| *** alenavolk has quit IRC | 04:10 | |
| *** tovin07_ has quit IRC | 04:22 | |
| *** zhurong has quit IRC | 04:27 | |
| mahatic | tdasilva: great to see you already pushed an update to https://review.openstack.org/#/c/371150 :) | 04:28 |
| patchbot | patch 371150 - swift - Return 404 on a GET if tombstone is newer | 04:28 |
| *** zhurong has joined #openstack-swift | 04:36 | |
| *** pcaruana has joined #openstack-swift | 04:43 | |
| *** zhengyin has quit IRC | 05:04 | |
| *** alenavolk has joined #openstack-swift | 05:13 | |
| *** zhengyin has joined #openstack-swift | 05:16 | |
| *** zhurong has quit IRC | 05:16 | |
| *** zhurong has joined #openstack-swift | 05:17 | |
| *** jaosorior_away is now known as jaosorior | 05:24 | |
| *** pcaruana has quit IRC | 05:49 | |
| *** aselius has quit IRC | 05:53 | |
| *** mingyu has joined #openstack-swift | 05:53 | |
| *** chsc has joined #openstack-swift | 05:54 | |
| *** mingyu has quit IRC | 05:55 | |
| *** rcernin has joined #openstack-swift | 05:58 | |
| *** pcaruana has joined #openstack-swift | 06:02 | |
| *** chsc has quit IRC | 06:15 | |
| *** Dinesh_Bhor has quit IRC | 06:23 | |
| *** hseipp has joined #openstack-swift | 06:44 | |
| *** tesseract has joined #openstack-swift | 06:51 | |
| *** ChubYann has quit IRC | 06:55 | |
| *** tovin07_ has joined #openstack-swift | 06:59 | |
| *** cschwede_ has quit IRC | 07:05 | |
| *** cschwede has joined #openstack-swift | 07:09 | |
| *** aselius has joined #openstack-swift | 07:20 | |
| *** cshastri has quit IRC | 07:23 | |
| *** silor has joined #openstack-swift | 07:23 | |
| *** cshastri has joined #openstack-swift | 07:24 | |
| *** zhengyin has quit IRC | 07:27 | |
| *** zhengyin has joined #openstack-swift | 07:27 | |
| *** oshritf has joined #openstack-swift | 07:29 | |
| *** geaaru has joined #openstack-swift | 07:44 | |
| *** silor has quit IRC | 08:01 | |
| *** silor has joined #openstack-swift | 08:01 | |
| *** geaaru has quit IRC | 08:05 | |
| *** hseipp has quit IRC | 08:08 | |
| acoles | good morning | 08:17 |
| *** geaaru has joined #openstack-swift | 08:19 | |
| *** silor has quit IRC | 08:19 | |
| *** silor has joined #openstack-swift | 08:20 | |
| *** zhurong has quit IRC | 08:29 | |
| *** alenavolk has quit IRC | 08:40 | |
| *** alenavolk has joined #openstack-swift | 08:42 | |
| *** alenavolk has joined #openstack-swift | 08:43 | |
| *** alenavolk has joined #openstack-swift | 08:43 | |
| *** alenavolk has quit IRC | 08:44 | |
| *** alenavolk has joined #openstack-swift | 08:45 | |
| *** alenavolk has quit IRC | 08:46 | |
| *** alenavolk has joined #openstack-swift | 08:47 | |
| *** alenavolk has quit IRC | 08:56 | |
| *** alenavolk has joined #openstack-swift | 08:56 | |
| *** alenavolk has quit IRC | 08:57 | |
| *** alenavolk has joined #openstack-swift | 08:57 | |
| *** alenavolk has quit IRC | 08:57 | |
| *** skudlik has quit IRC | 08:59 | |
| *** f0o has quit IRC | 09:03 | |
| *** f0o has joined #openstack-swift | 09:04 | |
| openstackgerrit | Van Hung Pham proposed openstack/swift master: Replace assertRaisesRegexp with assertRaisesRegex https://review.openstack.org/469811 | 09:06 |
| *** f0o has quit IRC | 09:09 | |
| *** f0o has joined #openstack-swift | 09:09 | |
| mahatic | acoles: good morning | 09:10 |
| *** silor1 has joined #openstack-swift | 09:13 | |
| *** silor has quit IRC | 09:15 | |
| *** silor1 is now known as silor | 09:15 | |
| acoles | mahatic: o/ | 09:16 |
| rledisez | hey, do you know cbartz? what timezone he's in? | 09:16 |
| openstackgerrit | Romain LE DISEZ proposed openstack/swift master: Preserve query params in tempurl https://review.openstack.org/333331 | 09:17 |
| rledisez | i'm looking at his patch ^, i want it in swift :) | 09:17 |
| acoles | rledisez: no but I recognise the nick, they're not in channel at the moment though | 09:19 |
| acoles | rledisez: email address suggests in EU | 09:20 |
| rledisez | acoles: yeah. he'll get notified by gerrit, eventually he'll join | 09:20 |
| *** hseipp has joined #openstack-swift | 09:25 | |
| *** aselius has quit IRC | 09:30 | |
| *** openstackstatus has quit IRC | 09:36 | |
| *** openstackstatus has joined #openstack-swift | 09:39 | |
| *** ChanServ sets mode: +v openstackstatus | 09:39 | |
| -openstackstatus- NOTICE: There is a known issue with setuptools 36.0.0 and errors about the "six" package. For current details see https://github.com/pypa/setuptools/issues/1042 and monitor #openstack-infra | 09:45 | |
| *** geaaru has quit IRC | 09:47 | |
| kota_ | acoles, rledisez, mahatic: morning | 09:56 |
| kota_ | and just preparing to go home. I finished https://review.openstack.org/#/c/302494 :) | 09:57 |
| patchbot | patch 302494 - swift - Sync metadata in 'rsync_then_merge' in db_replicator | 09:57 |
| *** tovin07_ has quit IRC | 09:57 | |
| *** geaaru has joined #openstack-swift | 10:00 | |
| *** ujjain has quit IRC | 10:11 | |
| *** zhurong has joined #openstack-swift | 10:35 | |
| *** Dinesh_Bhor has joined #openstack-swift | 10:49 | |
| *** zhengyin has quit IRC | 10:54 | |
| *** mat128 has joined #openstack-swift | 11:12 | |
| openstackgerrit | Alistair Coles proposed openstack/swift master: Ring doc cleanups https://review.openstack.org/465184 | 11:20 |
| *** cbartz has joined #openstack-swift | 11:20 | |
| *** mat128 has quit IRC | 11:33 | |
| *** cshastri has quit IRC | 11:33 | |
| *** mat128 has joined #openstack-swift | 11:33 | |
| *** psachin has quit IRC | 11:34 | |
| *** links has quit IRC | 11:37 | |
| *** skudlik has joined #openstack-swift | 11:38 | |
| *** kei_yama has quit IRC | 11:41 | |
| *** silor has quit IRC | 11:43 | |
| *** silor has joined #openstack-swift | 11:43 | |
| *** mat128 has quit IRC | 11:45 | |
| acoles | notmyname: re pike wsgi goal -> https://review.openstack.org/469847 | 11:47 |
| patchbot | patch 469847 - governance - Add swift artifacts for deploy-api-in-wsgi goal | 11:47 |
| acoles | notmyname: It is set to WIP pending you taking a look | 11:47 |
| *** MVenesio has joined #openstack-swift | 11:56 | |
| *** psachin has joined #openstack-swift | 11:58 | |
| *** cshastri has joined #openstack-swift | 12:05 | |
| *** NM has joined #openstack-swift | 12:13 | |
| *** jeffl has quit IRC | 12:14 | |
| tdasilva | good morning | 12:15 |
| *** zhurong has quit IRC | 12:16 | |
| psachin | Good Morning tdasilva | 12:21 |
| acoles | tdasilva: o/ | 12:23 |
| tdasilva | psachin, acoles: o/ | 12:24 |
| acoles | timburke: I revisited patch 466952 and also noticed patch 469269 both of which have the sorting method validation change - had you intended to remove it from 466952? | 12:25 |
| patchbot | https://review.openstack.org/#/c/466952/ - swift - Follow-up for per-policy proxy configs | 12:25 |
| patchbot | https://review.openstack.org/#/c/469269/ - swift - Add validation for sorting_method values | 12:25 |
| *** remixtj is now known as remix_tj | 12:25 | |
| acoles | timburke: IIRC it was the only real blocker for me on 466952 so if that was updated we could maybe get it merged?? | 12:26 |
| *** Dinesh_Bhor has quit IRC | 12:45 | |
| *** Dinesh_Bhor has joined #openstack-swift | 12:46 | |
| *** Dinesh_Bhor has quit IRC | 12:49 | |
| *** Dinesh_Bhor has joined #openstack-swift | 12:49 | |
| *** Dinesh_Bhor has quit IRC | 12:49 | |
| *** Dinesh_Bhor has joined #openstack-swift | 12:50 | |
| *** cshastri has quit IRC | 12:58 | |
| *** mingyu has joined #openstack-swift | 12:58 | |
| *** mingyu has quit IRC | 13:00 | |
| *** mdrabe has joined #openstack-swift | 13:08 | |
| *** lucasxu has joined #openstack-swift | 13:08 | |
| *** klamath has joined #openstack-swift | 13:09 | |
| *** klamath has quit IRC | 13:09 | |
| *** klamath has joined #openstack-swift | 13:09 | |
| *** psachin has quit IRC | 13:10 | |
| *** skudlik has quit IRC | 13:41 | |
| mahatic | kota_: thanks for your comments on patch 439572 | 13:41 |
| patchbot | https://review.openstack.org/#/c/439572/ - swift - Limit number of revert tombstone SSYNC requests | 13:41 |
| mahatic | clayg: you may want to jump in on that discussion (i.e. should we push tombstone to 'quorum nodes' or not) | 13:42 |
| rledisez | using python logger in swift is sometimes tricky, is there any tip to avoid that? | 13:43 |
| rledisez | object-server: STDERR: error: cannot switch to a different thread | 13:43 |
| acoles | mahatic: did you start reviewing any of the patches mentioned in the 0700 meeting? I was about to start on patch 435771 | 13:43 |
| patchbot | https://review.openstack.org/#/c/435771/ - swift - Rewrite redirection in cname_lookup & domain_remap | 13:43 |
| mahatic | acoles: sounds good. I'm looking into patch 435768 | 13:44 |
| patchbot | https://review.openstack.org/#/c/435768/ - swift - Allow to configure the nameservers in cname_lookup | 13:44 |
| acoles | mahatic: excellent, so no overlap | 13:45 |
| mahatic | nope | 13:45 |
| acoles | rledisez: do you see that error when you try to add a log call in a thread? | 13:46 |
| rledisez | acoles: alecuyer will answer you as he is having this issue :) | 13:48 |
| acoles | ok, sadly I don't have an answer :/ but I do recall seeing similar | 13:49 |
| rledisez | acoles: i want to fix https://bugs.launchpad.net/swift/+bug/1652323 | 13:49 |
| openstack | Launchpad bug 1652323 in OpenStack Object Storage (swift) "ssync syncs an expired object as a tombstone, probe test_object_expirer fails" [Medium,Confirmed] | 13:49 |
| rledisez | i'm thinking of sending all objects/fragments, expired or not. i think it should be doable now with the open_expired param added last night. what do you think about that? it would be transfering data thazt would be deleted eventually, but it's a fix, better than nothing | 13:49 |
| alecuyer | acoles: hello ! for example within the create() function of the BaseDiskFile - in the patched version, this uses a module to get a file like object from a "vfile" module. If I try log use a python logger there, I get the message that rledisez just posted | 13:50 |
| acoles | rledisez: the remaining problem I see is when needing to reconstruct an EC frag that has expired - the GETs for other frags will fail | 13:51 |
| alecuyer | so far I've been using my own dumb logger instead but I'm trying to figure out how to use the regular logging module without hanging the object server :) (I'm guessing this has something to do with eventlet but haven't figuret out how to fix it) | 13:52 |
| *** hseipp has quit IRC | 13:56 | |
| acoles | alecuyer: hi! I don't see anything obvious in the create() method to explain the error, but it is a context manager, so is it while using the writer class? e.g. the put ends up executes finalize_put in a thread? | 14:02 |
| acoles | rledisez: we could send a special header with those GETs to force object servers to return expired frags, or... we could just do nothing i.e. not sync with a tombstone (that is wrong) but also not send data - and just accept the inconsistent state until the object expires. IDK what is best. | 14:03 |
| acoles | I know I do no tlike have .ts in one place and .data in another with same timestamps! | 14:04 |
| alecuyer | Yes, this happens while using the writer class, which eventually does tpool_reraise(self._finalize_put .... ) | 14:04 |
| acoles | alecuyer: possibly related to this https://bitbucket.org/eventlet/eventlet/pull-requests/29/fix-use-of-semaphore-with-tpool-issue-137/diff | 14:06 |
| rledisez | acoles: there is already the "replication" header | 14:07 |
| rledisez | x-backend-replication if i remember well | 14:08 |
| rledisez | hum, maybe it's not sent when GET'ing a fragment | 14:09 |
| rledisez | this case looks a lot like dark data, each node think it's the only fragment available in the cluster | 14:10 |
| rledisez | *it has | 14:10 |
| alecuyer | acoles: thanks for the pointer ! reading that thread (hum.. :) ) | 14:10 |
| acoles | alecuyer: sorry I can't help more than to confirm that I think i have seem similar, maybe ask back later - perhaps clayg or torgomatic know of a workaround | 14:11 |
| acoles | rledisez: yeah, I don't think we send replication header with thos GETs. There is user-agent:obj-reconstructor header which is a clue but i think if we took this approach I would prefer to see a distinct header used (it could be 'x-backend-replication') and not overload an existing header | 14:13 |
| *** openstackgerrit has quit IRC | 14:34 | |
| -openstackstatus- NOTICE: python-setuptools 36.0.1 has been released and now making its way into jobs. Feel free to 'recheck' your failures. If you have any problems, please join #openstack-infra | 14:34 | |
| *** jaosorior has quit IRC | 14:36 | |
| clayg | rledisez: alecuyer: you've seen where utils has to monkey patch stdlib longings thread module to use green locking? Maybe avoid stdlib logging ... following existing patterns may have less resistance. | 14:38 |
| *** foutatoro has joined #openstack-swift | 14:42 | |
| alecuyer | clayg: I hadn't ! I'll try that | 14:43 |
| acoles | clayg: the existing pattern being to use utils.get_logger() ?? | 14:47 |
| *** openstackgerrit has joined #openstack-swift | 14:52 | |
| openstackgerrit | Doug Hellmann proposed openstack/swift master: remove remote qualifier from release notes branch scanning https://review.openstack.org/469880 | 14:52 |
| foutatoro | hi all I'm deploying openstack prod install using openstack-ansible project and I would like to know is there a way to backup switf object in order to retrieve data in case of switf node failure ? | 14:58 |
| clayg | :) | 15:00 |
| *** chsc has joined #openstack-swift | 15:08 | |
| *** aselius has joined #openstack-swift | 15:12 | |
| *** alenavolk has joined #openstack-swift | 15:12 | |
| openstackgerrit | Doug Hellmann proposed openstack/swift master: remove remote qualifier from release notes branch scanning https://review.openstack.org/469880 | 15:16 |
| *** dhellmann has joined #openstack-swift | 15:17 | |
| *** chsc has quit IRC | 15:18 | |
| *** gyee has joined #openstack-swift | 15:19 | |
| *** skudlik has joined #openstack-swift | 15:24 | |
| *** trevormc has joined #openstack-swift | 15:25 | |
| *** hseipp has joined #openstack-swift | 15:25 | |
| timburke | acoles: sorry, yeah, i split out the validation but didn't get around to pushing up a new patchset for patch 466952 -- part of me is still debating about whether it's worth renaming a whole bunch of things | 15:30 |
| patchbot | https://review.openstack.org/#/c/466952/ - swift - Follow-up for per-policy proxy configs | 15:30 |
| *** chsc has joined #openstack-swift | 15:32 | |
| *** chsc has joined #openstack-swift | 15:32 | |
| trevormc | hey swift noob here, I'm having trouble getting the auth to work for swift... is there anything I'm missing? http://paste.openstack.org/show/611218/ I think the devstack logs show v1 auth is being used, is this correct? | 15:33 |
| *** openstackgerrit has quit IRC | 15:33 | |
| acoles | timburke: let's do it (rename) where it gets us consistency, e.g. get rid of conf = get_options etc ...now is the time, in future it will just be noise | 15:34 |
| trevormc | .. maybe #openstack is better fit. | 15:34 |
| *** oshritf has quit IRC | 15:35 | |
| acoles | trevormc: looks like you have keystone style vars in the env (OS_*) so its looking for a keystone service at http://192.168.122.150/identity | 15:36 |
| timburke | acoles: but do we have a better name for OverrideConf? i kinda wanted proxy in there somewhere, with the expectation that there would later be object-* overrides as well... but idk if it's worth it? | 15:37 |
| trevormc | acoles, yeah thats normal right? | 15:37 |
| acoles | trevormc: https://docs.openstack.org/developer/python-swiftclient/cli.html#authentication has examples - v1 is equivalent to 'Legacy' | 15:38 |
| trevormc | acoles, thanks I think I'll find my answer there. sorry to interrupt :) | 15:39 |
| acoles | trevormc: also, add a --debug flag to the 'swift list' and you may get some more clues, but it looks like you are failing to get a token from keystone so client is not even talking to swift yet | 15:39 |
| trevormc | acoles, that sounds about right. I was trying to debug yesterday and saw similar signs. | 15:40 |
| clayg | foutatoro: you don't have to backup data you stored in swift if you have multiple disks/replicas it's very durable. If you have multiple nodes it's very available. I'm not aware of a good way to replicate the internals of a storage node - it's very "cattle" vs "pet". You could upload objects you store in your private cloud to an offsite cloud? | 15:40 |
| *** mdrabe has quit IRC | 15:41 | |
| acoles | trevormc: you can cross-check reaching identity service using the 'openstack' CLI (python-openstackclient) | 15:41 |
| acoles | IIRC something like 'openstack token issue' | 15:41 |
| trevormc | acoles, this is good info, I appreciate it. thanks. I think I got enough to start digging now. | 15:42 |
| acoles | trevormc: np. good luck | 15:42 |
| acoles | timburke: did you not already change the class name ... ProxyOverride or ProxyOverrideOptions?? ok, I need to go look at what was in the patch !) | 15:44 |
| timburke | acoles: i did -- i thought you weren't entirely happy with it though? | 15:44 |
| acoles | timburke: but that was last week ;) | 15:45 |
| acoles | timburke: OIC I had a rant, but there was smiley at the end | 15:45 |
| *** mingyu has joined #openstack-swift | 15:46 | |
| timburke | lol | 15:46 |
| *** MVenesio has quit IRC | 15:46 | |
| acoles | timburke: if we're going with s/conf/options/ or s/conf/opts/ then it makes sense to rename the class [Proxy]OverrideOptions . consistency wins IMHO. ... sorry if my comment stalled you | 15:48 |
| acoles | timburke: get_policy_options was the thin end of the wedge - in original patch I just left it there to avoid many line of change in the final review phase | 15:48 |
| acoles | s/left it there/stopped there/ | 15:48 |
| *** mingyu has quit IRC | 15:50 | |
| acoles | timburke: but... it's a naming decision... so I may feel differently next week :P | 15:50 |
| acoles | timburke: there's other good cleanup in that patch so worth keeping it alive IMO | 15:53 |
| *** skudlik has quit IRC | 15:53 | |
| timburke | yep -- i'll get another patchset up today | 16:00 |
| *** mingyu has joined #openstack-swift | 16:01 | |
| *** catintheroof has joined #openstack-swift | 16:02 | |
| *** rcernin has quit IRC | 16:04 | |
| *** mingyu has quit IRC | 16:05 | |
| *** alenavolk has quit IRC | 16:07 | |
| *** alenavolk has joined #openstack-swift | 16:08 | |
| *** chsc has quit IRC | 16:09 | |
| *** lucasxu has quit IRC | 16:10 | |
| *** alenavolk has quit IRC | 16:13 | |
| *** silor has quit IRC | 16:16 | |
| *** mdrabe has joined #openstack-swift | 16:18 | |
| *** alenavolk has joined #openstack-swift | 16:20 | |
| *** alenavolk has quit IRC | 16:21 | |
| *** cbartz has quit IRC | 16:21 | |
| *** alenavolk has joined #openstack-swift | 16:23 | |
| *** alenavolk has quit IRC | 16:24 | |
| *** hseipp has quit IRC | 16:25 | |
| *** catintheroof has quit IRC | 16:27 | |
| *** oshritf has joined #openstack-swift | 16:27 | |
| *** lucasxu has joined #openstack-swift | 16:29 | |
| *** lucasxu has quit IRC | 16:30 | |
| *** trevormc has left #openstack-swift | 16:31 | |
| *** oshritf has quit IRC | 16:31 | |
| *** silor has joined #openstack-swift | 16:32 | |
| *** foutatoro has quit IRC | 16:32 | |
| *** tonanhngo has joined #openstack-swift | 16:33 | |
| *** alenavolk has joined #openstack-swift | 16:39 | |
| *** silor has quit IRC | 16:43 | |
| *** NM has quit IRC | 16:47 | |
| *** NM has joined #openstack-swift | 16:50 | |
| acoles | timburke: here's another patch I revisited today - in this case it's close to squashing but I had one query https://review.openstack.org/#/c/450491/13/swift/container/sharder.py@a471 | 16:52 |
| patchbot | patch 450491 - swift - fixup! Add container sharding to Swift containers | 16:52 |
| acoles | timburke: tbh, if you say 'trust me' then that's fine | 16:52 |
| *** tesseract has quit IRC | 16:54 | |
| timburke | i have no idea anymore. apparently i did it in patchset 6... no, wait, *mattoliverau* did it in ps6... hmm... | 17:01 |
| acoles | timburke: ok let's ping mattoliverau on it | 17:01 |
| acoles | mattoliverau: can you shed any light on this change https://review.openstack.org/#/c/450491/13/swift/container/sharder.py@a471 i.e. does it make sense? thanks | 17:02 |
| patchbot | patch 450491 - swift - fixup! Add container sharding to Swift containers | 17:02 |
| *** klamath has quit IRC | 17:06 | |
| *** chsc has joined #openstack-swift | 17:07 | |
| *** chsc has joined #openstack-swift | 17:07 | |
| *** klamath has joined #openstack-swift | 17:07 | |
| *** alenavolk has quit IRC | 17:08 | |
| *** alenavolk has joined #openstack-swift | 17:09 | |
| *** alenavolk has quit IRC | 17:09 | |
| *** lucasxu has joined #openstack-swift | 17:12 | |
| *** alenavolk has joined #openstack-swift | 17:15 | |
| *** alenavolk has quit IRC | 17:17 | |
| *** alenavolk has joined #openstack-swift | 17:19 | |
| *** alenavolk has quit IRC | 17:19 | |
| clarkb | re liberasurecode being older on xenial, I checked the ocata UCA package and it is version 1.4.0-2~cloud0 so enabling uca in your jobs is one potential way to work around that if necessary | 17:22 |
| timburke | ah, cool! thanks clarkb | 17:22 |
| timburke | ...how does one do that? | 17:22 |
| clarkb | timburke: let me find an example | 17:22 |
| clarkb | https://git.openstack.org/cgit/openstack-dev/devstack/tree/tools/fixup_stuff.sh#n70 is what devstack does. Things like doc build jobs and unittests are different because they use fairly static job templates. I think they have a hook point where you can run something like what devstack runs though. Let me find out | 17:24 |
| *** openstackgerrit has joined #openstack-swift | 17:26 | |
| openstackgerrit | Merged openstack/swift master: Make probe tests work when policy-0 is EC https://review.openstack.org/469688 | 17:26 |
| clarkb | hrm the hook only exists for unittests not docs or pep8 though we can likely change that. if your repo has a tools/test-setup.sh file we run that before executing tox | 17:26 |
| clarkb | timburke: so you can possibly use ^ that script to pull in liberasurecode out of UCA | 17:26 |
| *** oshritf has joined #openstack-swift | 17:26 | |
| timburke | cool, thanks again clarkb! i'll give that a try later | 17:27 |
| *** klrmn has joined #openstack-swift | 17:28 | |
| *** oshritf has quit IRC | 17:33 | |
| *** mat128 has joined #openstack-swift | 17:34 | |
| *** mat128 has quit IRC | 17:35 | |
| *** alenavolk has joined #openstack-swift | 17:35 | |
| *** MVenesio has joined #openstack-swift | 17:35 | |
| *** alenavolk has quit IRC | 17:40 | |
| *** oshritf has joined #openstack-swift | 17:41 | |
| *** alenavolk has joined #openstack-swift | 17:42 | |
| *** mat128 has joined #openstack-swift | 17:46 | |
| *** mat128 has quit IRC | 17:47 | |
| *** oshritf has quit IRC | 17:47 | |
| *** mvk has quit IRC | 17:52 | |
| *** oshritf has joined #openstack-swift | 17:55 | |
| *** oshritf has quit IRC | 18:00 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift master: Updated from global requirements https://review.openstack.org/88736 | 18:02 |
| *** oshritf has joined #openstack-swift | 18:04 | |
| *** chsc has quit IRC | 18:04 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Make Sphinx treat warnings as errors (again) https://review.openstack.org/469315 | 18:04 |
| *** rcernin has joined #openstack-swift | 18:14 | |
| *** chsc has joined #openstack-swift | 18:15 | |
| *** chsc has joined #openstack-swift | 18:15 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: README: dedent the numbered list https://review.openstack.org/469967 | 18:15 |
| *** ChubYann has joined #openstack-swift | 18:17 | |
| *** oshritf has quit IRC | 18:25 | |
| *** gyee has quit IRC | 18:27 | |
| *** mat128 has joined #openstack-swift | 18:37 | |
| *** NM has quit IRC | 18:53 | |
| *** geaaru has quit IRC | 18:54 | |
| *** vint_bra1 has quit IRC | 18:55 | |
| *** NM has joined #openstack-swift | 18:58 | |
| *** vint_bra has joined #openstack-swift | 18:59 | |
| *** oshritf has joined #openstack-swift | 19:04 | |
| *** vint_bra has quit IRC | 19:05 | |
| *** NM has quit IRC | 19:06 | |
| *** oshritf has quit IRC | 19:20 | |
| *** chsc has quit IRC | 19:24 | |
| *** MVenesio has quit IRC | 19:30 | |
| mattoliverau | acoles, timburke: the meta timestamp is used to track the updates to the stats (bytes used and objectives count).. that method grabs the current pivot range but is also updating the stats from container info, so updates the timestamp. We could try and store the old versions and then only update on a change rather then sending new stats on each pass.. which might be a smart move... Though we'd be storing more | 19:32 |
| mattoliverau | metadata | 19:32 |
| timburke | ah, right... that get_info call... | 19:33 |
| timburke | thanks mattoliverau! and good morning :-) | 19:33 |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Prefer auth_uri in functests configs https://review.openstack.org/469984 | 19:33 |
| clayg | ^ I'm going to double check that example config on my saio - it works a treat on devstack! | 19:33 |
| timburke | (a very early morning, no?) | 19:33 |
| clayg | well... i guess well see what the gate dsvm job says about that - I don't want to sound too confident | 19:34 |
| timburke | clayg: i think we might also still need https://review.openstack.org/#/c/456791/ ? maybe | 19:34 |
| patchbot | patch 456791 - python-swiftclient - Allow swift functional tests to use keystone auth uri | 19:34 |
| notmyname | hello, world | 19:35 |
| clayg | not in my testing? Unless when I reverted back to master it didn't take some how... | 19:35 |
| clayg | oh... i mean we probably need that *for* swiftclient functests - just not for swift functests I don't think? | 19:35 |
| notmyname | timburke: help me catch up. are the sphinx errors fixed now? we can land stuff again? | 19:35 |
| timburke | notmyname: yup | 19:36 |
| notmyname | timburke: ack. thanks | 19:36 |
| *** chsc has joined #openstack-swift | 19:36 | |
| notmyname | timburke: oh my. CRITICAL PATCH!!111!!eleventy!!!11! https://review.openstack.org/#/c/469967 | 19:37 |
| patchbot | patch 469967 - swift - README: dedent the numbered list | 19:37 |
| timburke | mergeitmergeitmergeitmergeit | 19:37 |
| mattoliverau | morning :) | 19:41 |
| *** mat128 has quit IRC | 19:54 | |
| *** mat128 has joined #openstack-swift | 19:56 | |
| *** mat128 has quit IRC | 20:01 | |
| *** oshritf has joined #openstack-swift | 20:02 | |
| *** mat128 has joined #openstack-swift | 20:04 | |
| mattoliverau | timburke: yeah, it's very early. Lucy decided 5am was a good time to wake up, for I'm up too :) soz late reply playing with Lucy is busy work ;) | 20:10 |
| mattoliverau | s/for/so/ | 20:10 |
| *** mat128 has quit IRC | 20:21 | |
| *** oshritf has quit IRC | 20:26 | |
| *** alenavolk has quit IRC | 20:32 | |
| *** oshritf has joined #openstack-swift | 20:37 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Follow-up for per-policy proxy configs https://review.openstack.org/466952 | 20:39 |
| *** alenavolk has joined #openstack-swift | 20:42 | |
| openstackgerrit | Merged openstack/swift master: Rewrite redirection in cname_lookup & domain_remap https://review.openstack.org/435771 | 20:49 |
| openstackgerrit | Merged openstack/swift master: README: dedent the numbered list https://review.openstack.org/469967 | 20:54 |
| openstackgerrit | Tim Burke proposed openstack/swift master: domain_remap: stop mangling client-provided paths https://review.openstack.org/464217 | 20:58 |
| openstackgerrit | Tim Burke proposed openstack/swift master: domain_remap: be more careful about client-path mangling https://review.openstack.org/464216 | 20:58 |
| *** oshritf has quit IRC | 21:02 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Allow to configure the nameservers in cname_lookup https://review.openstack.org/435768 | 21:03 |
| *** lucasxu has quit IRC | 21:06 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Remove threads_per_disk from object-server.conf manpages https://review.openstack.org/470013 | 21:08 |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Prefer auth_uri in functests configs https://review.openstack.org/469984 | 21:10 |
| clarkb | clayg: is ^ related to http://logs.openstack.org/74/372374/26/check/gate-swift-dsvm-functional-ubuntu-xenial-nv/a49cbe9/logs/etc/swift/test.conf.txt.gz having both the auth_uri and auth_prefix overlap? | 21:13 |
| clarkb | clayg: I can add a depends on to your change to see how that does | 21:14 |
| *** mvk has joined #openstack-swift | 21:14 | |
| clarkb | clayg: https://review.openstack.org/#/c/372374/ will now run the swift functests using your change and the apache header size change you pointed out as well | 21:14 |
| patchbot | patch 372374 - openstack-dev/devstack - Turn on tls proxy by default for testing | 21:14 |
| clayg | clarkb: not sure I follow? my second patch was because the code I wrote in the first one was wrong when I tested it on a swift-all-in-one configuration without keystone ... I think it is still correct for devstack - and it will ignore/won't matter that the prefix crap overlaps | 21:15 |
| clarkb | clayg: got it | 21:15 |
| clayg | clarkb: oh, yes, patch 372374 should definitely depend on the auth uri fix - good work there | 21:16 |
| patchbot | https://review.openstack.org/#/c/372374/ - openstack-dev/devstack - Turn on tls proxy by default for testing | 21:16 |
| *** gyee has joined #openstack-swift | 21:16 | |
| clayg | clarkb: has there been any consideration to renaming devstack to gatestack? everytime I try to use a devstack environment for development I'm left with the impression that one does not simply "use" a devstack vm - you build them and throw them away - constantly | 21:17 |
| clayg | right now my keystone uwsgi won't start? | 21:18 |
| clayg | http://paste.openstack.org/show/611253/ | 21:18 |
| * clayg shrugs | 21:18 | |
| clarkb | renaming devstack would be a painful transition, we also want to move int to openstack/ instaed of openstack/ | 21:21 |
| clarkb | er instaed of openstack-dev/ | 21:21 |
| clarkb | clayg: I think most people you dev with devstack run it once then start and stop services more manually from then on rather than start stop start stop | 21:22 |
| clarkb | I personally do the throwaway setup... | 21:22 |
| clarkb | boot VM locally, run devstack, $whatever, delete | 21:22 |
| clayg | i wouldn't mind that *so* much - except that for me personally it takes a non-trivial wall time to *build* a devstack vm? I think it encourages "throw it at the gate" sort of workflows - which encourage "passes test; ship it" sort of workflows - which encourage "did you even run this?" sort of bugs :'( | 21:24 |
| clayg | does any of that ring true to you? | 21:24 |
| clarkb | clayg: I think we have a lot of throw it at the gate for various reasons, that is definitely part of it | 21:28 |
| clayg | `time vagrant up` - well see how it goes | 21:28 |
| clarkb | clayg: in theory the systemd changes should make it better to have a long lived devstack. But it is still early days on that | 21:29 |
| clayg | oh sorry, I meant to post this -> http://paste.openstack.org/show/611253/ | 21:30 |
| clayg | i got distracted looking for anything more interesting in /opt/stack/logs/* - but didn't find anything | 21:30 |
| clayg | decided to give up and rebuild from scratch tho | 21:31 |
| clarkb | clayg: with systemd/journald you want to do journalctl -u devstack@keystone.service | 21:31 |
| clayg | oh i did post that, sigh - i'm multitasking to much :'( | 21:31 |
| clayg | oic, I'll try that next time - thanks! | 21:31 |
| clarkb | which should dump the whole set of keystone service logs for you. Its new :) but should be better overall once people get into it | 21:32 |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: Return 404 on a GET if tombstone is newer https://review.openstack.org/371150 | 21:33 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift master: Updated from global requirements https://review.openstack.org/88736 | 21:41 |
| *** rcernin has quit IRC | 22:10 | |
| clarkb | clayg: progress now it fails on verifying the certificate. Looks like you can either tell swift to do insecure ssl (should be last resort) or we can somehow pass in the ca_cert path for swiftclient to trust that? | 22:12 |
| clayg | ok, didn't notice that when testing originally | 22:14 |
| clarkb | clayg: https://git.openstack.org/cgit/openstack-dev/devstack/tree/lib/swift#n466 swift3 alraedy seems to do that | 22:15 |
| clayg | ohai! real41m27.501s | 22:15 |
| clarkb | clayg: soshouldn't be too bad to get that into the test.conf? | 22:15 |
| * clayg shrugs | 22:15 | |
| *** mdrabe has quit IRC | 22:16 | |
| clarkb | clayg: actually I think I might be able to fix this in devstack. It has a step where it trusts the self made CA globally, but it only does that after starting services | 22:17 |
| clarkb | if I move it up to before service start the services should just trust the ca | 22:18 |
| clayg | maybe? i'm not sure where I see it failing on certs really... | 22:20 |
| clayg | with the devstack headers fix, and the auth_uri fix in swift my functests on my freshy-freshy devstack vm are doing pretty well... | 22:20 |
| clarkb | clayg: http://logs.openstack.org/74/372374/27/check/gate-swift-dsvm-functional-ubuntu-xenial-nv/360b584/console.html#_2017-06-01_21_44_52_774832 | 22:21 |
| clarkb | in your case probably beacuse you've run it multiple times so the CA is already trusted. I just pushed edit to https://review.openstack.org/469732 which I think may work | 22:22 |
| patchbot | patch 469732 - openstack-dev/devstack - Run devstack CA and cert setup early | 22:22 |
| clayg | ah, I was looking at results from the wrong patch set maybe | 22:23 |
| clayg | well - i bet this is going to go great - it's working great on my machine anyway :D | 22:23 |
| *** alenavolk has quit IRC | 22:26 | |
| *** mvk_ has joined #openstack-swift | 22:41 | |
| *** alenavolk has joined #openstack-swift | 22:42 | |
| *** alenavolk has quit IRC | 22:51 | |
| *** mvk_ has quit IRC | 22:52 | |
| clarkb | clayg: ok my initial thoguht was wrong. The problem there is we run the swift tests out of a virtualenv (via tox) so the global cert fixing doesn't affect it | 22:55 |
| clarkb | clayg: we will need to pass in the cafile path | 22:56 |
| clayg | oh neat - that probably explains why running the tests directly with nose wasn't hitting the issue for me - well spotted! | 22:56 |
| clarkb | clayg: in your change it is reading in the "real" swift config that devstack writes out, then emitting a new config for the functtests right? | 22:59 |
| clayg | I admit^Wemit nothing | 22:59 |
| clayg | clarkb: devstack wort out a config that had all the right bits of info - but in the wrong places - but it seemed sane to me the way devstack wrote it out - so I just fixed swift to expect that format and still work with the old | 23:00 |
| clayg | but i'm not aware if any plumbing exists for cacert into swiftclient currently | 23:00 |
| clayg | i'm kinda looking... | 23:00 |
| clarkb | clayg: right but is it reading test.conf or swift.conf? | 23:00 |
| clayg | oh - /etc/swift/test.conf yeah for sure | 23:01 |
| clarkb | clayg: get_auth_keystone() in swiftclient seems to accept a ca bundle to trust | 23:01 |
| clarkb | not sure if swift config plumbs that through though | 23:02 |
| timburke | clayg: and certainly the cli accepts a --os-cacert arg (or OS_CACERT env var) | 23:02 |
| * clayg shrugs | 23:02 | |
| clayg | yeah I was saying specifically dunno if functional/__init__.py has brains to take bits from .ini to swiftclient - less so that if swiftclient works and shit | 23:03 |
| timburke | service should pick up the env var for free... | 23:03 |
| clayg | timburke: are you sure that plumbing is for programatic access vs cli? maybe that'd be cool i guess... | 23:04 |
| clarkb | I will write the devstack side change to put it in test.conf | 23:04 |
| clarkb | and then let you sort out the swift side | 23:04 |
| clayg | could work | 23:04 |
| clarkb | clayg: something like https://review.openstack.org/470056 (I used cafile because that is what appears to be used elsewhere but not sure if that is an option already in swift with a possibly different name) | 23:09 |
| patchbot | patch 470056 - openstack-dev/devstack - Set the cafile in swift test.conf if tls enabled | 23:09 |
| clarkb | clayg: also left a comment on https://review.openstack.org/#/c/469984/2 I think a couple vars may be misnamed with the move to a function there | 23:11 |
| patchbot | patch 469984 - swift - Prefer auth_uri in functests configs | 23:11 |
| clayg | and oddly since it's a global it still works - thanks clarkb I'll try to clear that up | 23:12 |
| *** alenavolk has joined #openstack-swift | 23:13 | |
| *** alenavolk has quit IRC | 23:17 | |
| *** alenavolk has joined #openstack-swift | 23:17 | |
| *** alenavolk has quit IRC | 23:17 | |
| *** kei_yama has joined #openstack-swift | 23:21 | |
| *** chsc has quit IRC | 23:41 | |
| clayg | o | 23:50 |
| *** tonanhngo has quit IRC | 23:52 | |
| *** tonanhngo has joined #openstack-swift | 23:58 | |
| *** jamielennox|away is now known as jamielennox | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!