| *** zhill has joined #openstack-swift | 00:00 | |
| mattoliverau | well true, we could just use that :) | 00:00 |
|---|---|---|
| notmyname | ah. meeting time | 00:01 |
| *** haomaiwang has quit IRC | 00:01 | |
| mattoliverau | notmyname: tag team ;) | 00:02 |
| *** lcurtis has quit IRC | 00:02 | |
| *** haomaiwang has joined #openstack-swift | 00:03 | |
| openstackgerrit | Ben Martin proposed openstack/swift: Print min_part_hours lockout time remaining https://review.openstack.org/257577 | 00:03 |
| blmartin | And with that, good night | 00:05 |
| *** blmartin has quit IRC | 00:06 | |
| *** Zyric has quit IRC | 00:18 | |
| *** pgbridge has quit IRC | 00:30 | |
| kota_ | good morning | 00:45 |
| *** zhill has quit IRC | 00:50 | |
| notmyname | mattoliverau: back | 00:51 |
| notmyname | mattoliverau: did you find anything? | 00:52 |
| notmyname | looks like I did get an error on that test run | 00:53 |
| mattoliverau | still failed for me on the SAIO that was failing, and it's memory is 1/2 swapped out, rebooting and doing it again incase it was in a bad state to begin with. | 00:57 |
| mattoliverau | I wonder if something has upgraded and not cleaning itself up as well as it used to. | 00:57 |
| mattoliverau | I did notice in my getting it to work the other week that the .kill() to the proxy server in in_process wasn't killing it properly, so I have a hack to close the socket.. but maybe the proxy still taking up memory. | 00:57 |
| mattoliverau | actaully maybe it's not swapped out... I'm getting confused with all my open tabs :P | 00:57 |
| *** haomaiwang has quit IRC | 01:01 | |
| *** haomaiwang has joined #openstack-swift | 01:01 | |
| kota_ | looks like gates has an issue? | 01:02 |
| *** openstack has joined #openstack-swift | 01:08 | |
| *** haomaiwang has quit IRC | 01:10 | |
| *** dmorita has quit IRC | 01:12 | |
| *** gyee has quit IRC | 01:17 | |
| then3rd | notmyname: I saw your message from a couple days ago about upgrading. I forgot this was currently icehous. | 01:20 |
| notmyname | then3rd: have you done an upgrade on the cluster since you took it over? | 01:20 |
| then3rd | I have not. From what I've read it *should* be mostly straight forward, no? | 01:21 |
| notmyname | yeah, it is | 01:22 |
| notmyname | https://swiftstack.com/blog/2013/12/20/upgrade-openstack-swift-no-downtime/ | 01:22 |
| then3rd | Yep, looked at that earlier. | 01:22 |
| notmyname | then3rd: you're upgrading from a pretty old version (relatively speaking), so you'll need to take some extra care with checking for config updates | 01:23 |
| then3rd | You beat me to my next question. XD | 01:23 |
| notmyname | then3rd: but there's no reason to do every upgrade | 01:23 |
| notmyname | then3rd: and there's not any "upgrade here first, then go to the next version" | 01:24 |
| then3rd | So, it is feasible to go directly to liberty? | 01:24 |
| notmyname | then3rd: yes. but the further you upgrade, the more care you should take (in general) | 01:24 |
| notmyname | then3rd: eg with 2.4.0 there were some dependency and config changes called out https://github.com/openstack/swift/blob/master/CHANGELOG | 01:24 |
| notmyname | then3rd: and there have been various things moved to middleware (eg dlo and versioning) since the version you're running | 01:25 |
| notmyname | then3rd: and there's some great stuff too that you'll probably want to use, like the "Allow 1+ object-servers-per-disk deployment" (also in 2.4.0) | 01:26 |
| notmyname | then3rd: but the first goal is to upgrade without adding anything new. get that working, only then make it work better | 01:26 |
| *** dmorita has joined #openstack-swift | 01:27 | |
| notmyname | then3rd: if you didn't want to go in one big upgrade, then I'd recommend going to 2.0 first, then 2.5.0 | 01:27 |
| then3rd | Exactly. I'd like to get it upgraded without it getting fubar'd | 01:27 |
| notmyname | then3rd: of course :-) | 01:27 |
| notmyname | then3rd: I'd suggest doing in in a lab or VM first | 01:28 |
| then3rd | If I had that luxury.. another one of my goals. | 01:28 |
| notmyname | heh | 01:28 |
| notmyname | then3rd: are you on 1.11 or 1.13? | 01:29 |
| then3rd | yeah.. someone else built this, quit, and left zero notes. | 01:29 |
| then3rd | 1.13.1 | 01:29 |
| notmyname | ok. so 2.0 is easy and will "just work" | 01:29 |
| then3rd | From everyting I have seen, this is a pretty standard install. I believe he followed the openstack.org install guide, so configs will probably be copy-edit-paste. Ultimately, I'd like to deploye/manage this with Ansible. | 01:31 |
| notmyname | then3rd: then you should talk to mattoliverau or tdasilva about that. i think they both have ansible playbooks for swift | 01:31 |
| notmyname | then3rd: ok, I reviewed the changelog. nothing jumps out that would cause your existing configs to break with newer code | 01:32 |
| notmyname | then3rd: so upgrading to the new version should be relatively painless | 01:32 |
| notmyname | then3rd: and after than you can update the config file for new settings | 01:33 |
| then3rd | That's good news. | 01:33 |
| notmyname | then3rd: of course, standard disclaimer, some guy on the internet telling you it will work is no guarantee that it will. be sure to check it yourself ;-) | 01:33 |
| notmyname | then3rd: looks like the notes for 2.0 and 2.4.0 are probably the most important ones | 01:34 |
| notmyname | for upgrades | 01:34 |
| then3rd | Random guy on the internet is still more reputable than anything else I can come up with. ;-) | 01:34 |
| *** lyrrad has quit IRC | 01:34 | |
| notmyname | then3rd: and please ask about anythign you need more clarity on. measure twice, cut once, and all that | 01:34 |
| then3rd | Certainly. I really apreaciate the help. | 01:35 |
| notmyname | ok, time for me to bike home | 01:36 |
| then3rd | Be safe, ttyl! | 01:37 |
| *** rminmin has joined #openstack-swift | 01:37 | |
| *** garthb_ has quit IRC | 01:39 | |
| *** haomaiwang has joined #openstack-swift | 01:40 | |
| mattoliverau | then3rd: we (rackspace private cloud) deploy using ansible, so feel free to ask me questions too ;) | 01:40 |
| openstackgerrit | Merged openstack/swift: Merge branch 'master' into feature/crypto https://review.openstack.org/257416 | 01:42 |
| then3rd | mattoliverau: thanks! I think my trouble has been finding playbooks that are complete and easy to use, yet not overly complex. I'm still probably less than 100 hours hands-on with asnible, so it takes a while to review what exactly the roles are doing. | 01:43 |
| *** Zyric_ is now known as Zyric | 01:46 | |
| then3rd | Swiftacular looked promising but seems to revolve around vagrant. the official openstack-ansible project seems like its best supported but looks like a chore to get working. | 01:46 |
| then3rd | I think it was openstack-ansible I got working though. Only successfull with the all-in-one lxc build. | 01:48 |
| *** rminmin has quit IRC | 01:48 | |
| *** badari has quit IRC | 01:52 | |
| mattoliverau | then3rd: openstack-ansible is what we RPC use. I have my own simpler ansible setup I use for building test clusters, but it's definitely not feature full, infact it's only what I've getten around to writting as I need to test things :) | 01:52 |
| mattoliverau | then3rd: but if you want to follow the openstack-ansibe root, I can introduce you to the people who work on it day to day :) | 01:54 |
| *** badari has joined #openstack-swift | 01:54 | |
| then3rd | I've thought about doing something similar. I don't think we need to be able to turn every knob right now.. it could be easier to add in thigs as-needed. | 01:54 |
| mattoliverau | then3rd: where are you based? | 01:54 |
| then3rd | Utah, USA. | 01:54 |
| then3rd | Since I'm in a time crunch, I won't be able to work on stuff like automation until February. I'll make it a goal to automate the dev environment. Should lead to automating production. | 01:56 |
| mattoliverau | then3rd: good idea :) | 01:56 |
| mattoliverau | then3rd: if you want to go down the OSA (openstack ansible path) then in Feb I can introduce you to the people in the know ;) | 01:57 |
| then3rd | That is, if I don't find somewhere that actually pays their admins enough first. :D | 01:58 |
| mattoliverau | but we're always avalabile here to help in anyway we can | 01:58 |
| mattoliverau | lo | 01:58 |
| then3rd | Certainly! | 01:58 |
| mattoliverau | then3rd: *cough* rackspace *cough* :P | 01:58 |
| then3rd | I do have at least one friend in rackspace devops, San Antonio... I should see if he has the hookup. hehe | 02:00 |
| *** haomaiwang has quit IRC | 02:01 | |
| *** eikke has quit IRC | 02:02 | |
| mattoliverau | then3rd: cool, and if you want to continue doing openstack devops (OSA even) I can always pass something on or make initial intros ;) | 02:02 |
| *** kei_yama has quit IRC | 02:02 | |
| *** kei_yama has joined #openstack-swift | 02:03 | |
| mattoliverau | anyway I've had lunch so am heading home to debug exciting testr issues. bbs. | 02:03 |
| then3rd | I'm all for exploring my options.. this place struggles to really be truly innovative. | 02:04 |
| then3rd | And so far, I'm loving everything that openstack can do. Anyway, catch you later | 02:05 |
| *** rvasilets__ has quit IRC | 02:05 | |
| *** asettle has quit IRC | 02:11 | |
| *** dewsday has joined #openstack-swift | 02:11 | |
| *** dewsday is now known as asettle | 02:11 | |
| *** klrmn has quit IRC | 02:14 | |
| *** asettle has quit IRC | 02:14 | |
| *** asettle has joined #openstack-swift | 02:15 | |
| *** noark9 has joined #openstack-swift | 02:20 | |
| *** mtreinish has quit IRC | 02:20 | |
| *** mtreinish has joined #openstack-swift | 02:20 | |
| *** changbl has quit IRC | 02:21 | |
| *** alexandra has joined #openstack-swift | 02:26 | |
| *** haomaiwang has joined #openstack-swift | 02:29 | |
| *** asettle has quit IRC | 02:30 | |
| *** alexandra is now known as asettle | 02:30 | |
| ho | mattoliverau: quick question about doc/source/development_auth.rst. in the doc there is a word "Cross Origin RequestS". does it mean "Corss Origin Requests" or does it have special meaning? | 02:32 |
| notmyname | ho: it's for CORS, but it's supposed to be Cross-Origin Resource Sharing | 02:34 |
| ho | notmyname: my question is why last charactor is large S | 02:35 |
| notmyname | ho: right. it's a mistake | 02:35 |
| *** changbl has joined #openstack-swift | 02:35 | |
| ho | notmyname: i see. i would like to know there is a special meaning. thanks! | 02:35 |
| notmyname | ho: someone was trying to fit it to CORS | 02:35 |
| ho | notmyname: oh, i see. is it better to fix it or keep it?. | 02:37 |
| notmyname | fix it. it's wrong the way it is now | 02:37 |
| ho | notmyname: OK. i will submit it soon. | 02:39 |
| then3rd | notmyname: Dumb question.. I finally managed to locate the ubuntu-cloud repos for the different releases (http://ubuntu-cloud.archive.canonical.com/ubuntu/ trusty-updates/juno/main). Are there packages for other "minor" releases too? i.e: 2.0 mentioner earlier? (https://launchpad.net/swift/juno/2.0.0) | 02:39 |
| notmyname | then3rd: no. it seems that most ditros have only chosen to package (most of) the openstack major releases | 02:39 |
| notmyname | most deployers end up building their own packages to have some more control over it | 02:40 |
| notmyname | then3rd: what OS are you on/ | 02:40 |
| then3rd | ubuntu 14.04 | 02:41 |
| openstackgerrit | Hisashi Osanai proposed openstack/swift: Fix a typo in development_auth.rst https://review.openstack.org/257654 | 02:41 |
| notmyname | ho: could you make it say "Cross-Origin Resource Sharing (CORS)"? | 02:43 |
| ho | notmyname: OK i will change it. | 02:43 |
| notmyname | then3rd: hmm.. the place I was going to point you to hasn't been updated a while | 02:44 |
| notmyname | then3rd: do they have a 2.2.0? | 02:45 |
| notmyname | then3rd: here's some info that might be handy https://wiki.openstack.org/wiki/Swift/version_map | 02:45 |
| *** ianbrown has quit IRC | 02:46 | |
| *** ianbrown has joined #openstack-swift | 02:47 | |
| notmyname | then3rd: so if you wanted to take 2 upgrade steps, you might try 2.2.0 and 2.5.0 | 02:47 |
| then3rd | Yep. Only thing I'm concerned about may be keystone v3. I won't break my v2 auth, right? | 02:47 |
| notmyname | nope | 02:47 |
| then3rd | Cool. In the event I do break something, is there any sort of backup (simply make a copy of /etc/swift ?) I should be making first? If it doesn't work, should I expect any downgrade complications? | 02:49 |
| then3rd | notmyname: I know these are pretty generic questions, but they've been hard to find answers to. | 02:50 |
| notmyname | then3rd: from looking at the changelog earlier, and from memory, there aren't any on-disk format changes in the range you're looking at. so I don't anticipate any downgrade complications | 02:50 |
| notmyname | however, if you add new storage policies, you won't be able to get at those anymore if you downgrade | 02:50 |
| notmyname | hmm | 02:51 |
| openstackgerrit | Hisashi Osanai proposed openstack/swift: Fix a typo in development_auth.rst https://review.openstack.org/257654 | 02:51 |
| notmyname | there's a not in the 2.4.0 release about versioning and downgrading there | 02:51 |
| notmyname | ho: thanks | 02:52 |
| notmyname | then3rd: that's the kind of thing you'll have to watch out for on downgrades | 02:52 |
| notmyname | then3rd: we're generally very careful at keeping an upgrade path and letting you do that without any issue | 02:53 |
| then3rd | I see. The client would like to run whatevever is latest anyway, so downgrading is just a backout plan if I discover I'm completely incompetent. | 02:53 |
| then3rd | :P | 02:54 |
| *** sanchitmalhotra has joined #openstack-swift | 02:56 | |
| then3rd | I think I'll chew on this a little more tonight, put together a change control procedure and try this tomorrow. Again, thanks for the pointers. | 02:57 |
| *** haomaiwang has quit IRC | 03:01 | |
| *** haomaiwang has joined #openstack-swift | 03:01 | |
| janonymous | morning | 03:03 |
| mattoliverau | janonymous: morning | 03:05 |
| ho | janonymous: morning! | 03:05 |
| janonymous | Please find links for review if somebody can lend some time: https://review.openstack.org/#/c/208222/ , https://review.openstack.org/#/c/254276/ | 03:09 |
| janonymous | :) | 03:13 |
| openstackgerrit | Merged openstack/python-swiftclient: Delete python bytecode before every test run https://review.openstack.org/252210 | 03:25 |
| then3rd | notmyname: another question I'm not having much luck with. My object-server nodes don't seem to have any swift logs. It seems that log_requests defaults to true, but I don't have anything swifty in /var/log. There are no extra logging options set in any of my configs. | 03:26 |
| then3rd | oh hey. it goest to syslog aka /dev/log. mystery solved. | 03:28 |
| then3rd | s/goest/goes/ | 03:28 |
| *** klrmn has joined #openstack-swift | 03:30 | |
| *** pgbridge has joined #openstack-swift | 03:33 | |
| *** haomaiwang has quit IRC | 03:40 | |
| *** kei_yama_ has joined #openstack-swift | 03:44 | |
| *** kei_yama has quit IRC | 03:44 | |
| *** venkat has joined #openstack-swift | 03:45 | |
| *** ppai has joined #openstack-swift | 03:50 | |
| *** yatin has joined #openstack-swift | 03:50 | |
| *** david-lyle has quit IRC | 03:50 | |
| *** david-lyle has joined #openstack-swift | 03:58 | |
| *** david-lyle has quit IRC | 04:01 | |
| *** david-lyle has joined #openstack-swift | 04:02 | |
| *** km has quit IRC | 04:02 | |
| *** km has joined #openstack-swift | 04:03 | |
| *** links has joined #openstack-swift | 04:07 | |
| *** proteusguy_ has quit IRC | 04:18 | |
| *** haomaiwang has joined #openstack-swift | 04:23 | |
| *** openstackstatus has quit IRC | 04:24 | |
| *** openstack has joined #openstack-swift | 04:26 | |
| *** ChanServ sets mode: +v openstackstatus | 04:26 | |
| *** SkyRocknRoll has joined #openstack-swift | 04:29 | |
| *** proteusguy_ has joined #openstack-swift | 04:35 | |
| *** haomaiwang has quit IRC | 04:36 | |
| *** haomaiwang has joined #openstack-swift | 04:38 | |
| *** haomaiwang has quit IRC | 04:39 | |
| *** 17WABA6GG has joined #openstack-swift | 04:39 | |
| *** 17WABA6GG has quit IRC | 04:40 | |
| *** 21WAAHVK2 has joined #openstack-swift | 04:40 | |
| *** 21WAAHVK2 has quit IRC | 04:41 | |
| *** 14WAAGEZK has joined #openstack-swift | 04:41 | |
| *** 14WAAGEZK has quit IRC | 04:42 | |
| *** haomaiwa_ has joined #openstack-swift | 04:42 | |
| *** haomaiwa_ has quit IRC | 04:43 | |
| *** haomaiwang has joined #openstack-swift | 04:43 | |
| *** haomaiwang has quit IRC | 04:44 | |
| *** haomaiwang has joined #openstack-swift | 04:44 | |
| *** haomaiwang has quit IRC | 04:45 | |
| *** 21WAAHVNZ has joined #openstack-swift | 04:45 | |
| *** 21WAAHVNZ has quit IRC | 04:46 | |
| *** 14WAAGE11 has joined #openstack-swift | 04:46 | |
| *** 14WAAGE11 has quit IRC | 04:47 | |
| *** 21WAAHVO2 has joined #openstack-swift | 04:47 | |
| *** 21WAAHVO2 has quit IRC | 04:48 | |
| *** 21WAAHVPK has joined #openstack-swift | 04:48 | |
| *** 21WAAHVPK has quit IRC | 04:49 | |
| *** 14WAAGE3N has joined #openstack-swift | 04:49 | |
| *** 14WAAGE3N has quit IRC | 04:50 | |
| *** haomaiwa_ has joined #openstack-swift | 04:50 | |
| *** haomaiwa_ has quit IRC | 04:51 | |
| *** 18WABIL7Z has joined #openstack-swift | 04:51 | |
| *** 18WABIL7Z has quit IRC | 04:52 | |
| *** haomaiwang has joined #openstack-swift | 04:52 | |
| *** haomaiwang has quit IRC | 04:53 | |
| *** haomaiwa_ has joined #openstack-swift | 04:53 | |
| *** haomaiwa_ has quit IRC | 04:54 | |
| *** km has quit IRC | 05:31 | |
| *** km has joined #openstack-swift | 05:32 | |
| *** takashi has joined #openstack-swift | 05:55 | |
| *** pgbridge has quit IRC | 05:58 | |
| *** noark9 has quit IRC | 06:07 | |
| *** badari has quit IRC | 06:14 | |
| *** mac_ified has quit IRC | 06:19 | |
| *** Jeffrey4l has joined #openstack-swift | 06:19 | |
| *** dmorita has quit IRC | 06:19 | |
| janonymous | I am still getting trouble with tox -e py27 test.unit.container.test_reconciler:TestReconcilerUtils.test_parse_raw_obj , when i run the whole UT , i.e. tox -e py27 everythin is ok. | 06:29 |
| *** ppai has quit IRC | 06:42 | |
| *** ppai has joined #openstack-swift | 06:56 | |
| *** takashi_ has joined #openstack-swift | 07:17 | |
| *** takashi has quit IRC | 07:19 | |
| *** takashi_ has quit IRC | 07:19 | |
| *** takashi has joined #openstack-swift | 07:20 | |
| *** takashi_ has joined #openstack-swift | 07:20 | |
| *** zaitcev has quit IRC | 07:27 | |
| openstackgerrit | Shu Muto proposed openstack/python-swiftclient: Drop py33 support https://review.openstack.org/257707 | 07:38 |
| openstackgerrit | Christian Schwede proposed openstack/swift: Fix full_listing in internal_client https://review.openstack.org/257502 | 07:38 |
| *** ntt has joined #openstack-swift | 07:46 | |
| *** trifon has joined #openstack-swift | 07:48 | |
| *** akle has quit IRC | 07:54 | |
| *** wbhuber has quit IRC | 08:03 | |
| *** arnox has joined #openstack-swift | 08:06 | |
| *** rledisez has joined #openstack-swift | 08:08 | |
| *** klrmn has quit IRC | 08:15 | |
| *** hseipp has joined #openstack-swift | 08:28 | |
| *** admin0 has joined #openstack-swift | 08:29 | |
| *** rvasilets__ has joined #openstack-swift | 08:30 | |
| *** acoles_ is now known as acoles | 08:52 | |
| *** ho has quit IRC | 08:54 | |
| *** jistr has joined #openstack-swift | 08:56 | |
| acoles | blmartin: did you get an answer re viewkeys problem in unit tests - i saw same yesterday but conveniced myself i had too old six version. you need six>=1.9.0, maybe use tox -r. | 09:00 |
| acoles | kota_: thanks for reviewing the diskfile patch, i will respond to your comments asap | 09:08 |
| *** admin0 has quit IRC | 09:09 | |
| openstackgerrit | Merged openstack/swift: Deprecated tox -downloadcache option removed https://review.openstack.org/256597 | 09:10 |
| openstackgerrit | Merged openstack/python-swiftclient: Deprecated tox -downloadcache option removed https://review.openstack.org/256715 | 09:13 |
| *** admin0 has joined #openstack-swift | 09:20 | |
| *** acoles is now known as acoles_ | 09:31 | |
| *** janonymous has quit IRC | 09:33 | |
| *** takashi_ has quit IRC | 09:34 | |
| *** takashi has quit IRC | 09:34 | |
| *** openstackgerrit has quit IRC | 09:47 | |
| *** openstackgerrit has joined #openstack-swift | 09:48 | |
| *** dewsday has joined #openstack-swift | 09:58 | |
| *** jmccarthy has joined #openstack-swift | 09:59 | |
| *** aix has joined #openstack-swift | 10:01 | |
| *** venkat has quit IRC | 10:12 | |
| *** aix has quit IRC | 10:26 | |
| *** aix has joined #openstack-swift | 10:28 | |
| *** openstackgerrit has quit IRC | 10:32 | |
| *** openstackgerrit has joined #openstack-swift | 10:33 | |
| kota_ | acoles: great to hrear, I'm looking forward to it :-) | 10:47 |
| *** admin0 has quit IRC | 11:11 | |
| *** admin0 has joined #openstack-swift | 11:12 | |
| *** ppai has quit IRC | 11:14 | |
| *** haomaiwang has joined #openstack-swift | 11:19 | |
| *** haomaiwang has quit IRC | 11:20 | |
| *** ig0r_ has quit IRC | 11:20 | |
| *** haomaiwang has joined #openstack-swift | 11:20 | |
| *** haomaiwang has quit IRC | 11:21 | |
| *** haomaiwang has joined #openstack-swift | 11:22 | |
| *** haomaiwang has quit IRC | 11:23 | |
| *** haomaiwang has joined #openstack-swift | 11:23 | |
| *** haomaiwang has quit IRC | 11:24 | |
| *** 14WAAGM1V has joined #openstack-swift | 11:24 | |
| *** 14WAAGM1V has quit IRC | 11:25 | |
| *** haomaiwa_ has joined #openstack-swift | 11:25 | |
| *** haomaiwa_ has quit IRC | 11:26 | |
| *** ppai has joined #openstack-swift | 11:26 | |
| *** haomaiwang has joined #openstack-swift | 11:26 | |
| *** haomaiwang has quit IRC | 11:27 | |
| *** haomaiwang has joined #openstack-swift | 11:27 | |
| *** haomaiwang has quit IRC | 11:28 | |
| *** haomaiwang has joined #openstack-swift | 11:28 | |
| *** haomaiwang has quit IRC | 11:29 | |
| *** 14WAAGM54 has joined #openstack-swift | 11:29 | |
| *** 14WAAGM54 has quit IRC | 11:30 | |
| *** sanchitmalhotra has quit IRC | 11:31 | |
| *** haomaiwang has joined #openstack-swift | 11:32 | |
| *** haomaiwang has quit IRC | 11:33 | |
| *** 21WAAH3GU has joined #openstack-swift | 11:33 | |
| *** 21WAAH3GU has quit IRC | 11:34 | |
| *** haomaiwang has joined #openstack-swift | 11:34 | |
| *** haomaiwang has quit IRC | 11:35 | |
| *** haomaiwang has joined #openstack-swift | 11:35 | |
| *** haomaiwang has quit IRC | 11:36 | |
| *** haomaiwang has joined #openstack-swift | 11:36 | |
| *** haomaiwang has quit IRC | 11:37 | |
| *** haomaiwang has joined #openstack-swift | 11:37 | |
| *** haomaiwang has quit IRC | 11:38 | |
| *** haomaiwang has joined #openstack-swift | 11:38 | |
| *** haomaiwang has quit IRC | 11:39 | |
| *** haomaiwang has joined #openstack-swift | 11:39 | |
| *** haomaiwang has quit IRC | 11:40 | |
| *** haomaiwang has joined #openstack-swift | 11:40 | |
| *** haomaiwang has quit IRC | 11:41 | |
| *** haomaiwang has joined #openstack-swift | 11:41 | |
| *** haomaiwang has quit IRC | 11:42 | |
| *** 14WAAGNG5 has joined #openstack-swift | 11:42 | |
| *** 14WAAGNG5 has quit IRC | 11:43 | |
| *** haomaiwang has joined #openstack-swift | 11:46 | |
| *** haomaiwang has quit IRC | 11:47 | |
| *** haomaiwang has joined #openstack-swift | 11:47 | |
| *** haomaiwang has quit IRC | 11:48 | |
| *** 17WABBDO5 has joined #openstack-swift | 11:48 | |
| *** 17WABBDO5 has quit IRC | 11:49 | |
| *** haomaiwang has joined #openstack-swift | 11:49 | |
| *** haomaiwang has quit IRC | 11:50 | |
| *** haomaiwang has joined #openstack-swift | 11:50 | |
| *** haomaiwang has quit IRC | 11:51 | |
| *** haomaiwang has joined #openstack-swift | 11:51 | |
| *** haomaiwang has quit IRC | 11:52 | |
| *** haomaiwang has joined #openstack-swift | 11:52 | |
| *** haomaiwang has quit IRC | 11:53 | |
| *** 18WABIUKW has joined #openstack-swift | 11:54 | |
| *** 18WABIUKW has quit IRC | 11:55 | |
| *** haomaiwa_ has joined #openstack-swift | 11:55 | |
| *** mattoliverau has quit IRC | 11:55 | |
| *** haomaiwa_ has quit IRC | 11:56 | |
| *** yatin has quit IRC | 11:56 | |
| *** haomaiwa_ has joined #openstack-swift | 11:56 | |
| *** matt6434 has joined #openstack-swift | 11:56 | |
| *** haomaiwa_ has quit IRC | 11:57 | |
| *** haomaiwa_ has joined #openstack-swift | 11:57 | |
| *** haomaiwa_ has quit IRC | 11:58 | |
| *** haomaiwa_ has joined #openstack-swift | 11:58 | |
| *** Jeffrey4l has quit IRC | 11:58 | |
| *** haomaiwa_ has quit IRC | 11:59 | |
| *** 14WAAGNTQ has joined #openstack-swift | 12:00 | |
| *** 14WAAGNTQ has quit IRC | 12:01 | |
| *** 14WAAGNU9 has joined #openstack-swift | 12:01 | |
| *** 14WAAGNU9 has quit IRC | 12:02 | |
| *** haomaiwang has joined #openstack-swift | 12:02 | |
| *** haomaiwang has quit IRC | 12:03 | |
| *** haomaiwang has joined #openstack-swift | 12:03 | |
| *** haomaiwang has quit IRC | 12:04 | |
| *** acoles_ is now known as acoles | 12:04 | |
| acoles | kota_: are you still here? | 12:04 |
| *** rvasilets__ has quit IRC | 12:05 | |
| *** haomaiwang has joined #openstack-swift | 12:06 | |
| *** haomaiwang has quit IRC | 12:07 | |
| *** 18WABIUWU has joined #openstack-swift | 12:07 | |
| *** 18WABIUWU has quit IRC | 12:08 | |
| *** 18WABIUXV has joined #openstack-swift | 12:08 | |
| *** 18WABIUXV has quit IRC | 12:09 | |
| *** haomaiwang has joined #openstack-swift | 12:10 | |
| *** haomaiwang has quit IRC | 12:11 | |
| *** haomaiwang has joined #openstack-swift | 12:11 | |
| *** haomaiwang has quit IRC | 12:12 | |
| *** haomaiwang has joined #openstack-swift | 12:12 | |
| *** CaioBrentano has joined #openstack-swift | 12:12 | |
| *** haomaiwang has quit IRC | 12:13 | |
| *** haomaiwang has joined #openstack-swift | 12:15 | |
| *** haomaiwang has quit IRC | 12:16 | |
| *** haomaiwang has joined #openstack-swift | 12:17 | |
| *** haomaiwang has quit IRC | 12:18 | |
| *** dewsday has quit IRC | 12:18 | |
| *** haomaiwang has joined #openstack-swift | 12:18 | |
| *** haomaiwang has quit IRC | 12:19 | |
| *** ppai has quit IRC | 12:19 | |
| *** haomaiwang has joined #openstack-swift | 12:19 | |
| *** haomaiwang has quit IRC | 12:20 | |
| *** haomaiwang has joined #openstack-swift | 12:20 | |
| *** haomaiwang has quit IRC | 12:21 | |
| *** haomaiwang has joined #openstack-swift | 12:21 | |
| *** haomaiwang has quit IRC | 12:22 | |
| *** 17WABBED5 has joined #openstack-swift | 12:22 | |
| *** 17WABBED5 has quit IRC | 12:23 | |
| *** haomaiwang has joined #openstack-swift | 12:23 | |
| *** haomaiwang has quit IRC | 12:24 | |
| *** haomaiwang has joined #openstack-swift | 12:24 | |
| *** haomaiwang has quit IRC | 12:25 | |
| *** haomaiwang has joined #openstack-swift | 12:25 | |
| *** haomaiwang has quit IRC | 12:26 | |
| *** 17WABBEGT has joined #openstack-swift | 12:26 | |
| *** 17WABBEGT has quit IRC | 12:27 | |
| *** haomaiwa_ has joined #openstack-swift | 12:27 | |
| *** haomaiwa_ has quit IRC | 12:28 | |
| *** 14WAAGOIE has joined #openstack-swift | 12:28 | |
| *** 14WAAGOIE has quit IRC | 12:29 | |
| *** haomaiwang has joined #openstack-swift | 12:29 | |
| *** km has quit IRC | 12:29 | |
| *** haomaiwang has quit IRC | 12:30 | |
| *** haomaiwa_ has joined #openstack-swift | 12:32 | |
| *** haomaiwa_ has quit IRC | 12:33 | |
| *** ppai has joined #openstack-swift | 12:33 | |
| *** 18WABIVH4 has joined #openstack-swift | 12:34 | |
| *** 18WABIVH4 has quit IRC | 12:34 | |
| *** admin0 has quit IRC | 12:34 | |
| *** km has joined #openstack-swift | 12:34 | |
| *** haomaiwa_ has joined #openstack-swift | 12:35 | |
| *** haomaiwa_ has quit IRC | 12:36 | |
| *** haomaiwang has joined #openstack-swift | 12:37 | |
| *** haomaiwang has quit IRC | 12:38 | |
| *** haomaiwa_ has joined #openstack-swift | 12:38 | |
| *** haomaiwa_ has quit IRC | 12:39 | |
| *** haomaiwang has joined #openstack-swift | 12:40 | |
| *** haomaiwang has quit IRC | 12:41 | |
| *** 21WAAH43O has joined #openstack-swift | 12:41 | |
| *** 21WAAH43O has quit IRC | 12:42 | |
| *** haomaiwa_ has joined #openstack-swift | 12:42 | |
| *** haomaiwa_ has quit IRC | 12:43 | |
| *** haomaiwa_ has joined #openstack-swift | 12:43 | |
| *** haomaiwa_ has quit IRC | 12:44 | |
| *** haomaiwang has joined #openstack-swift | 12:44 | |
| *** haomaiwang has quit IRC | 12:45 | |
| *** SkyRocknRoll has quit IRC | 12:46 | |
| *** haomaiwa_ has joined #openstack-swift | 12:46 | |
| *** haomaiwa_ has quit IRC | 12:47 | |
| *** haomaiwang has joined #openstack-swift | 12:47 | |
| *** haomaiwang has quit IRC | 12:48 | |
| *** haomaiwang has joined #openstack-swift | 12:48 | |
| *** haomaiwang has quit IRC | 12:49 | |
| *** ppai has quit IRC | 12:49 | |
| *** admin0 has joined #openstack-swift | 12:52 | |
| *** km has quit IRC | 12:53 | |
| *** sayali_ has quit IRC | 12:56 | |
| *** sayali has joined #openstack-swift | 12:58 | |
| *** kei_yama_ has quit IRC | 13:01 | |
| openstackgerrit | Merged openstack/swift: Fix a typo in development_auth.rst https://review.openstack.org/257654 | 13:02 |
| *** ppai has joined #openstack-swift | 13:03 | |
| *** links has quit IRC | 13:09 | |
| *** sayali has quit IRC | 13:18 | |
| *** sayali has joined #openstack-swift | 13:20 | |
| *** petertr7_away has quit IRC | 13:23 | |
| *** petertr7_away has joined #openstack-swift | 13:24 | |
| *** petertr7_away is now known as petertr7 | 13:24 | |
| *** lpabon has joined #openstack-swift | 13:25 | |
| *** sayali has quit IRC | 13:26 | |
| *** ppai has quit IRC | 13:27 | |
| *** sayali has joined #openstack-swift | 13:28 | |
| *** sayali has quit IRC | 13:30 | |
| *** breitz has quit IRC | 13:31 | |
| *** breitz has joined #openstack-swift | 13:32 | |
| *** sayali has joined #openstack-swift | 13:32 | |
| kota_ | Back | 13:33 |
| *** sayali has quit IRC | 13:33 | |
| *** whydidyoustealmy has joined #openstack-swift | 13:35 | |
| *** sayali has joined #openstack-swift | 13:35 | |
| *** shakamunyi has joined #openstack-swift | 13:35 | |
| *** whydidyoustealmy is now known as superflyy | 13:36 | |
| *** petertr7 has quit IRC | 13:38 | |
| *** sayali has quit IRC | 13:40 | |
| *** sayali has joined #openstack-swift | 13:42 | |
| *** petertr7_away has joined #openstack-swift | 13:43 | |
| *** petertr7_away is now known as petertr7 | 13:43 | |
| *** janonymous has joined #openstack-swift | 13:48 | |
| *** links has joined #openstack-swift | 13:51 | |
| *** aix has quit IRC | 13:52 | |
| kota_ | acoles: what's up? | 13:57 |
| acoles | kota_: hi! i had a question about one of your comments , but I figured it out now :) about durable_timestamp. Actually your comment caused me to realise there was some tests missing. | 13:58 |
| acoles | kota_: thanks! | 13:58 |
| *** sayali has quit IRC | 13:59 | |
| kota_ | Great! | 13:59 |
| *** rcernin has joined #openstack-swift | 14:02 | |
| kota_ | Sorry, i had a party today for late response | 14:04 |
| *** aix has joined #openstack-swift | 14:05 | |
| acoles | kota_: enjoy the party :D | 14:10 |
| *** sayali has joined #openstack-swift | 14:11 | |
| kota_ | acoles: thanks! | 14:14 |
| *** sayali has quit IRC | 14:24 | |
| *** sayali has joined #openstack-swift | 14:27 | |
| *** sayali has quit IRC | 14:32 | |
| *** badari has joined #openstack-swift | 14:33 | |
| *** sayali has joined #openstack-swift | 14:36 | |
| *** Jeffrey4l has joined #openstack-swift | 14:41 | |
| *** dslevin_ has quit IRC | 14:42 | |
| *** dslevin has quit IRC | 14:42 | |
| *** hseipp has quit IRC | 14:44 | |
| *** sayali has quit IRC | 14:50 | |
| *** janonymous has quit IRC | 14:51 | |
| *** links has quit IRC | 14:58 | |
| *** links has joined #openstack-swift | 14:59 | |
| *** blmartin has joined #openstack-swift | 15:03 | |
| *** dslevin has joined #openstack-swift | 15:06 | |
| *** rvasilets__ has joined #openstack-swift | 15:08 | |
| *** aix has quit IRC | 15:09 | |
| *** aix has joined #openstack-swift | 15:09 | |
| *** noark9 has joined #openstack-swift | 15:18 | |
| *** joeljwright has joined #openstack-swift | 15:22 | |
| *** ChanServ sets mode: +v joeljwright | 15:22 | |
| *** janonymous has joined #openstack-swift | 15:24 | |
| *** tsg has joined #openstack-swift | 15:30 | |
| *** changbl has quit IRC | 15:35 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Don't ssync data when only a durable is missing https://review.openstack.org/232684 | 15:36 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Make ECDiskFile report all fragments found on disk https://review.openstack.org/231121 | 15:36 |
| *** wbhuber has joined #openstack-swift | 15:44 | |
| *** Guest56642 has joined #openstack-swift | 15:47 | |
| *** Guest56642 has left #openstack-swift | 15:47 | |
| *** wbhuber_ has joined #openstack-swift | 15:47 | |
| *** wbhuber has quit IRC | 15:48 | |
| *** wbhuber_ is now known as wbhuber | 15:48 | |
| *** admin0 has quit IRC | 15:53 | |
| *** rcernin has quit IRC | 15:58 | |
| *** Jeffrey4l has quit IRC | 16:00 | |
| *** dslevin has quit IRC | 16:06 | |
| *** diazjf has joined #openstack-swift | 16:07 | |
| *** eranrom has joined #openstack-swift | 16:07 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Minor cleanup of repeated identical test assertions https://review.openstack.org/257984 | 16:12 |
| blmartin | cschwede: Thanks for reviewing the patch, It looks like the error you were seeing is the following: rebalance is attempted, as min_part_hours has passed the ring is rebalanced but no partitions need to be reassigned. last_part_moves_epoch is updated and then after that we check if min_part_seconds_left is > 0 which it always will be as we just updated the moves_epoch. This is easily fixed by moving the check before attempting the rebalance (just l | 16:14 |
| *** pgbridge has joined #openstack-swift | 16:22 | |
| blmartin | acoles: Thanks for the answer about viewkeys. I ran that and it may have solved it (when I activate the venv I can import six and call viewkeys) I'll let you know later if tox plays nicely with it | 16:23 |
| acoles | blmartin: good luck! | 16:24 |
| *** garthb has joined #openstack-swift | 16:26 | |
| *** arnox has quit IRC | 16:33 | |
| *** eranrom has quit IRC | 16:39 | |
| *** arnox has joined #openstack-swift | 16:40 | |
| *** arnox has quit IRC | 16:40 | |
| *** rledisez1 has joined #openstack-swift | 16:52 | |
| *** rledisez1 has quit IRC | 16:53 | |
| *** zhill has joined #openstack-swift | 16:53 | |
| *** janonymous_ has joined #openstack-swift | 16:57 | |
| *** lcurtis has joined #openstack-swift | 16:59 | |
| notmyname | good morning | 16:59 |
| vinsh | cschwede: Hello, I want to talk about https://review.openstack.org/#/c/230914/ with you when you are around. | 17:00 |
| vinsh | cschwede: I like your idea behind the patch and I would like to rebase it for you if you don't mind. | 17:01 |
| vinsh | Its in conflict with some larger changes I made the last few weeks and I could bring it up to speed quickly. | 17:01 |
| *** rledisez has quit IRC | 17:02 | |
| *** david-lyle has quit IRC | 17:02 | |
| *** gyee has joined #openstack-swift | 17:03 | |
| *** yatin has joined #openstack-swift | 17:08 | |
| *** lyrrad has joined #openstack-swift | 17:08 | |
| *** yatin has quit IRC | 17:09 | |
| pdardeau | good morning! | 17:11 |
| diazjf | notmyname, I will be having a hangout session with the Barbican team tomorrow to discuss what is needed for Castellan in Swift. Would be cool if you could attend 8-) | 17:12 |
| notmyname | when? | 17:12 |
| *** david-lyle has joined #openstack-swift | 17:12 | |
| diazjf | notmyname, its most likely gonna be 2PM Central, but I can move it to 3 if you need. | 17:13 |
| notmyname | I can do that | 17:13 |
| diazjf | notmyname awesome. I'll post the hangout link tomorrow. | 17:14 |
| openstackgerrit | Stuart McLaren proposed openstack/python-swiftclient: Retry download of object body https://review.openstack.org/159208 | 17:16 |
| openstackgerrit | Stuart McLaren proposed openstack/python-swiftclient: Test 'string' behaviour of get_object https://review.openstack.org/258018 | 17:16 |
| openstackgerrit | Ben Martin proposed openstack/swift: Print min_part_hours lockout time remaining https://review.openstack.org/257577 | 17:16 |
| openstackgerrit | Shashirekha Gundur proposed openstack/swift: Update SAIO setup for Ubuntu 14.04 https://review.openstack.org/207790 | 17:22 |
| *** nadeem has joined #openstack-swift | 17:27 | |
| *** garthb_ has joined #openstack-swift | 17:27 | |
| *** dmorita has joined #openstack-swift | 17:27 | |
| *** dmorita has quit IRC | 17:28 | |
| *** nadeem has quit IRC | 17:29 | |
| *** garthb has quit IRC | 17:29 | |
| *** nadeem has joined #openstack-swift | 17:29 | |
| *** dmorita has joined #openstack-swift | 17:29 | |
| *** petertr7 is now known as petertr7_away | 17:30 | |
| *** noark9 has quit IRC | 17:31 | |
| *** acoles is now known as acoles_ | 17:38 | |
| hurricanerix | yo | 17:41 |
| hurricanerix | err, morning =) | 17:41 |
| *** blmartin has quit IRC | 17:47 | |
| *** blmartin has joined #openstack-swift | 17:48 | |
| *** links has quit IRC | 17:50 | |
| *** klrmn has joined #openstack-swift | 17:55 | |
| notmyname | hurricanerix: so what are we going to do about the testr patch? | 17:59 |
| *** garthb_ has quit IRC | 18:00 | |
| *** garthb_ has joined #openstack-swift | 18:00 | |
| notmyname | hurricanerix: with the in-process tests, I'm seeing the same as mattoliverau. tons of memory usage | 18:01 |
| hurricanerix | notmyname i will take a look at it again today. | 18:01 |
| notmyname | hurricanerix: in general, it seems to work ok for me when doing `restswift && startmain && tox -efunc` | 18:01 |
| notmyname | but the in-process tests don't work | 18:02 |
| *** joeljwright has quit IRC | 18:03 | |
| notmyname | gmmaha: were you looking at the testr stuff? | 18:08 |
| *** diazjf has quit IRC | 18:17 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 18:18 |
| *** tsg has quit IRC | 18:20 | |
| *** blmartin has quit IRC | 18:20 | |
| *** jistr has quit IRC | 18:22 | |
| openstackgerrit | Tim Burke proposed openstack/python-swiftclient: Move python-keystoneclient to "extras" https://review.openstack.org/254346 | 18:35 |
| * jrichli and others who work with her will be at holiday festivities the rest of today | 18:44 | |
| gmmaha | notmyname: sorry wasnt at desk. yes, i was looking at it | 18:47 |
| gmmaha | if i run the patch 10 (testr) with --isolated i am able to get them all to pass with no memory burn | 18:47 |
| patchbot | gmmaha: https://review.openstack.org/#/c/10/ | 18:47 |
| gmmaha | but it takes an hour to run on my machine :( | 18:47 |
| gmmaha | or we have to option of snubbing the logger and we should be fine | 18:48 |
| gmmaha | and tell folks that if you want logs, you can run per test or if you want the big one we need some above 4G | 18:49 |
| notmyname | wow | 18:49 |
| gmmaha | spoke to mtreinish about it and there isnt a nice way of snubbing out to the subunit stream out of ostestr into a file even as its needed for ostestr to assess the test results | 18:49 |
| notmyname | seems rather terrible to go from a 2 minute test run to a 60 minute test run | 18:50 |
| gmmaha | yeah | 18:50 |
| gmmaha | there is one another thing, not sure if thats ok | 18:50 |
| gmmaha | the debug_logger, i can make it write to a tmp file | 18:50 |
| gmmaha | and then throw it out if all tests are good and leave it there if its not! | 18:50 |
| gmmaha | that might reduce the trash on the stdout which floods the subunit stream | 18:51 |
| notmyname | right. that seems reasonable I think | 18:51 |
| notmyname | but the test run time is a really big deal | 18:51 |
| notmyname | any ideas why it takes so much longer to run? | 18:52 |
| gmmaha | if i remove the --isolated, we should be back | 18:52 |
| gmmaha | the --isolated will setup and teardown for each test, reducing the load on testr and keeping the mem usage down. | 18:52 |
| gmmaha | but the overhead of a setup and teardown is the time increase to an hour | 18:52 |
| notmyname | ah | 18:53 |
| notmyname | isn't there a 40 minute timer in the gate? after that it gets killed? | 18:53 |
| gmmaha | ohhh i didnt know about that :D | 18:54 |
| gmmaha | but i have a feeling that the gate machines are all running something over 4G which is why it passes for jenkins fine | 18:54 |
| gmmaha | while our tests all fail | 18:54 |
| notmyname | gmmaha: I'd really like to see the result of the debug_logger change you're proposing | 18:55 |
| notmyname | is that somethign that will be in the test wrapper or in swift itself? | 18:55 |
| gmmaha | notmyname: right now i am thinking about swift.. | 18:55 |
| notmyname | well, both are "in swift" wrt the repo. I mean somethign int he logging in swift | 18:55 |
| gmmaha | the call to make the FakeLogger which write now writes to stdour with print(...) | 18:55 |
| gmmaha | its actually test/unit/__init__.py | 18:55 |
| gmmaha | aah i see your point. How do we make sure we can check the test result before blowing away the file :( | 18:59 |
| notmyname | ...right...that was my point...totally what I meant...of course | 18:59 |
| notmyname | ;-) | 18:59 |
| gmmaha | haha | 18:59 |
| *** aix has quit IRC | 18:59 | |
| gmmaha | can i just write a same named file in the swift repo itself | 19:00 |
| gmmaha | ? | 19:00 |
| gmmaha | and asdd it to .gitignore? | 19:00 |
| * gmmaha feels yucky saying that | 19:00 | |
| notmyname | yeah, that doesn't sound nice | 19:00 |
| notmyname | gmmaha: but you may have found something interesting. if the log output can be modified in the test harness, that could be something cleaner to implement | 19:00 |
| *** eranrom has joined #openstack-swift | 19:01 | |
| gmmaha | yeah | 19:01 |
| gmmaha | do you think we can take my debug log snubbing patch now and file a follow on bug for a cleanup? | 19:01 |
| gmmaha | not sure how pressed on time we are on this | 19:01 |
| *** tsg has joined #openstack-swift | 19:02 | |
| notmyname | gmmaha: you're looking in unit/__init__.py? but we're looking at functional tests | 19:02 |
| gmmaha | the functional test __init__.py calls the debug_logger from test/unit/__init__.py to create the fake logger | 19:03 |
| gmmaha | that basically just does a print statement | 19:03 |
| * gmmaha goes to double check that claim | 19:03 | |
| notmyname | of course it does ;-) | 19:03 |
| notmyname | yup | 19:04 |
| notmyname | line 618 | 19:04 |
| gmmaha | yeah | 19:04 |
| notmyname | what about an envvar that redirects debug_logger? or a different setup of the in-process stuff that doesn't use debug_logger? | 19:05 |
| notmyname | hmm..why do we need debug logger for in-process stuff anyway? | 19:06 |
| gmmaha | thats what i tried with this patch 255457 | 19:06 |
| patchbot | gmmaha: https://review.openstack.org/#/c/255457/ - Option to supress logs for in_process test | 19:06 |
| gmmaha | basically env variable andno logs.. no env variable and i unleash the kracken | 19:07 |
| notmyname | ah, right | 19:09 |
| notmyname | and i like that, but I'm wondering if there's another option besides stdout or nothing | 19:10 |
| gmmaha | aaah | 19:10 |
| notmyname | like maybe doing a "real" logger to syslog? why not? | 19:10 |
| gmmaha | yeah my patch is more of the band-aid than a real fix | 19:10 |
| notmyname | but I like where you're going with it | 19:10 |
| gmmaha | i thought the reason we didnt do real log was since its all in_process, leave no trace | 19:10 |
| notmyname | but askign that question makes me wonder why we need a stdout logger at all. | 19:10 |
| gmmaha | :D | 19:10 |
| pdardeau | notmyname: on the topic on long-running tests, it appears to take quite a while to create 60000 devices in ring? | 19:11 |
| notmyname | pdardeau: it's more of a function of the part power you're using, but yeah. math is hard. go buy better CPUs ;-) | 19:11 |
| tdasilva | notmyname, gmmaha: so, in the case of tox, where did the in-process logging go? | 19:11 |
| notmyname | pdardeau: wow. that sounded a whole lot more snarky that I meant it to sound. ;-) | 19:12 |
| gmmaha | tdasilva: its iwth tox, it goes to stdout | 19:12 |
| gmmaha | with the new switch to ostestr/testr | 19:12 |
| tdasilva | was it being suppressed by tox? | 19:12 |
| gmmaha | with nosetests( i am not expert) it seems to keep it within itself happily | 19:12 |
| ahale | i was thinking that too pdardeau : it would be nice if ring-builder took ranges so you could easily add a bunch | 19:12 |
| gmmaha | tdasilva: no sir. tox isnt doing any supressions.. | 19:12 |
| tdasilva | gmmaha: I mean, we didn't have the problem of the logging being dumped on stdout, right? | 19:12 |
| gmmaha | i am running all the tests with ttox | 19:12 |
| pdardeau | notmyname: no worries! | 19:12 |
| ahale | or an input file or something | 19:12 |
| gmmaha | tdasilva: we have been dumping to stdout all this while.. dont think that code changes.. guessing its how nosetests vs testr/ostestr handle the stdout data | 19:13 |
| tdasilva | sorry, i mean nosetests | 19:13 |
| notmyname | tdasilva: gmmaha: or is the reality that they were dumped to stdout and nose suppressed that therefore they were just unused? | 19:13 |
| notmyname | tdasilva: gmmaha: if that's the case, then turning them off is effectively the same thing,r ight? | 19:13 |
| tdasilva | notmyname: that's what i was trying to get ot | 19:13 |
| tdasilva | sounds like it | 19:14 |
| gmmaha | :D | 19:14 |
| pdardeau | ahale: i might try doing that. i'll be running these a lot, and at current rate it'll take forever | 19:14 |
| ahale | but then theres quite a few things wrong with ring-builder | 19:14 |
| gmmaha | notmyname: tdasilva: with an option to bring it out or without an option? | 19:14 |
| tdasilva | altought, if there's an option to do proper logging, then i'd prefer that | 19:14 |
| notmyname | pdardeau: what is it that's slow? adding devices or rebalancing? | 19:14 |
| ahale | yah, one of my test rings took a day to build :( | 19:14 |
| notmyname | ahale: get a better CPU! ;-) | 19:15 |
| pdardeau | pdardeau: adding devices | 19:15 |
| notmyname | pdardeau: ah, ok | 19:15 |
| ahale | notmyname: use more cpu cores and i might ! | 19:15 |
| notmyname | pdardeau: I was thinking the rebalance operation | 19:15 |
| notmyname | ahale: rewrite it in go! | 19:15 |
| ahale | i might just ! | 19:15 |
| tdasilva | honestly, if what we have today is effectively no logging, then I wouldn't mind for now changing it to use 'None' instead of the DebugLogger | 19:15 |
| notmyname | tdasilva: yes! | 19:15 |
| notmyname | ahale: no not really. don't do that ;-) | 19:15 |
| ahale | theres a few other things wrong with ring builder actually, main one is it will happily make rings it knows it will traceback when rebalancing | 19:16 |
| tdasilva | emphasis on 'for now'. ideally we would have proper logging...eventually ;) | 19:16 |
| * ahale needs to make bug report | 19:16 | |
| notmyname | tdasilva: which we still have if you aren't doing in-process | 19:16 |
| gmmaha | notmyname: tdasilva: so completely pull out the logging and then add a bug to follow-on | 19:16 |
| notmyname | ahale: pdardeau: I seem to remember something about batch adds to the ring | 19:16 |
| tdasilva | gmmaha: sounds reasonable | 19:16 |
| gmmaha | i am ok with that, though one part of me wants to have the flag in case there is someone more crazy than me | 19:17 |
| notmyname | gmmaha: yeah | 19:17 |
| notmyname | gmmaha: actually, just rolling your patch into the testr patch is probably a good idea | 19:17 |
| tdasilva | so, off by default, on with flag? | 19:17 |
| pdardeau | notmyname: so something better than builder.add_dev in loop? | 19:17 |
| gmmaha | tdasilva: yeah thats what i was thinking | 19:18 |
| ahale | also the holes remove makes in rings are annoying, they effecively reduce the max number of devices you can have :( | 19:18 |
| gmmaha | notmyname: okie. let me merge the patches then | 19:18 |
| notmyname | pdardeau: yeah, trying to find it | 19:18 |
| gmmaha | sending anew one up in a bit | 19:18 |
| tdasilva | gmmaha: cool | 19:18 |
| notmyname | ahale: have you used the latest version that jsut landed? any better there? | 19:18 |
| ahale | so i have a ring with 30k devices in it that i cant rebalance cos theres >64k devs | 19:18 |
| notmyname | ahale: right. that's what pdardeau is fixing right? :-) | 19:18 |
| ahale | nah havent tried that - the slowness of adding stuff to ring builder is so bad :) | 19:18 |
| ahale | i dunno, maybe? | 19:19 |
| pdardeau | notmyname: yes, that's what i'm currently working on | 19:19 |
| ahale | oh cool :) | 19:19 |
| tdasilva | gmmaha: just for reference, with the current patch, running against SAIO took 5 min. for me, running in-process took 20 minutes | 19:20 |
| tdasilva | notmyname: ^ | 19:20 |
| ahale | pdardeau: which bits are you thinking about fixing ? just out of interest - as theres been a few different things i've thought about | 19:22 |
| notmyname | tdasilva: yeah, that's not good | 19:22 |
| pdardeau | ahale: first up is to support more than 65535 devices (currently working on this) | 19:23 |
| notmyname | pdardeau: I'm nto sure what I was rememebering about adding a bunch of things to the ring at once. what about adding a device is taking a long time? | 19:23 |
| notmyname | pdardeau: I still hope you're looking at a dynamic number of bits there instead of a larger static number :-) | 19:23 |
| pdardeau | ahale: and then allow increase of part power | 19:24 |
| ahale | oh cool, I need that too :) | 19:24 |
| pdardeau | notmyname: yeah, i'm just trying to get a test going up to the current limit right now | 19:24 |
| notmyname | :-) | 19:24 |
| notmyname | pdardeau: just get ahale to send you his ring ;-) | 19:24 |
| ahale | notmyname: its slow because it only adds one at a time and loads up all the structures each time, while things like set weight operate on searches all at once, i guess | 19:25 |
| pdardeau | ahale: yeah, it's a memory management exercise for a single core ;-) | 19:25 |
| notmyname | yeah " max(d['id'] for d in self.devs if d)" could be a little slow ;-) | 19:25 |
| notmyname | hmm..there are 2 loops over self.devs to add a device | 19:26 |
| pdardeau | notmyname: it's all the list growing that happens 1 at a time | 19:27 |
| notmyname | gmmaha: cool. looking forward to it :-) | 19:31 |
| notmyname | pdardeau: yeah. :-( | 19:31 |
| ahale | pdardeau: if you need a big ring, i have a test one with 65476 devices. though its made with oldish ring code and has a lot of null in it | 19:31 |
| *** openstackgerrit has quit IRC | 19:32 | |
| *** openstackgerrit has joined #openstack-swift | 19:33 | |
| pdardeau | ahale: i just made one with 60000 devices, laptop is still glowing | 19:33 |
| openstackgerrit | Ganesh Maharaj Mahalingam proposed openstack/swift: Modify functional tests to use testr https://review.openstack.org/214206 | 19:33 |
| gmmaha | notmyname: hurricanerix: tdasilva: ^^ | 19:33 |
| gmmaha | can you give the patch and go and see how you feel about it | 19:33 |
| ahale | lol | 19:33 |
| hurricanerix | gmmaha sure | 19:34 |
| gmmaha | thanks a lot | 19:34 |
| * gmmaha just realized he is an idiot. :) | 19:36 | |
| notmyname | peluse: now that there is a new patch set, can you drop your -2 from patch 214206? no more danger of it landing from your +A misclick | 19:37 |
| patchbot | notmyname: https://review.openstack.org/#/c/214206/ - Modify functional tests to use testr | 19:37 |
| gmmaha | notmyname: just messed it up a bit.. fixing and send a new one up in a few | 19:37 |
| notmyname | ok | 19:37 |
| notmyname | gmmaha: I'm about to eat lunch, and then there's a company all-hands, so I probably won't look at it for a couple of hours | 19:38 |
| gmmaha | notmyname: no worries. and no rush. now that we have a solution | 19:38 |
| gmmaha | :) | 19:38 |
| notmyname | we hope :-) | 19:38 |
| openstackgerrit | Ganesh Maharaj Mahalingam proposed openstack/swift: Modify functional tests to use testr https://review.openstack.org/214206 | 19:38 |
| gmmaha | :D | 19:39 |
| gmmaha | pushed it up.. the envvar name changed between .testr.conf and the init.py file | 19:39 |
| notmyname | gmmaha: ok. can you leave a comment explaining the changes from patch set 12? | 19:39 |
| gmmaha | notmyname: sure can | 19:40 |
| notmyname | thanks | 19:40 |
| notmyname | dfg: thanks for the SLO ratelimit review | 19:41 |
| tdasilva | gmmaha: cool, in-process down to 3 min. | 19:42 |
| notmyname | yay | 19:42 |
| gmmaha | tdasilva: awesome :D | 19:43 |
| *** changbl has joined #openstack-swift | 19:44 | |
| gmmaha | notmyname: added my comments.. please let me know if you think thats good.. but one last thing is that we havent addressed acoles_ comment | 19:46 |
| gmmaha | how do we run a single test? :( | 19:46 |
| jlvillal | gmmaha, If using os-testr I thought it was: tox -efunctional -- -r "SOME_REGEX" | 19:49 |
| pdardeau | notmyname: ahale: looks like _set_parts_wanted is the guilty party | 19:50 |
| gmmaha | jlvillal: yeah but the --regex is defaulted in the script.. trying to see if i can put it a logic to look for -n/--no-discover to remove the regex | 19:50 |
| jlvillal | gmmaha, huh? Why is it in script? | 19:50 |
| * jlvillal hasn't read the patch... | 19:50 | |
| gmmaha | jlvillal: casue by default we only want the funcational tests to run.. without that, it will try and run all the tests | 19:51 |
| jlvillal | gmmaha, Aren't the functional tests in their own directory? | 19:51 |
| gmmaha | under test/* yes | 19:51 |
| jlvillal | gmmaha, Not test/func/* ? | 19:51 |
| jlvillal | and test/unit/* | 19:52 |
| jlvillal | :) | 19:52 |
| dfg | notmyname: np | 19:52 |
| gmmaha | yes thats how they are laid out | 19:52 |
| jlvillal | Then should set the directory to run the tests | 19:52 |
| jlvillal | Let me find how ironic does it | 19:52 |
| mtreinish | gmmaha, jlvillal: so that's why tempest started setting custom discovery paths with an env var | 19:53 |
| gmmaha | thanks.. that might help | 19:53 |
| mtreinish | that pattern has been cargo culted to a lot of different projects | 19:53 |
| jlvillal | gmmaha, https://github.com/openstack/ironic/blob/master/.testr.conf#L2 At the end of the line TESTS_DIR | 19:53 |
| mtreinish | gmmaha: look at the.testr.conf in tempest or novaclient for an example | 19:53 |
| mtreinish | or that link from jlvillal :) | 19:53 |
| jlvillal | gmmaha, https://github.com/openstack/ironic/blob/master/tox.ini#L15 | 19:54 |
| gmmaha | mtreinish: thanks. checking the link, it still that the discover. I am trying to see if i can make it mutually exclusive with --no-discover | 19:54 |
| *** gyee has quit IRC | 19:54 | |
| mtreinish | gmmaha: the no discover flag doesn't use testr | 19:55 |
| mtreinish | thats a shortcut to run a specific test | 19:55 |
| mtreinish | it's the same as invoking subunit.run manually | 19:55 |
| gmmaha | mtreinish: aah gotcha.. sorry was confusing the tests.. thanks. will fix the patch | 19:55 |
| jlvillal | gmmaha, We also run 'ostestr' from tox.ini | 19:56 |
| jlvillal | gmmaha, https://github.com/openstack/ironic/blob/master/tox.ini#L17 | 19:56 |
| openstackgerrit | Ondřej Nový proposed openstack/swift: Show only directories not files in recon device list https://review.openstack.org/258131 | 20:10 |
| *** tsg has quit IRC | 20:26 | |
| *** zaitcev has joined #openstack-swift | 20:26 | |
| *** ChanServ sets mode: +v zaitcev | 20:26 | |
| pdardeau | notmyname: ahale: just took a quick stab at implementing a batch mechanism in builder.py | 20:36 |
| pdardeau | speeds things up dramatically | 20:36 |
| ahale | cool :) | 20:37 |
| *** changbl has quit IRC | 20:37 | |
| pdardeau | ahale: can now create 65535 devices in under 1 second | 20:39 |
| ahale | oh wow ! thats amazingly faster | 20:39 |
| pdardeau | the good thing is it's a very simple change | 20:39 |
| zaitcev | what's the partition power | 20:41 |
| pdardeau | 10 | 20:42 |
| *** changbl has joined #openstack-swift | 20:44 | |
| *** dslevin has joined #openstack-swift | 20:47 | |
| *** matt6434 is now known as mattoliverau | 20:48 | |
| *** ChanServ sets mode: +v mattoliverau | 20:48 | |
| mattoliverau | Morning | 20:48 |
| openstackgerrit | Ganesh Maharaj Mahalingam proposed openstack/swift: Modify functional tests to use testr https://review.openstack.org/214206 | 20:50 |
| gmmaha | good morning mattoliverau | 20:50 |
| *** changbl has quit IRC | 20:53 | |
| *** tsg has joined #openstack-swift | 20:56 | |
| *** gyee has joined #openstack-swift | 21:01 | |
| *** nadeem has quit IRC | 21:13 | |
| *** nadeem has joined #openstack-swift | 21:13 | |
| gmmaha | mattoliverau: morning. Any chance you could check the latest patch and see if that works for you? | 21:15 |
| gmmaha | also i removed the regex search. that will help with a single line test | 21:15 |
| *** dmorita has quit IRC | 21:16 | |
| mattoliverau | gmmaha: just about to look at it :) nice work on finding fixes :) | 21:16 |
| gmmaha | mattoliverau: thanks.. :) more of a band-aid.. | 21:16 |
| gmmaha | tdasilva: mattoliverau: notmyname: the bug that i create to make logging less painful, should that be a wishlist entry or a bug? | 21:17 |
| tdasilva | gmmaha: i'm ok with wishlist | 21:18 |
| gmmaha | tdasilva: thanks. | 21:18 |
| gmmaha | was leaning towards that as well. | 21:19 |
| onovy | hey, swift is cool http://postimg.org/image/nn16ge657/ 1200 req/sec just now :) | 21:21 |
| *** admin0 has joined #openstack-swift | 21:22 | |
| vinsh | Anyone that might know..... :) : I have a customer, they uploaded a large (6gig) object to swift. This worked fine.. the client broke it into segments. This customer then set a referer/key on that container containing that object. | 21:24 |
| vinsh | that container is private | 21:24 |
| vinsh | They can not GET this object if the container_segments container is set to private | 21:25 |
| vinsh | if they set container_segments to public. it all works | 21:25 |
| vinsh | I wonder, does the segments container require a similar/specific acl? | 21:25 |
| gmmaha | tdasilva: i created it.. but apparently i dont have the rights to set it to wishlilst (importance) :/ | 21:25 |
| gmmaha | https://bugs.launchpad.net/swift/+bug/1526538 | 21:25 |
| openstack | Launchpad bug 1526538 in OpenStack Object Storage (swift) "Enhancements to debug logger" [Undecided,New] | 21:25 |
| tdasilva | gmmaha: done | 21:28 |
| gmmaha | tdasilva: thanks. | 21:28 |
| tdasilva | gmmaha: no no, thank you! ;) | 21:29 |
| gmmaha | ;) | 21:29 |
| mattoliverau | vinsh: yeah, if you give access to the container holding an *LO then you also need to give them access to the container holding the segments, otherwise the sub requests to get the segments will fail. | 21:29 |
| mattoliverau | vinsh: also if your using a DLO then container listing is important so they may need the rlisting set as well | 21:30 |
| vinsh | Ah.. good idea. I'll test now. | 21:32 |
| vinsh | with the acl set on the segments container.. a wget yields "Username/Password Authentication Failed." | 21:33 |
| vinsh | without the acl it just responds with a 401 | 21:33 |
| vinsh | I guess those are both 401 :) | 21:33 |
| *** dmorita has joined #openstack-swift | 21:34 | |
| tdasilva | gmmaha: in master, i'm able to run a specific test like this: tox -e func -- test.functional.tests:TestContainer.testDelete , but with the testr patch it fails, is there a new way to run specific tests? | 21:39 |
| vinsh | mattoliverau: I set .r:<key>.example.com,.rlistings on both the container and the segements container. That what you were suggesting? | 21:40 |
| *** ntt has quit IRC | 21:41 | |
| vinsh | ah.. setting the segments container to .r:*,.rlistings works though | 21:42 |
| vinsh | setting the same domain/key as the main container doesn't though | 21:42 |
| vinsh | hrm | 21:42 |
| mtreinish | tdasilva: try 'tox -efunc -- --no-discover test.functional.tests:TestContainer.testDelete' | 21:43 |
| gmmaha | tdasilva: can you try tox -e func -- -r test.functional.tests.TestContainer.testDelete$ | 21:43 |
| mattoliverau | gmmaha: Nice work, it works for me, nice work finding the logs being the cause of the memory usage. | 21:43 |
| mattoliverau | gmmaha: left a comment regarding default settings in .testr.conf. | 21:43 |
| openstackgerrit | Ondřej Nový proposed openstack/swift: Keystone middleware deprecated option is_admin removed https://review.openstack.org/258158 | 21:43 |
| onovy | it's cleanup time :) | 21:44 |
| gmmaha | mattoliverau: checking the patch now.. seems there is another comment as well. :D | 21:44 |
| *** ntt has joined #openstack-swift | 21:45 | |
| mtreinish | mattoliverau: for running a single test --no-discover/-n will be a bit better because it skips the discovery phase, it just directly calls subunit.run on the path you give it | 21:47 |
| gmmaha | mattoliverau: got loads of help from you guys and others :) | 21:47 |
| gmmaha | mattoliverau: for the default in the testr.conf, would you prefer it be test/functional since we know that it works today | 21:49 |
| gmmaha | and then cahnge it to unit on a lter date? | 21:49 |
| gmmaha | later* | 21:49 |
| mattoliverau | yup, and since that's what the change does | 21:49 |
| mattoliverau | it's and easy change/patch for later. | 21:49 |
| *** ntt has quit IRC | 21:49 | |
| gmmaha | mtreinish: the -n was causing an expcetion when we ran it that way | 21:49 |
| mtreinish | gmmaha: did you file a bug? | 21:51 |
| gmmaha | mtreinish: not yet. | 21:51 |
| mtreinish | do you have a pastebin of the exception, because that's how it's supposed to work. You give it a python object path with tests and it runs it | 21:51 |
| gmmaha | mtreinish: i can get you one in a few secs | 21:52 |
| openstackgerrit | Ganesh Maharaj Mahalingam proposed openstack/swift: Modify functional tests to use ostestr/testr https://review.openstack.org/214206 | 21:52 |
| openstackgerrit | Ganesh Maharaj Mahalingam proposed openstack/swift: Modify functional tests to use ostestr/testr https://review.openstack.org/214206 | 21:53 |
| gmmaha | mtreinish: http://paste.openstack.org/show/482017/ | 21:53 |
| tdasilva | gmmaha, mtreinish: so the closest I could get is this: tox -efunc -- --no-discover test.functional.tests.TestContainer | 21:53 |
| notmyname | onovy: nice! | 21:53 |
| gmmaha | tdasilva: didnt -- -r test.funcation.tests.TestContainer$ work for you? | 21:53 |
| * gmmaha goes to test it out | 21:53 | |
| jlvillal | mattoliverau, I wasn't sure how someone would use the default test/unit path. Since unit tests haven't been setup to use testr | 21:54 |
| tdasilva | gmmaha: one sec, let me try that again | 21:54 |
| mtreinish | gmmaha: the difference is -r does a regex match against discovery. So testr runs the discover command from your .testr.conf builds a list of tests and regex matches the arg against that list and then runs all the tests in that list | 21:55 |
| mtreinish | while -n litterally runs subunit.run $path | 21:55 |
| mtreinish | so the regex is a bit more forgiving to user error | 21:55 |
| tdasilva | gmmaha: yeah, that also works, I thought the $ was a typo, but I guess it is also important | 21:55 |
| *** petertr7_away is now known as petertr7 | 21:56 | |
| gmmaha | mtreinish: does -n let me run the exact test itself. i think tdasilva confirmed and i am able to run the class test as well | 21:56 |
| gmmaha | tox -efunc -v -- -n test.functional.tests.TestContainer works but tox -efunc -v -- -n test.functional.tests.TestContainer.testDelete fails | 21:56 |
| tdasilva | gmmaha: naahh...nevermind, the $ is not needed | 21:57 |
| mtreinish | gmmaha: there is a typo in you command, looking at the code it should be -n test.functional.test_container.TestContainer | 21:57 |
| gmmaha | tdasilva: aaah :) | 21:57 |
| *** david-lyle has quit IRC | 21:58 | |
| mtreinish | gmmaha: it takes a module, a class, or a method (or a test suite if you're manually constructing one anywhere) | 21:58 |
| tdasilva | gmmaha: oops, i guess it's late for me and I'm not paying attention....the $ does make a difference. with it, only one test runs testDelete, without $ 3 tests that begin with testDelete runs, so it's really a regex | 21:59 |
| *** david-lyle has joined #openstack-swift | 22:00 | |
| gmmaha | tdasilva: ok.. thanks for confirming | 22:01 |
| jlvillal | mtreinish, There is a 'test.py' file that has TestContainer | 22:01 |
| gmmaha | mtreinish: there is a test_container file and also a tests.py which has a class TestContainer | 22:01 |
| gmmaha | but you confirmed that it has to be a class/method/module to be called with -n | 22:01 |
| jlvillal | But isn't a method a class method? | 22:02 |
| gmmaha | notmyname: tdasilva: have a feeling all this has to go into one of our docs :) | 22:02 |
| tdasilva | gmmaha: now, i'm going to test with pdb | 22:02 |
| jlvillal | tdasilva, pdb++ ;) | 22:02 |
| gmmaha | tdasilva: awesome.. thanks a lot.. did not think about that at all.. :D | 22:02 |
| *** tsg has quit IRC | 22:02 | |
| jlvillal | tdasilva, I mean pudb ... | 22:03 |
| mtreinish | gmmaha: it's calling subunit.run without doing any discovery. (hence the name) You have to pass it a python module path to an object that it can run | 22:03 |
| mtreinish | just like the unittest cli interface | 22:03 |
| tdasilva | jlvillal: haven't used pudb yet :( | 22:04 |
| jlvillal | tdasilva, https://pypi.python.org/pypi/pudb | 22:04 |
| mattoliverau | jlvillal: pudb +1 | 22:04 |
| jlvillal | mtreinish, So why can it run: test.functional.tests.TestContainer but not test.functional.tests.TestContainer.testDelete | 22:05 |
| jlvillal | Is it because ...testDelete is not an object? | 22:05 |
| mattoliverau | jlvillal: .test.conf is run by testr by deafult, so by running: ostestr --serial it will use the default path used in the .test.conf. And we have haven't or will not make any effort to make sure the unit test pass in testr, at least not yet. | 22:05 |
| jlvillal | mattoliverau, Okay. /me wonders who would do that.... | 22:06 |
| jlvillal | :) | 22:06 |
| mattoliverau | anyone looking at the testr config and go cool, unit tests must work | 22:06 |
| mattoliverau | so anyone who like to explore.. ie. most engineers :P | 22:07 |
| jlvillal | mattoliverau, Okay. I'm not sure how likely that is. But already has gotten changed in .testr.conf. | 22:08 |
| *** ianbrown has quit IRC | 22:08 | |
| jlvillal | mattoliverau, So I think all is good :) | 22:08 |
| gmmaha | jlvillal: i think there is a WIP patch to do that. patch 257053 | 22:09 |
| patchbot | gmmaha: https://review.openstack.org/#/c/257053/ - [WIP] testr to run unittests | 22:09 |
| *** ianbrown has joined #openstack-swift | 22:09 | |
| mattoliverau | jlvillal: it doesn't matter how likely it is, the change was to get functaional tests to use testr, so it needs to be complete. But yeah, it's been changed so it doesn't matter anymore :) | 22:09 |
| mattoliverau | well it needs to be complete for me to put my +2 on it anyway :P | 22:10 |
| jlvillal | :) | 22:10 |
| tdasilva | jlvillal: so is pudb is just a better version of pdb or is the only way that will work with testr? | 22:12 |
| tdasilva | pdb is not working for me :/ | 22:12 |
| *** asettle is now known as asettle-afk | 22:13 | |
| mattoliverau | tdasilva: it's a ncurses interface so easier to use and looks better. | 22:15 |
| mattoliverau | and you can still use: import pudb .. pudb.set_trace() | 22:15 |
| *** rcernin has joined #openstack-swift | 22:16 | |
| tdasilva | mattoliverau: mm..currently, neither one is working...let me paste errors somewhere... | 22:17 |
| mattoliverau | gmmaha: made one last comment regarding .testr.conf.. sorry my bad, I notice the path was incorrect. .testr.conf is in the swift dir, so from there the functional tests live under test/functional not swift/test/functional | 22:20 |
| jlvillal | tdasilva, Just a better pdb I believe. | 22:21 |
| jlvillal | tdasilva, To use pdb have to pass in a certain flag. mtreinish Knows it off the top of his head :) | 22:22 |
| * jlvillal has to google | 22:22 | |
| mtreinish | jlvillal: --pdb or --no-discover | 22:22 |
| mtreinish | they do the same thing under the covers (they didn't use to which is why there are 2 flags) | 22:22 |
| mattoliverau | mtreinish: good to know ;) | 22:23 |
| jlvillal | mtreinish, Thanks. tdasilva ^^^ | 22:23 |
| *** trifon has quit IRC | 22:23 | |
| jlvillal | tdasilva, tox -efunctional -- --pdb | 22:23 |
| tdasilva | mtreinish, jlvillal: thanks, trying that now | 22:23 |
| jlvillal | or something like that | 22:23 |
| mtreinish | jlvillal: you need to give it an object path too | 22:24 |
| mtreinish | it doesn't know what to run otherwise | 22:24 |
| mtreinish | just like --no-discover | 22:24 |
| gmmaha | mattoliverau: i can fix that for the patch. let me check pdb as well and upload the final patch | 22:25 |
| mtreinish | iirc, the issue is testr starts the actual runners in subprocesses which interferes with pdb catching breakpoints or whatever it's doing (I don't really use pdb so it's all secondhand) | 22:25 |
| *** barker has joined #openstack-swift | 22:25 | |
| tdasilva | so, pdb sort of worked for me, but I couldn't provide the test name, only the class name (same error as before), so I had to run it like this: tox -e func -- --pdb test.functional.tests.TestContainer | 22:27 |
| *** asettle-afk is now known as asettle | 22:27 | |
| tdasilva | otherwise I get the error: TypeError: unbound method testDelete() must be called with TestContainer instance as first argument (got nothing instead) | 22:27 |
| *** admin0 has quit IRC | 22:28 | |
| gmmaha | tdasilva: ok. same happening here | 22:29 |
| openstackgerrit | Ganesh Maharaj Mahalingam proposed openstack/swift: Modify functional tests to use ostestr/testr https://review.openstack.org/214206 | 22:31 |
| mattoliverau | I just used the old 'python -m testtools.run discover --load-list my-list' like mentioned in https://wiki.openstack.org/wiki/Testr to degug testr tests. And you can replace pdb with pudb if you so desire. But if it works from ostestr that's even better. | 22:36 |
| gmmaha | mattoliverau: i think what tdasilva did worked for me but not with pudb. the display was restricted with 1/4 of the scren on the left leaving it not so easily usable | 22:37 |
| gmmaha | i remember reading something about that in one of the wiki pages | 22:38 |
| mattoliverau | ahh, though I guess patches are welcome :P | 22:38 |
| openstackgerrit | Paul Dardeau proposed openstack/swift: Add batch mode feature to RingBuilder https://review.openstack.org/258172 | 22:40 |
| vinsh | mattoliverau: I'm stumped. Just want to verify what I am doing makes sense. I applied the same acl of ".r:<key>.example.com,.rlistings" that was on the main container.. to the matching segments container. However we are still hitting a 401 trying to get that object in that container. ONLY when we set the segments container to ".r:*,.rlistings" can we get the object... the problem then is that the segments container is p | 22:40 |
| vinsh | ublic. | 22:40 |
| *** barker has quit IRC | 22:53 | |
| *** barker has joined #openstack-swift | 22:53 | |
| *** garthb__ has joined #openstack-swift | 22:54 | |
| *** garthb_ has quit IRC | 22:55 | |
| *** hurricanerix has quit IRC | 22:56 | |
| *** hurricanerix has joined #openstack-swift | 22:56 | |
| pdardeau | my latest patch is hosed, please ignore | 22:59 |
| then3rd | notmyname: This isn't normal, is it? It looks like all my storage nodes are doing it: http://pastebin.com/raw.php?i=qvDx3uwu | 22:59 |
| *** dustins is now known as dustins|out | 22:59 | |
| *** petertr7 is now known as petertr7_away | 22:59 | |
| *** david-lyle has quit IRC | 23:01 | |
| *** garthb__ has quit IRC | 23:01 | |
| *** david-lyle has joined #openstack-swift | 23:01 | |
| then3rd | heh, looks like one of the nodes has bad permissions on the object storage directory and hasn't been syncing. Fun. | 23:03 |
| *** barker has quit IRC | 23:07 | |
| *** barker has joined #openstack-swift | 23:08 | |
| *** km has joined #openstack-swift | 23:10 | |
| notmyname | then3rd: I was in a meeting. glad you found the issue | 23:14 |
| notmyname | then3rd: a good heuristic when seeing errors in swift after a deployment is first check permissions, then check drive health | 23:15 |
| *** eranrom has quit IRC | 23:16 | |
| *** eranrom has joined #openstack-swift | 23:17 | |
| *** rcernin has quit IRC | 23:20 | |
| *** barker_ has joined #openstack-swift | 23:20 | |
| notmyname | then3rd: it's probably a good idea to raise the max connections in rsync. looks like you're erroring with just 2 connections. probably want to raise that somewhat to 10 or 20 | 23:21 |
| notmyname | then3rd: depends a little on your cluster size and the activity in it | 23:21 |
| then3rd | No worries! Yeah, I corrected perms and restarted storage/background processes. Took a couple minutes to sync, but erros went away on all nodes. It's magic. | 23:21 |
| *** barker has quit IRC | 23:22 | |
| gmmaha | tdasilva: thanks for the comment on the patch :) | 23:23 |
| *** barker_ has quit IRC | 23:24 | |
| then3rd | notmyname: Right now it's only 5 storage nodes with four 3TB raids. We have four more nodes with 4tb arrays in another reigon ready to go. I have no idea how large it'll end up... 2 connections did seem a little limiting. | 23:25 |
| pdardeau | notmyname: ahale: my earlier posts regarding the slowness of adding devices was wrong! | 23:25 |
| pdardeau | my environment was jacked up - i was picking up OLD code | 23:26 |
| pdardeau | the slowness was fixed by clayg patches on Dec 7 | 23:26 |
| notmyname | cool | 23:26 |
| pdardeau | i'm an idiot! | 23:26 |
| pdardeau | but it was slow once upon a time | 23:26 |
| notmyname | I believe it! | 23:27 |
| *** kei_yama has joined #openstack-swift | 23:30 | |
| notmyname | pdardeau: do you still want the batch mode patch? | 23:31 |
| pdardeau | notmyname: no, it's not needed | 23:31 |
| notmyname | ack. would you mind abandoning it, then? | 23:32 |
| *** nadeem has quit IRC | 23:33 | |
| *** david-lyle_ has joined #openstack-swift | 23:33 | |
| pdardeau | notmyname: sure will. gotta figure out how to do that in gerrit :) | 23:33 |
| *** nadeem has joined #openstack-swift | 23:33 | |
| notmyname | there's a button | 23:33 |
| pdardeau | got it, thx | 23:34 |
| *** david-lyle has quit IRC | 23:36 | |
| notmyname | gmmaha: tdasilva: I step out to a meeting and I see comments in here, comments in gerrit, and two new patch sets | 23:41 |
| *** nexusz99 has joined #openstack-swift | 23:43 | |
| *** nexusz99 has quit IRC | 23:43 | |
| gmmaha | notmyname: :) the latest patch seems to be good and tdasilva has a nice comment on how to run single tests and pdb with it | 23:44 |
| notmyname | good | 23:44 |
| notmyname | I'll take a look (and I'd love to have mattoliverau's comment too) | 23:45 |
| gmmaha | notmyname: awesome. thanks | 23:45 |
| mattoliverau | Just moving from hacking place number 1, brb | 23:45 |
| gmmaha | just one thing is pudb doesnt work. Not sure if its an issue with how testr is. i do remeber reading something about it. I will check on that later | 23:45 |
| *** ho has joined #openstack-swift | 23:50 | |
| *** chlong has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!