| *** mikecmpbll has joined #openstack-swift | 00:05 | |
| *** itlinux has joined #openstack-swift | 00:10 | |
| *** kei-ichi has joined #openstack-swift | 00:30 | |
| *** gyee has quit IRC | 00:51 | |
| *** mikecmpbll has quit IRC | 02:12 | |
| *** mikecmpbll has joined #openstack-swift | 02:16 | |
| *** openstackgerrit has joined #openstack-swift | 02:20 | |
| openstackgerrit | Cyril Roelandt proposed openstack/swift master: Fix test/unit/common/test_container_sync_realms.py https://review.openstack.org/633644 | 02:20 |
|---|---|---|
| openstackgerrit | Cyril Roelandt proposed openstack/swift master: Python3: Fix test/unit/common/test_container_sync_realms.py https://review.openstack.org/633644 | 02:20 |
| openstackgerrit | Cyril Roelandt proposed openstack/swift master: Python3: fix test/unit/obj/test_replicator.py https://review.openstack.org/633645 | 02:28 |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: Change how O_TMPFILE support is detected https://review.openstack.org/633646 | 02:35 |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: Change how O_TMPFILE support is detected https://review.openstack.org/633646 | 02:46 |
| *** psachin has joined #openstack-swift | 02:52 | |
| *** gkadam has joined #openstack-swift | 03:11 | |
| *** itlinux has quit IRC | 04:41 | |
| *** spsurya has joined #openstack-swift | 05:05 | |
| *** ccamacho has quit IRC | 06:44 | |
| *** mrjk has joined #openstack-swift | 07:05 | |
| *** mrjk has quit IRC | 07:05 | |
| *** mrjk has joined #openstack-swift | 07:06 | |
| *** mrjk has quit IRC | 07:07 | |
| *** mikecmpbll has quit IRC | 07:18 | |
| *** ccamacho has joined #openstack-swift | 07:20 | |
| *** mrjk has joined #openstack-swift | 07:40 | |
| *** mrjk has quit IRC | 07:40 | |
| *** mrjk has joined #openstack-swift | 07:40 | |
| *** e0ne has joined #openstack-swift | 07:59 | |
| openstackgerrit | Simeon Gourlin proposed openstack/swift master: Fix decryption for broken objects https://review.openstack.org/633671 | 08:21 |
| openstackgerrit | Simeon Gourlin proposed openstack/swift master: Fix decryption for broken objects https://review.openstack.org/633671 | 08:22 |
| zigo | timburke: We filed the bug: https://bugs.launchpad.net/swift/+bug/1813725 | 08:28 |
| openstack | Launchpad bug 1813725 in OpenStack Object Storage (swift) "Object encryption key truncated sometimes when used with Barbican" [Undecided,New] | 08:28 |
| *** tkajinam has quit IRC | 08:48 | |
| *** pcaruana has joined #openstack-swift | 08:51 | |
| *** mikecmpbll has joined #openstack-swift | 09:03 | |
| *** mvkr has joined #openstack-swift | 09:33 | |
| *** mvkr has quit IRC | 10:46 | |
| *** mvkr has joined #openstack-swift | 11:03 | |
| *** nguyenhai_ has quit IRC | 11:47 | |
| *** nguyenhai has joined #openstack-swift | 11:59 | |
| *** kukacz has quit IRC | 12:04 | |
| *** kukacz has joined #openstack-swift | 12:04 | |
| *** pcaruana has quit IRC | 12:40 | |
| *** pcaruana has joined #openstack-swift | 12:50 | |
| *** e0ne has quit IRC | 13:13 | |
| *** gkadam has quit IRC | 13:20 | |
| *** e0ne has joined #openstack-swift | 13:29 | |
| *** psachin has quit IRC | 13:56 | |
| *** pcaruana has quit IRC | 14:45 | |
| *** e0ne has quit IRC | 14:51 | |
| *** pcaruana has joined #openstack-swift | 14:53 | |
| *** e0ne has joined #openstack-swift | 14:54 | |
| openstackgerrit | Pete Zaitcev proposed openstack/swift master: Python3: Fix test/unit/common/test_container_sync_realms.py https://review.openstack.org/633644 | 15:05 |
| clayg | tdasilva: so you're saying that you think the old check was broken - or at least in the mixed filesystem case the try and fallback code was being hit | 15:24 |
| clayg | tdasilva: I still think the best thing to do is make _get_tempfile just try O_TMPFILE, catch the error and cache .fallback_mkstemp | 15:24 |
| clayg | if one fifle-system sucks they can all suck | 15:25 |
| tdasilva | clayg: i don't think it was broken, but it was going based on the fact that we only support three fs (ext4, xfs and brtfs) and since brtfs was the last one to add support (3.16) the code was just checking for that kernel version | 15:28 |
| clayg | so on a 3.15 system for example, maybe you *could* have used o_tmpfile on your xfs filesystems and not brtfs - ok... so... | 15:29 |
| tdasilva | right, or even in the case of rhel/centos systems....i *think* they only backport the support to their 3.10 version for a specific fs (i.e., xfs) | 15:30 |
| tdasilva | fyi..i'm saying the backports happen individually, not that they haven't backported other fs. Also I only checked XFS | 15:31 |
| clayg | Can you clarify if your conclusion is advocating "we need to have more complex code to support the mixed-filesystem-support-case" or if... you're just like telling me FYI? | 15:35 |
| zaitcev | It didn't look more complex to me. | 15:38 |
| clayg | full disclosure, I'm advocating every call to _get_tempfile should just try O_TMPFILE first and if it fails cache a per-process (or per manager instance at least) flag that says "well I guess don't bother with that syscall next time and go straight to fallback_mkstemp mode" | 15:38 |
| tdasilva | clayg: i'm ok with that approach (the only small piece i'm uncertain is "do we cache per disk or not" ? | 15:39 |
| tdasilva | I think in the way you are describing, it is not bad to cache per device | 15:39 |
| clayg | I definitely think we do not need to cache per disk - we never have before and no one complained - we're making up problems that people don't have | 15:40 |
| clayg | I think if it weren't for the kernel check we'd probably already be doing the right thing | 15:40 |
| clayg | the fallback code looks solid | 15:40 |
| clayg | but on old systems I guess we get the double syscall ever PUT - so... so cache fallback mode... I guess | 15:41 |
| clayg | it's not like it's going to get better w/o restarting the process | 15:41 |
| tdasilva | right, the wasted syscall on old systems was what i was trying to avoid | 15:41 |
| tdasilva | let's go with caching one result only (not per device) and if it becomes we can cache per disk later | 15:42 |
| tdasilva | "if it becomes an issue" | 15:42 |
| *** ccamacho has quit IRC | 15:45 | |
| zaitcev | but muh BSD on rPI | 15:45 |
| *** ccamacho has joined #openstack-swift | 15:46 | |
| clayg | lol | 15:46 |
| clayg | zaitcev: I mean... if you run mixed file-system you just get the LCD behavior mkstemp - that's probably acceptable isn't it? | 15:47 |
| clayg | timburke: uhhh... do you have any idea on p 633671 - the bug report... how the?! | 15:47 |
| patchbot | https://review.openstack.org/#/c/633671/ - swift - Fix decryption for broken objects - 2 patch sets | 15:47 |
| zaitcev | clayg: I agree that we don't care about mixed-filesystem thing. | 15:47 |
| *** openstackgerrit has quit IRC | 15:51 | |
| tdasilva | gotta run, will send new patchset | 15:54 |
| *** e0ne has quit IRC | 16:31 | |
| *** pcaruana has quit IRC | 17:01 | |
| *** gyee has joined #openstack-swift | 17:03 | |
| *** ccamacho has quit IRC | 17:18 | |
| *** mikecmpbll has quit IRC | 17:32 | |
| *** NM has joined #openstack-swift | 17:36 | |
| *** e0ne has joined #openstack-swift | 17:46 | |
| *** NM has quit IRC | 17:50 | |
| *** NM has joined #openstack-swift | 17:51 | |
| *** NM has quit IRC | 17:52 | |
| *** NM has joined #openstack-swift | 17:52 | |
| *** NM has quit IRC | 18:03 | |
| *** NM has joined #openstack-swift | 18:06 | |
| *** openstackgerrit has joined #openstack-swift | 18:07 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Rebuild frags for unmounted disks https://review.openstack.org/629056 | 18:07 |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Quiet down a unittest https://review.openstack.org/633793 | 18:08 |
| clayg | yay! 🤞 | 18:08 |
| *** mikecmpbll has joined #openstack-swift | 18:20 | |
| *** e0ne has quit IRC | 18:53 | |
| *** mvkr has quit IRC | 19:06 | |
| *** e0ne has joined #openstack-swift | 19:07 | |
| zaitcev | timburke: I need your attention at this: https://review.openstack.org/#/c/570320/ | 20:09 |
| patchbot | patch 570320 - swift - py3: port the container sharder - 12 patch sets | 20:09 |
| zaitcev | timburke: It started as a mundane fix-up for sharder CLI, but it grew and turned into a real headache. It's a dependency for anything in the container, even the backend. Although, they are actually interlinked. Also, it affects py2, so needs extra care. | 20:11 |
| timburke | zaitcev: i'll see what i can do | 20:11 |
| zaitcev | I think I would like Matt have a look, but he's frolicking at LCA | 20:11 |
| *** e0ne has quit IRC | 20:42 | |
| *** NM has quit IRC | 20:54 | |
| *** NM has joined #openstack-swift | 21:00 | |
| *** NM has quit IRC | 21:04 | |
| *** gyee has quit IRC | 22:35 | |
| *** tkajinam has joined #openstack-swift | 22:55 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: Change how O_TMPFILE support is detected https://review.openstack.org/633646 | 22:56 |
| -openstackstatus- NOTICE: http://zuul.openstack.org is not working. https://zuul.openstack.org does work. Please use that while we investigate. | 23:14 | |
| *** rcernin has quit IRC | 23:53 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!