| *** gabor_antal has quit IRC | 00:04 | |
| *** gabor_antal has joined #openstack-swift | 00:04 | |
| *** tqtran has joined #openstack-swift | 00:09 | |
| *** tqtran has quit IRC | 00:13 | |
| *** gabor_antal_ has joined #openstack-swift | 00:14 | |
| *** gabor_antal has quit IRC | 00:18 | |
| kota_ | good morning | 01:08 |
|---|---|---|
| mattoliverau | kota_: morning | 01:22 |
| kota_ | mattoliverau: hi :-D | 01:22 |
| *** davideagle has quit IRC | 01:37 | |
| *** vint_bra has joined #openstack-swift | 01:51 | |
| *** Jeffrey4l has joined #openstack-swift | 02:11 | |
| *** vint_bra has quit IRC | 02:29 | |
| *** links has joined #openstack-swift | 03:48 | |
| *** ppai has joined #openstack-swift | 03:54 | |
| mahatic_ | good morning | 04:00 |
| *** psachin has joined #openstack-swift | 04:09 | |
| *** tqtran has joined #openstack-swift | 04:10 | |
| *** tqtran has quit IRC | 04:15 | |
| mattoliverau | mahatic_: morning | 04:21 |
| *** SkyRocknRoll has joined #openstack-swift | 05:03 | |
| *** siva_krish has joined #openstack-swift | 06:01 | |
| openstackgerrit | Kazuhiro MIYAHARA proposed openstack/swift: Remove 'X-Static-Large-Object' from .meta files https://review.openstack.org/385412 | 06:16 |
| *** SkyRocknRoll has quit IRC | 06:23 | |
| *** SkyRocknRoll has joined #openstack-swift | 06:40 | |
| *** dmorita has joined #openstack-swift | 06:59 | |
| *** sams-gleb has joined #openstack-swift | 07:04 | |
| *** rcernin has joined #openstack-swift | 07:05 | |
| openstackgerrit | Pavel Kvasnička proposed openstack/swift: WIP: Fix missing and corrupted suffixes during replication https://review.openstack.org/402324 | 07:17 |
| *** rcernin has quit IRC | 07:26 | |
| *** rcernin has joined #openstack-swift | 07:28 | |
| *** hseipp has joined #openstack-swift | 07:40 | |
| *** dmorita has quit IRC | 07:47 | |
| *** pavel has joined #openstack-swift | 07:50 | |
| *** pavel is now known as PavelK | 07:51 | |
| *** rcernin has quit IRC | 07:51 | |
| *** silor has joined #openstack-swift | 07:51 | |
| PavelK | cschwede: here? | 07:52 |
| *** silor1 has joined #openstack-swift | 07:54 | |
| *** rcernin has joined #openstack-swift | 07:54 | |
| PavelK | cschwede: I'd like to fix missing hashes in pkl and sometimes hard-recalculate all hashes. Do you working on patch 402204 now? | 07:55 |
| patchbot | https://review.openstack.org/#/c/402204/ - swift - Add missing suffixes to hashes.pkl during replication | 07:55 |
| *** silor has quit IRC | 07:55 | |
| *** silor1 is now known as silor | 07:55 | |
| *** Jeffrey4l has quit IRC | 07:57 | |
| *** silor has quit IRC | 08:02 | |
| *** siva_krish has quit IRC | 08:03 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Confirm receipt of SLO PUT with etag https://review.openstack.org/390901 | 08:05 |
| *** silor has joined #openstack-swift | 08:11 | |
| *** amoralej|off is now known as amoralej | 08:11 | |
| cschwede | PavelK: Hi! I’m working on making the listdir more deterministic right now | 08:11 |
| *** tqtran has joined #openstack-swift | 08:12 | |
| *** silor1 has joined #openstack-swift | 08:14 | |
| *** silor has quit IRC | 08:15 | |
| *** silor1 is now known as silor | 08:15 | |
| PavelK | cschwede: I offer more deterministic listing in patch 402324 - see swift/obj/replicator.py file without renaming do_list parameter | 08:17 |
| patchbot | https://review.openstack.org/#/c/402324/ - swift - WIP: Fix missing and corrupted suffixes during rep... | 08:17 |
| *** tqtran has quit IRC | 08:17 | |
| cschwede | PavelK: i already started working on deterministc listings yesterday, told that to to onovy? | 08:18 |
| cschwede | PavelK: also, looking at that patch it also has some random shuffling, and that’s likely not enough? | 08:19 |
| *** pcaruana has joined #openstack-swift | 08:19 | |
| PavelK | cschwede: I'm waiting for path from you; please, ignore my commit - it's a draft | 08:20 |
| *** geaaru has joined #openstack-swift | 08:20 | |
| onovy | PavelK: cschwede: I think we need to sync what to do :) i have 2 ideas | 08:21 |
| onovy | do_list_dir only checks MISSING suffixies | 08:21 |
| onovy | and it's called every 10th run - non-deterministic | 08:21 |
| onovy | make it deterministic, which is my first idea | 08:21 |
| onovy | cschwede: you are working on this, right? | 08:21 |
| cschwede | onovy: yes | 08:21 |
| onovy | cool | 08:21 |
| onovy | second idea: | 08:21 |
| onovy | add new param, check_all | 08:21 |
| onovy | run it every 100th run, deterministric | 08:22 |
| onovy | this is nice-to-have | 08:22 |
| onovy | and it can fix currently corrupted suffixies, unrelated to current bug we are trying to fix | 08:22 |
| onovy | because now, if hashes.pkl is not consistend with suffix dirs, nothing will fix it | 08:22 |
| cschwede | onovy: PavelK: i think the most important thing is to land PavelK first patch pretty soon | 08:23 |
| onovy | ok, so we will land it to our production and then we can confirm it helps | 08:23 |
| onovy | (rsync metrics should jump down) | 08:23 |
| cschwede | onovy: i’m not sure about the re-computing of every suffix every 100th run. we need to check if there is currently an issue; if the suffix is already in the hashes.pkl, it is likely that it gets recomputed inbetween? | 08:24 |
| onovy | who recompute it? | 08:24 |
| onovy | i think there is no code for recompute it, if there isn't any write traffic to it | 08:25 |
| openstackgerrit | Pavel Kvasnička proposed openstack/swift: Fix missing suffixes during replication https://review.openstack.org/402324 | 08:27 |
| *** rledisez has joined #openstack-swift | 08:35 | |
| *** geaaru has quit IRC | 08:39 | |
| *** asettle has joined #openstack-swift | 08:39 | |
| *** geaaru has joined #openstack-swift | 08:55 | |
| *** cbartz has joined #openstack-swift | 09:00 | |
| *** cbartz has quit IRC | 09:01 | |
| *** dmorita has joined #openstack-swift | 09:08 | |
| *** sams-gleb has quit IRC | 09:12 | |
| *** dmorita has quit IRC | 09:12 | |
| *** sams-gleb has joined #openstack-swift | 09:12 | |
| *** sams-gleb has quit IRC | 09:16 | |
| *** acoles_ is now known as acoles | 09:20 | |
| acoles | cschwede: onovy good morning | 09:32 |
| onovy | morning | 09:34 |
| onovy | acoles: we are going to upgrade 1 (of 4) zone in production cluster with Pavel's patch | 09:35 |
| onovy | acoles: and Pavel is here on IRC, PavelK | 09:35 |
| acoles | hi PavelK | 09:36 |
| PavelK | acoles: Hi | 09:36 |
| cschwede | acoles: Hello Alistair! | 09:36 |
| openstackgerrit | Christian Schwede proposed openstack/swift: Deterministic suffix update in hashes.pkl https://review.openstack.org/402376 | 09:36 |
| cschwede | onovy: PavelK: ^^ the deterministic patch | 09:37 |
| cschwede | i run nearly 2000 replications with that, and this is the distribution: http://paste.openstack.org/show/590429/ | 09:37 |
| cschwede | first column is number of updated suffix, second one is the partition id | 09:38 |
| *** sams-gleb has joined #openstack-swift | 09:38 | |
| cschwede | eh, sorry, 6650 replications | 09:38 |
| cschwede | whatever. it shows at least that every partition is updated as often as all others | 09:38 |
| acoles | ok. O I have been thinking overnight and still not sure why your rsync stats stay high - I read PavelK comment on the gerrit review. At first, yes there will be zero-file-rsyncs because replicator recalculates a partition (random 10%) and destination does not have the "lost suffix", but after each of those rsyncs the destination gets a REPLICATE to recalc sync'd suffixes so after that both sides should be in sync? | 09:39 |
| *** sams-gleb has quit IRC | 09:39 | |
| *** sams-gleb has joined #openstack-swift | 09:39 | |
| onovy | acoles: yep, and they are | 09:40 |
| onovy | and we are writting to cluster, right? | 09:41 |
| acoles | onovy: yeah, so new writes regenerate the rpoblem | 09:41 |
| onovy | yep | 09:41 |
| acoles | problem* | 09:41 |
| acoles | k | 09:41 |
| acoles | but the unnecessary syncs should only occur when sender does a listdir, so approx 10% | 09:42 |
| acoles | does that make sense? | 09:42 |
| *** psachin has quit IRC | 09:43 | |
| *** psachin has joined #openstack-swift | 09:43 | |
| acoles | hmmm, actually new object writes AND reclaimable tombstones may trigger a suffix hash mismatch when the sender does listdirs | 09:45 |
| *** davideagle_ has joined #openstack-swift | 09:47 | |
| acoles | cschwede: PavelK onovy re. the status of patch 401330...I have no problem merging that fix. I want to be sure we do not miss another lurking bug re. rsync rates, but that does not impede the patch fix. | 09:49 |
| patchbot | https://review.openstack.org/#/c/401330/ - swift - Fixed regression in consolidate_hashes | 09:49 |
| onovy | acoles: i think we all ack it's regression and this is fix working | 09:50 |
| onovy | so let's marge it. if it will not lower rsync, we will continue in investigation | 09:50 |
| onovy | and we can reopen my bug later | 09:50 |
| acoles | BUT, we will want to backport this and I'd prefer to have some of the test cleanup from my follow up backported too, and backport may be easier from a single commit. | 09:50 |
| onovy | ack | 09:50 |
| onovy | so should we merge it first? | 09:50 |
| onovy | squash it | 09:50 |
| acoles | onovy: yeah, if we can that would be ideal IMHO. but don't squash yet, it may be easier for cschwede to review my follow up while its separate | 09:52 |
| acoles | cschwede: what do you think? should we try to review and squash https://review.openstack.org/#/c/402199/1 so we have single commit to backport from | 09:53 |
| patchbot | patch 402199 - swift - Add assertions to invalidate_hashes tests | 09:53 |
| acoles | ? | 09:53 |
| cschwede | acoles: onovy: sounds good, looking into this right now | 09:53 |
| cschwede | acoles: we probably want to backport the deterministic patch 402376 too | 09:53 |
| patchbot | https://review.openstack.org/#/c/402376/ - swift - Deterministic suffix update in hashes.pkl | 09:53 |
| onovy | PavelK: please review patch 402199 | 09:54 |
| *** abalfour has quit IRC | 09:54 | |
| patchbot | https://review.openstack.org/#/c/402199/ - swift - Add assertions to invalidate_hashes tests | 09:54 |
| acoles | TBH the one thing I'd really like fixed is the flawed search for distinct suffixdir in the test. But its probably easier to keep change on master as close to a backport as we can. | 09:54 |
| onovy | cschwede: ack for backport of deterministic patch too, but that should bem solo patch | 09:54 |
| onovy | *be | 09:55 |
| cschwede | yep | 09:55 |
| *** abalfour has joined #openstack-swift | 09:55 | |
| acoles | cschwede: yes but lets take that separately. PavelK patch is most urgent. but great you are working on determinism :) | 09:55 |
| acoles | I just squashed those two patches locally and cherry-picked back to mitaka, not completely clean but no probs, one of the modified test was new since mitaka | 10:02 |
| onovy | acoles: yep, have same conflict inside my local backport :) | 10:05 |
| acoles | I think we just include the new tests, can't hurt | 10:05 |
| onovy | right | 10:09 |
| cschwede | acoles: hmm, i see tests failing locally on your patch 402199 | 10:15 |
| *** sams-gleb has quit IRC | 10:15 | |
| patchbot | https://review.openstack.org/#/c/402199/ - swift - Add assertions to invalidate_hashes tests | 10:15 |
| cschwede | http://paste.openstack.org/show/590434/ | 10:16 |
| cschwede | nevermind, my failure | 10:16 |
| *** sams-gleb has joined #openstack-swift | 10:16 | |
| acoles | cschwede: oh? had you made a change? | 10:17 |
| cschwede | acoles: i messed up my saio… fixed now, tests are passing. sorry for the noise | 10:18 |
| *** davideagle_ has quit IRC | 10:19 | |
| *** 18WAAUL0E has joined #openstack-swift | 10:22 | |
| acoles | cschwede: I did notice that the unit test will pick up whatever HASH_PATH_PREFIX etc you have in a local /etc/swift, which is why I care about the test that searches for a different suffixdir, because its not predictable | 10:23 |
| acoles | cschwede: from your paste I infer that your SAIO, like mine, has swift_hash_path_prefix = changeme | 10:23 |
| acoles | :) | 10:23 |
| cschwede | acoles: agreed, i was just checking that out :) | 10:23 |
| cschwede | and with your test i also agree that there is a possible test issue in test_delete_when_suffix_hash_not_in_existing_pkl, and suffixes might be the same there - so i’m fine with merging both patches | 10:24 |
| cschwede | acoles: i left a +2 on your patch 402199, just as a reference | 10:25 |
| patchbot | https://review.openstack.org/#/c/402199/ - swift - Add assertions to invalidate_hashes tests | 10:25 |
| *** asettle__ has joined #openstack-swift | 10:25 | |
| acoles | cschwede: thanks, I'll wait for PavelK or onovy to look over it | 10:25 |
| acoles | cschwede: yeah, test_delete_when_suffix_hash_not_in_existing_pkl would always pass, but was not guaranteed to be using two distinct suffixes | 10:26 |
| acoles | if the test env had the "wrong" HASH_PATH_PREFIX | 10:27 |
| *** asettle has quit IRC | 10:29 | |
| acoles | afk in search of coffee | 10:31 |
| *** asettle__ is now known as asettle | 10:39 | |
| openstackgerrit | Merged openstack/swift: use the new upper constraints infra features https://review.openstack.org/354291 | 10:40 |
| acoles | back, so PavelK if you are happy with patch 402199 then lets squash it into your then cschwede and I can get that merged | 10:44 |
| patchbot | https://review.openstack.org/#/c/402199/ - swift - Add assertions to invalidate_hashes tests | 10:44 |
| acoles | I can push it if you want, I have done the squash already | 10:45 |
| *** SkyRocknRoll has quit IRC | 11:19 | |
| onovy | acoles: Pavel will be back soon | 11:29 |
| onovy | give as few minutes :) | 11:29 |
| onovy | us | 11:29 |
| onovy | acoles: you can squash, pavel ack yor patch | 11:35 |
| acoles | onovy: ok | 11:36 |
| PavelK | \quit | 11:37 |
| *** PavelK has quit IRC | 11:37 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Fixed regression in consolidate_hashes https://review.openstack.org/401330 | 11:38 |
| onovy | acoles: i think you should add yourself as co-author? | 11:39 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Fixed regression in consolidate_hashes https://review.openstack.org/401330 | 11:41 |
| acoles | oh did Pavel quit? | 11:42 |
| acoles | onovy: say "great work" to Pavel! :) | 11:42 |
| acoles | cschwede: I added +2 now ^^ | 11:46 |
| cschwede | acoles: onovy: thx, approved the patch - should be in the gate soon | 11:53 |
| cschwede | acoles: does it make sense to create another bugreport about the non-deterministic update of hashes.pkl? | 11:53 |
| acoles | I think so, particularly because you pointed out that in extreme (low part power) the listdir might never happen | 11:54 |
| cschwede | acoles: yeah… alright, i’ll do that | 11:55 |
| *** oshritf has joined #openstack-swift | 12:03 | |
| *** vint_bra has joined #openstack-swift | 12:12 | |
| cschwede | acoles: hmm, even worse, it happens when a replicator has less than 10 partitions to replicate, which might be much more likely | 12:15 |
| cschwede | https://bugs.launchpad.net/swift/+bug/1644807 | 12:16 |
| openstack | Launchpad bug 1644807 in OpenStack Object Storage (swift) "hashes.pkl suffix update is not deterministic and might be never triggered" [High,In progress] - Assigned to Christian Schwede (cschwede) | 12:16 |
| acoles | cschwede: ack | 12:16 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Add test for object updater user-agent header https://review.openstack.org/398523 | 12:21 |
| openstackgerrit | Christian Schwede proposed openstack/swift: Deterministic suffix update in hashes.pkl https://review.openstack.org/402376 | 12:22 |
| cschwede | dumb gerrit, i just edited the commit msg… | 12:22 |
| *** 18WAAUL0E has quit IRC | 12:29 | |
| acoles | huh, I just discovered this patch 250022 which got +"A back in March and never landed :/ | 12:30 |
| patchbot | https://review.openstack.org/#/c/250022/ - swift - Add missing unittests for some recon methods | 12:30 |
| acoles | I guess none of our dashboards are showing reviews that have +A | 12:31 |
| *** pavel_ has joined #openstack-swift | 12:35 | |
| *** pavel1 has joined #openstack-swift | 12:35 | |
| *** pavel1 is now known as PavelK | 12:36 | |
| PavelK | cschwede: acoles: Thanks for CR, we release to one locality in production | 12:39 |
| *** oshritf has quit IRC | 12:40 | |
| acoles | cschwede: onovy PavelK backports here https://review.openstack.org/402509 and https://review.openstack.org/402508 | 12:43 |
| patchbot | patch 402509 - swift (stable/newton) - Fixed regression in consolidate_hashes | 12:43 |
| patchbot | patch 402508 - swift (stable/mitaka) - Fixed regression in consolidate_hashes | 12:43 |
| openstackgerrit | Sachin Patil proposed openstack/swift: get_part_nodes should raise error on invalid part https://review.openstack.org/402522 | 12:48 |
| psachin | ping clayg. Please review https://review.openstack.org/402522 | 12:53 |
| patchbot | patch 402522 - swift - get_part_nodes should raise error on invalid part | 12:53 |
| *** davideagle_ has joined #openstack-swift | 12:55 | |
| onovy | acoles: ad patch 250022 // it depends on other never-landed patch | 13:00 |
| patchbot | https://review.openstack.org/#/c/250022/ - swift - Add missing unittests for some recon methods | 13:00 |
| onovy | this: https://review.openstack.org/#/c/240036/ | 13:00 |
| patchbot | patch 240036 - swift - Add unit test to test cli recon for the commands | 13:00 |
| acoles | oic | 13:04 |
| onovy | Peter is sad about it :( | 13:04 |
| *** davideagle_ has quit IRC | 13:16 | |
| *** amoralej is now known as amoralej|lunch | 13:17 | |
| *** davideagle has joined #openstack-swift | 13:17 | |
| *** sams-gleb has quit IRC | 13:35 | |
| *** sams-gleb has joined #openstack-swift | 13:37 | |
| *** rcernin has quit IRC | 13:39 | |
| *** rcernin has joined #openstack-swift | 13:49 | |
| *** PavelK has quit IRC | 13:53 | |
| *** hogepodge has quit IRC | 13:59 | |
| *** tqtran has joined #openstack-swift | 14:14 | |
| *** psachin has quit IRC | 14:17 | |
| mahatic_ | so I update my liberasurecode to 1.3.1 and I still see this error on starting swift services -> http://paste.openstack.org/show/590463/ | 14:18 |
| mahatic_ | kota_: or anyone else: would you happen to be familiar with this error? | 14:18 |
| *** tqtran has quit IRC | 14:19 | |
| onovy | mahatic_: it looks like your python-pyeclib doesn't see version from liberasurecode | 14:23 |
| mahatic_ | I got that when I upgraded my pyeclib and then installed latest liberasurecode | 14:23 |
| *** hogepodge has joined #openstack-swift | 14:23 | |
| onovy | but you have liberasurecode, because there are errors directly from it :) | 14:23 |
| onovy | how you installed liberasurecode? | 14:23 |
| onovy | http://paste.openstack.org/show/590466/ | 14:25 |
| openstackgerrit | Christian Schwede proposed openstack/swift: Deterministic suffix update in hashes.pkl https://review.openstack.org/402376 | 14:25 |
| mahatic_ | Installed from here https://launchpad.net/ubuntu/+source/liberasurecode | 14:25 |
| cschwede | acoles: now with better tests ^^ | 14:25 |
| onovy | mahatic_: so you downloaded binary .deb package ? | 14:25 |
| onovy | what's your current distro? | 14:26 |
| *** amoralej|lunch is now known as amoralej | 14:26 | |
| onovy | mahatic_: call this pls: python -c 'from pyeclib_c import get_liberasurecode_version ; print get_liberasurecode_version()' | 14:28 |
| *** hogepodge has quit IRC | 14:28 | |
| onovy | i guess: you need to reinstall python-pyeclib, because pyeclib_c is not correctly linked to new liberasurecode | 14:29 |
| mahatic_ | onovy: hi, here you go: http://paste.openstack.org/show/590469/ | 14:30 |
| onovy | this is my: 66048 | 14:31 |
| onovy | (1.3.1) | 14:31 |
| mahatic_ | hmm there are reports of this issue (but in syslog I believe) https://bitbucket.org/tsg-/liberasurecode/issues/23/spurious-messages-in-syslog | 14:31 |
| *** links has quit IRC | 14:31 | |
| mahatic_ | I think I need to update the installed path in conf | 14:32 |
| mahatic_ | will check that | 14:32 |
| onovy | your pyeclib_c is definetly using wrong liberasurecode C lib | 14:32 |
| onovy | really old one... | 14:32 |
| onovy | hmm it's funny. this version is "packed" inside python-pyeclib | 14:37 |
| onovy | so it just says: when you built python-pyeclib, this version of liberasurecode was installed on system | 14:37 |
| onovy | so rebuilt python-pyeclib -> fixed. | 14:37 |
| onovy | mahatic_: must go now. reinstall/rebuild python-pyeclib and it will work. bye :) | 14:39 |
| mahatic_ | onovy: hey sure. looks like my pkg is installed in the local lib path, need to fix it, thanks and bye! :) | 14:40 |
| acoles | mahatic_: hi mahatic_ , I am no expert, but I built liberasurecode from source follwing instructions on the README | 14:41 |
| acoles | first I purged old liberasurecode that i had installed with apt-get | 14:41 |
| acoles | and I used pip to upgrade pyeclib | 14:42 |
| mahatic_ | acoles: hi, what was your source of liberasurecode? I picked it up from here https://launchpad.net/ubuntu/+source/liberasurecode and then did what README said | 14:42 |
| kota_ | hello | 14:43 |
| kota_ | mahatic: maybe you need to rebuild pyeclib *after* liberasurecode installation | 14:43 |
| acoles | mahatic_: https://github.com/openstack/liberasurecode | 14:44 |
| acoles | which version did you grab from launchpad, you need 1.3.1? | 14:44 |
| kota_ | mahatic: via `python setup.py install` at pyeclib or `pip uninstall pyeclib` and `pip install pyeclib` agagin | 14:44 |
| mahatic_ | I took 1.3.1 | 14:45 |
| mahatic_ | kota_: I didn't build *after* upgrading, will check | 14:45 |
| kota_ | ah, onovy has already suggested that. | 14:45 |
| mahatic_ | kota_: hi btw, shouldn't you be off for the weekend? :) | 14:46 |
| onovy | kota_: why is liberasurecode packed inside python-pyeclib binary when it is build? | 14:46 |
| onovy | this way i can't upgrade liberasurecode package in Debian without rebuilding python-pyeclib after every minor upgrade of lib | 14:46 |
| kota_ | onovy: actually, not packed | 14:46 |
| acoles | I didnt think it was any more | 14:47 |
| onovy | kota_: it's | 14:47 |
| onovy | if i install python-pyeclib from debian/sid, get_versions says i have 1.2.0 version | 14:47 |
| kota_ | onovy: pyeclib just refers the liberasurecode_version.h which is installed at liberasurecode installation | 14:47 |
| onovy | even i have liberasurecode version 1.3.1 | 14:47 |
| onovy | no :) | 14:47 |
| onovy | pyeclib refers liberasurecode_version.h, which is DURING BUILD read | 14:47 |
| onovy | so if i have binary distribution of pyeclib, i have packed version | 14:48 |
| kota_ | onovy: exacctly it's better to refer the actual version from .so lib which is used by pyeclib | 14:48 |
| onovy | pyeclib_c is build during package build | 14:48 |
| onovy | right :) | 14:48 |
| *** dmorita has joined #openstack-swift | 14:48 | |
| onovy | is it possible to fix it? :) | 14:48 |
| kota_ | yes, add the functionality to liberasurecode and then make pyeclib to use it. | 14:49 |
| onovy | yep, there aren't ABI inside liberasurecode, right? | 14:49 |
| onovy | (for version read) | 14:49 |
| *** sams-gleb has quit IRC | 14:49 | |
| kota_ | onovy: however, probably it will make hard version constraint for us like recent isa-l issues | 14:50 |
| *** sams-gleb has joined #openstack-swift | 14:50 | |
| kota_ | onovy: so that, I'm now pending it to decide how we should handle the version limitation. | 14:50 |
| kota_ | s/to decide/until deciding/ | 14:51 |
| onovy | make function inside liberasurecode for "gimme version" | 14:51 |
| kota_ | yes | 14:51 |
| onovy | pyeclib_c will ask this function | 14:51 |
| mahatic_ | acoles: kota_ onovy rebuilding pyeclib fixed it, thanks! | 14:51 |
| onovy | and check constrains | 14:51 |
| kota_ | onovy: yeah | 14:51 |
| onovy | because using .h can cause big problems | 14:51 |
| kota_ | mahatic_: congrats | 14:51 |
| onovy | like mahatic_ now had | 14:51 |
| kota_ | i think i had description for the constraint on the 1.3.1-ish commit log... | 14:52 |
| *** dmorita has quit IRC | 14:52 | |
| onovy | kota_: sry i must go now. can we continue discussion later pls? | 14:53 |
| acoles | mahatic_: also when you next run any tox job you'll probably need to add the -r flag to reinstall pyeclib in the venvs | 14:53 |
| *** sams-gleb has quit IRC | 14:54 | |
| kota_ | onovy: sure, could we make it in the next week? it's getting midnight in my timezone in Fri | 14:55 |
| mahatic_ | acoles: oic | 14:55 |
| kota_ | hmm... https://review.openstack.org/#/c/398744/ has not laned yet :/ | 14:55 |
| patchbot | patch 398744 - pyeclib - Add soft warning log line when using liberasurecod... | 14:55 |
| mahatic_ | acoles: thank you, I could really use that info :) | 14:56 |
| mahatic_ | kota_: yeah, I was looking for a patch like that, I think that needs to be landed | 14:56 |
| *** hogepodge has joined #openstack-swift | 15:00 | |
| kota_ | k, JST is now in Sat, so leaving off from online | 15:01 |
| kota_ | bye, enjoy everyone your weekend. | 15:01 |
| * kota_ is going to offline | 15:01 | |
| *** hogepodge has quit IRC | 15:05 | |
| openstackgerrit | Merged openstack/swift: Fixed regression in consolidate_hashes https://review.openstack.org/401330 | 15:09 |
| mahatic_ | kota_: good night, have a nice weekend | 15:12 |
| *** Jeffrey4l has joined #openstack-swift | 15:14 | |
| *** sams-gleb has joined #openstack-swift | 15:17 | |
| *** rcernin has quit IRC | 15:17 | |
| *** openstackgerrit has quit IRC | 15:18 | |
| *** openstackgerrit has joined #openstack-swift | 15:19 | |
| *** ppai has quit IRC | 15:21 | |
| *** rcernin has joined #openstack-swift | 15:22 | |
| *** rcernin has quit IRC | 15:30 | |
| openstackgerrit | Flavio Percoco proposed openstack/swift-bench: Show team and repo badges on README https://review.openstack.org/402850 | 15:38 |
| openstackgerrit | Flavio Percoco proposed openstack/python-swiftclient: Show team and repo badges on README https://review.openstack.org/402853 | 15:38 |
| *** asettle has quit IRC | 15:38 | |
| openstackgerrit | Flavio Percoco proposed openstack/swift: Show team and repo badges on README https://review.openstack.org/402856 | 15:39 |
| openstackgerrit | Flavio Percoco proposed openstack/swift-specs: Show team and repo badges on README https://review.openstack.org/402859 | 15:40 |
| *** bkopilov has joined #openstack-swift | 15:58 | |
| *** admin6 has left #openstack-swift | 15:59 | |
| *** cbartz has joined #openstack-swift | 16:01 | |
| *** geaaru has quit IRC | 16:34 | |
| *** thebloggu has joined #openstack-swift | 16:42 | |
| *** PavelK has joined #openstack-swift | 16:42 | |
| *** PavelK has quit IRC | 16:48 | |
| *** links has joined #openstack-swift | 16:51 | |
| *** rledisez has quit IRC | 17:10 | |
| *** Jeffrey4l has quit IRC | 17:12 | |
| *** links has quit IRC | 17:15 | |
| *** delewis has joined #openstack-swift | 17:21 | |
| *** hogepodge has joined #openstack-swift | 17:24 | |
| *** hseipp has quit IRC | 17:30 | |
| *** pcaruana has quit IRC | 17:33 | |
| *** tqtran has joined #openstack-swift | 18:16 | |
| *** hseipp has joined #openstack-swift | 18:20 | |
| *** tqtran has quit IRC | 18:21 | |
| *** amoralej is now known as amoralej|off | 18:23 | |
| *** hseipp has quit IRC | 18:27 | |
| *** davideagle has quit IRC | 18:31 | |
| *** ChubYann has joined #openstack-swift | 18:47 | |
| *** acoles is now known as acoles_ | 18:52 | |
| *** asettle has joined #openstack-swift | 19:06 | |
| *** cbartz has quit IRC | 19:11 | |
| *** hogepodge has quit IRC | 19:11 | |
| *** asettle has quit IRC | 19:25 | |
| *** dmorita has joined #openstack-swift | 19:48 | |
| *** dmorita has quit IRC | 19:53 | |
| *** thebloggu has quit IRC | 19:55 | |
| *** hogepodge has joined #openstack-swift | 20:08 | |
| *** hseipp has joined #openstack-swift | 20:30 | |
| *** silor has quit IRC | 20:34 | |
| *** hseipp has quit IRC | 21:09 | |
| *** vint_bra has quit IRC | 21:41 | |
| *** nikivi has joined #openstack-swift | 21:53 | |
| *** nikivi has quit IRC | 21:58 | |
| *** pavel has joined #openstack-swift | 21:59 | |
| *** tqtran has joined #openstack-swift | 22:17 | |
| pavel | onovy: I can't understand why rsync runs more often after restarting replicator daemon (I mean BEFORE fix). Partitions are ordered randomly each replication cycle so I can't understand why there is "better" order when replicator restarted. Any idea? | 22:18 |
| *** pavel is now known as PavelK | 22:18 | |
| *** tqtran has quit IRC | 22:22 | |
| *** PavelK has quit IRC | 22:23 | |
| *** PavelK has joined #openstack-swift | 22:30 | |
| *** PavelK has quit IRC | 22:31 | |
| *** catintheroof has joined #openstack-swift | 22:34 | |
| *** hogepodge has quit IRC | 22:40 | |
| *** nikivi has joined #openstack-swift | 22:43 | |
| *** hogepodge has joined #openstack-swift | 22:51 | |
| *** nikivi has quit IRC | 23:02 | |
| *** davideagle has joined #openstack-swift | 23:28 | |
| *** sams-gleb has quit IRC | 23:40 | |
| *** sams-gleb has joined #openstack-swift | 23:41 | |
| *** sams-gleb has quit IRC | 23:45 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!