*** gyee has quit IRC | 00:07 | |
*** itlinux has quit IRC | 00:16 | |
openstackgerrit | Matthew Oliver proposed openstack/swift feature/deep: sharding - Make swift-container-info shard aware https://review.openstack.org/505580 | 00:23 |
---|---|---|
mattoliverau | ^^ just a rebase, but /me suspects it'll need to be rebased again soon with the merge from master patch ;) | 00:25 |
*** bigdogstl has joined #openstack-swift | 00:29 | |
openstackgerrit | Tim Burke proposed openstack/swift feature/deep: Bring us closer to master https://review.openstack.org/505858 | 00:31 |
timburke | sorry mattoliverau, gonna push for the merge-from-master first :-) | 00:33 |
mattoliverau | I know :) | 00:33 |
timburke | that follow-up i just pushed would be pretty nice, too | 00:34 |
timburke | though... maybe there aren't actually any conflicts | 00:34 |
mattoliverau | timburke: nice! | 00:36 |
*** bigdogstl has quit IRC | 00:36 | |
mattoliverau | I love it when things get simpler :) | 00:36 |
* mattoliverau might go from a walk on the beach as I wait for tests to run ;) | 00:37 | |
*** itlinux has joined #openstack-swift | 00:38 | |
timburke | you can go ahead and merge, but maybe wait until after the patch in the gate lands | 00:39 |
*** tovin07_ has joined #openstack-swift | 00:41 | |
*** openstackgerrit has quit IRC | 00:41 | |
*** openstackgerrit has joined #openstack-swift | 00:44 | |
*** bigdogstl has joined #openstack-swift | 00:53 | |
*** hoonetorg has joined #openstack-swift | 00:56 | |
*** klrmn has quit IRC | 01:06 | |
*** bigdogstl has quit IRC | 01:15 | |
*** cshastri has joined #openstack-swift | 01:25 | |
*** itlinux has quit IRC | 01:28 | |
*** Sukhdev has quit IRC | 01:39 | |
*** itlinux has joined #openstack-swift | 01:49 | |
*** itlinux has quit IRC | 01:51 | |
*** mingyu has quit IRC | 01:53 | |
*** kota_ has quit IRC | 01:56 | |
*** kota_ has joined #openstack-swift | 01:56 | |
*** ChanServ sets mode: +v kota_ | 01:56 | |
*** kei_yama has quit IRC | 01:57 | |
*** mingyu has joined #openstack-swift | 01:59 | |
*** kei_yama has joined #openstack-swift | 02:04 | |
*** vint_bra has joined #openstack-swift | 02:11 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/python-swiftclient master: Updated from global requirements https://review.openstack.org/89250 | 02:17 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/swift master: Updated from global requirements https://review.openstack.org/88736 | 02:18 |
*** bigdogstl has joined #openstack-swift | 02:25 | |
*** bigdogstl has quit IRC | 02:33 | |
*** mingyu has quit IRC | 02:45 | |
*** mingyu has joined #openstack-swift | 02:47 | |
*** mingyu has quit IRC | 02:48 | |
openstackgerrit | Merged openstack/swift feature/deep: Merge branch 'master' into merge-from-master https://review.openstack.org/505844 | 02:49 |
*** gkadam-away is now known as gkadam | 02:55 | |
*** janonymous has joined #openstack-swift | 03:00 | |
*** vint_bra has quit IRC | 03:04 | |
*** _ix has joined #openstack-swift | 03:10 | |
*** Sukhdev has joined #openstack-swift | 03:24 | |
*** bigdogstl has joined #openstack-swift | 03:31 | |
*** bigdogstl has quit IRC | 03:35 | |
*** kei_yama has quit IRC | 03:44 | |
openstackgerrit | Matthew Oliver proposed openstack/swift feature/deep: Bring us closer to master https://review.openstack.org/505858 | 03:55 |
mattoliverau | ^^ fixes the pep8 error so I can merge it :) | 03:55 |
*** kei_yama has joined #openstack-swift | 03:59 | |
*** bigdogstl has joined #openstack-swift | 04:00 | |
*** bigdogstl has quit IRC | 04:11 | |
*** silor has joined #openstack-swift | 04:23 | |
*** SkyRocknRoll has joined #openstack-swift | 04:33 | |
*** SkyRocknRoll has quit IRC | 04:34 | |
*** bigdogstl has joined #openstack-swift | 04:38 | |
*** psachin has joined #openstack-swift | 04:41 | |
*** bigdogstl has quit IRC | 04:43 | |
*** links has joined #openstack-swift | 04:43 | |
*** SkyRocknRoll has joined #openstack-swift | 04:52 | |
*** itlinux has joined #openstack-swift | 05:00 | |
openstackgerrit | Matthew Oliver proposed openstack/swift feature/deep: Merge Shard-{Account, Container} to 1 metadata https://review.openstack.org/505976 | 05:01 |
*** pcaruana has joined #openstack-swift | 05:09 | |
*** SkyRocknRoll_ has joined #openstack-swift | 05:15 | |
*** SkyRocknRoll_ has joined #openstack-swift | 05:15 | |
*** Sukhdev has quit IRC | 05:23 | |
*** silor has quit IRC | 05:33 | |
*** bigdogstl has joined #openstack-swift | 05:40 | |
*** bigdogstl has quit IRC | 05:44 | |
*** rmcall has quit IRC | 05:57 | |
*** rmcall has joined #openstack-swift | 06:02 | |
*** vito_ has joined #openstack-swift | 06:04 | |
*** vito_ is now known as Guest92796 | 06:04 | |
*** bigdogstl has joined #openstack-swift | 06:05 | |
*** rmcall has quit IRC | 06:06 | |
*** bigdogstl has quit IRC | 06:09 | |
*** Guest92796 has quit IRC | 06:25 | |
*** rcernin has joined #openstack-swift | 06:28 | |
openstackgerrit | Matthew Oliver proposed openstack/swift feature/deep: Use temp db and a rename for set_sharding_state https://review.openstack.org/506002 | 06:33 |
*** itlinux has quit IRC | 06:37 | |
*** itlinux has joined #openstack-swift | 06:40 | |
*** itlinux has quit IRC | 06:40 | |
*** SkyRocknRoll_ has quit IRC | 06:46 | |
*** cschwede has joined #openstack-swift | 06:47 | |
*** ChanServ sets mode: +v cschwede | 06:47 | |
*** SkyRocknRoll has quit IRC | 06:47 | |
*** SkyRocknRoll_ has joined #openstack-swift | 07:00 | |
*** SkyRocknRoll has joined #openstack-swift | 07:03 | |
*** ChubYann has quit IRC | 07:04 | |
*** pcaruana has quit IRC | 07:05 | |
openstackgerrit | Merged openstack/swift feature/deep: sharding - Make swift-container-info shard aware https://review.openstack.org/505580 | 07:05 |
*** skudlik has joined #openstack-swift | 07:09 | |
*** mingyu has joined #openstack-swift | 07:13 | |
*** ChanServ sets mode: +o openstack | 07:14 | |
*** mingyu has quit IRC | 07:14 | |
*** bigdogstl has joined #openstack-swift | 07:17 | |
*** tesseract has joined #openstack-swift | 07:19 | |
*** bigdogstl has quit IRC | 07:22 | |
*** pcaruana has joined #openstack-swift | 07:23 | |
acoles | good morning | 07:25 |
*** cbartz has joined #openstack-swift | 07:27 | |
acoles | mattoliverau: are you still around? | 07:27 |
acoles | timburke: thanks for the merge from master | 07:28 |
mattoliverau | yeah, but I need to go save my wife from the toddler, mind if I go deal with the baby and get back to you in aabout and hour or so? | 07:30 |
acoles | mattoliverau: no problem at all, thanks | 07:37 |
*** ypbg has joined #openstack-swift | 07:45 | |
*** cschwede_ has joined #openstack-swift | 07:49 | |
*** cschwede_ has quit IRC | 07:52 | |
*** geaaru has joined #openstack-swift | 07:52 | |
*** cschwede_ has joined #openstack-swift | 07:53 | |
*** gkadam has quit IRC | 07:54 | |
*** cschwede_ has quit IRC | 07:54 | |
*** cschwede_ has joined #openstack-swift | 07:55 | |
*** cschwede_ has quit IRC | 07:56 | |
*** cschwede_ has joined #openstack-swift | 07:59 | |
*** cschwede has quit IRC | 08:03 | |
*** cschwede_ has quit IRC | 08:04 | |
*** cschwede has joined #openstack-swift | 08:05 | |
openstackgerrit | Merged openstack/swift feature/deep: Bring us closer to master https://review.openstack.org/505858 | 08:14 |
*** bigdogstl has joined #openstack-swift | 08:28 | |
*** bigdogstl has quit IRC | 08:34 | |
openstackgerrit | Joel Wright proposed openstack/swift master: Add Preamble and Postamble to SLO and SegmentedIterable https://review.openstack.org/365371 | 08:47 |
*** gkadam has joined #openstack-swift | 09:01 | |
*** bigdogstl has joined #openstack-swift | 09:14 | |
*** bigdogstl has quit IRC | 09:19 | |
mattoliverau | acoles: o/ | 09:21 |
acoles | mattoliverau: hi | 09:22 |
acoles | mattoliverau: so I had a couple of questions around this patch https://review.openstack.org/#/c/505763/ | 09:23 |
patchbot | patch 505763 - swift (feature/deep) - sharding: WIP make find_pivot_ranges a broker method | 09:23 |
mattoliverau | oh yeah, I started making comments, but never pushed and on my other computer. opps | 09:24 |
mattoliverau | oh cool, chrome has remembered them on another computer.. thats kinda scary | 09:24 |
acoles | mattoliverau: the main thing was here https://review.openstack.org/#/c/505763/1/swift/container/sharder.py@904 | 09:28 |
patchbot | patch 505763 - swift (feature/deep) - sharding: WIP make find_pivot_ranges a broker method | 09:28 |
acoles | 2 things actually: 1. should we just make sure that find_shard_ranges always gives us back the last range, so the special case is not needed and 2. what to do when creating a shard db fails | 09:29 |
mattoliverau | the initial version of the find shardranges would only find uppers, until it reached the end.. in which case the last upper to None (or now '') would be the last node.. but yeah, nothing stopping us from the finding ranges to add it. | 09:31 |
acoles | working backwards, I was thinking more about 2 and wonder if we should persist *all* the shard ranges that have been found even when we may fail to create the DBs. otherwise we'd waste all that scanning work that has been done to find the pivots | 09:31 |
*** SkyRocknRoll_ has quit IRC | 09:31 | |
acoles | but that implies having some state in ShardRange to indicate that it has been found but not yet been created. | 09:32 |
mattoliverau | hmm, well I agree we don't want to waste the scanning. So we make sure they're written to the shardranges table | 09:33 |
acoles | yeah and in Denver we were talking about maybe having one daemon find the shard ranges and another (the replicator) process them, so persisting them as they are found would make sense for that too. | 09:34 |
mattoliverau | currently we have an upto metadata for each index, which it updates once sharding.. if it can't and doesn't write it then this wont be updated and thus the cleave to the container will happen again. | 09:34 |
mattoliverau | although the initial PUT wont. meaning it wont exist, but the cleave currently is a create and replicate so it will be created | 09:34 |
acoles | 'upto' is sysmeta on the root? | 09:35 |
mattoliverau | If we are doing cascade deletes where the shard audit will create missing containers inthe root container ranges, then this will work for missing PUTs during cleaving too | 09:35 |
mattoliverau | upto on whatever container is cleaving | 09:35 |
mattoliverau | so root initially, and on the shards | 09:35 |
mattoliverau | just so the nodes can coordinate. ie, I'm not finished sharding, don't unlink if my index hasn't finished cleaving otherwise we might miss data | 09:36 |
mattoliverau | however, maybe adding a state isn't a bad idea... maybe it can be used when deleteting, marked as gone.. so we know if we need to recreate. But maybe that doesnt matter. | 09:38 |
acoles | is that 'X-Container-Sysmeta-Shard-Last-%d' ? | 09:38 |
mattoliverau | yup | 09:38 |
acoles | got it | 09:38 |
mattoliverau | where %d is in index of the node. | 09:38 |
acoles | yeah | 09:38 |
mattoliverau | and that should persist beyond rebalances which is nice (ie. idx remains the same). | 09:39 |
acoles | ok, so in principle we need a way to differentiate between ShardRanges that have been found and those that have been cleaved. maybe that sysmeta does the job already, maybe we add state to the ShardRange. | 09:40 |
acoles | for now i just wanted to clarify the principle :) | 09:40 |
*** tesseract has quit IRC | 09:41 | |
*** tesseract has joined #openstack-swift | 09:41 | |
acoles | I still need to climb the learning curve more before jumping on a solution | 09:41 |
mattoliverau | +1 yeah, we need to keep track of where wrere up to.. but it gets complicated when we also want to keep track of each primary nodes progress in the off chance one contains data the others dont | 09:42 |
acoles | yup | 09:42 |
acoles | so another thing I noticed, when scanning for pivots, is that the 'progress' is calculated assuming container_shard_size doesn't change from one cycle to next. But it may change in config :( | 09:45 |
mattoliverau | yeah.. that's a problem :( | 09:46 |
acoles | which led to another question...when does the ShardRange object count first get updated? | 09:46 |
acoles | cos it seems to be zero when the range is first found | 09:46 |
mattoliverau | at the moment the shardrange object count is being updated after every visit from a sharder. It sends and update to the root container. | 09:47 |
acoles | (hmmm, the object count probably doesn't help cos it will change as more updates land in the shard?) | 09:47 |
mattoliverau | hmm, yeah | 09:47 |
mattoliverau | nothing stopping us from adding the inital count tho | 09:48 |
acoles | right, that's what I wondered, whether each ShardRange needs to stash a shard_size which is its initial size. Then we can sum them to establish 'progress so far' | 09:49 |
*** mvk has quit IRC | 09:49 | |
acoles | or stash progress in sysmeta | 09:49 |
mattoliverau | if we have the last pivot, do we need to know the progress? we can just search until we've hit the end.. or not (hit batch) | 09:49 |
acoles | ^^ that works too, but the progress allows an optimisation where you can tell you're going to land past the end of the container without doing the query (if progress + shard_size > object_count) | 09:50 |
mattoliverau | I can't remember what the offset will return if beyond the list. | 09:50 |
mattoliverau | and maybe that was the problem | 09:51 |
mattoliverau | yeah | 09:51 |
mattoliverau | just thinking out loud :) | 09:51 |
acoles | me too! | 09:51 |
acoles | ok I'll ponder that some more. | 09:53 |
*** tovin07_ has quit IRC | 09:53 | |
mattoliverau | adding another field just to store counts while cleaving is annoying. | 09:54 |
acoles | we don't have to fix all this right now, I just want to build a picture of where we need to get to. | 09:54 |
mattoliverau | when cleaving shards, the obj counts will not updated, only the ranges in the root will be.. so we can just use those counts to track | 09:54 |
mattoliverau | when cleaving a root, we could maybe not update root range counts while cleaving | 09:55 |
mattoliverau | the progress is then easy to sum | 09:55 |
mattoliverau | just like we do for summing up ShardRanges | 09:55 |
acoles | but if all we have done is found (not yet cleaved) some ranges, then won't object counts still be zero? | 09:55 |
mattoliverau | unless we put the shard_size in there | 09:56 |
acoles | into object_count? | 09:56 |
mattoliverau | yeah, but dirty | 09:56 |
mattoliverau | but reusing what we have. | 09:56 |
acoles | OIC | 09:57 |
mattoliverau | rather then adding a feild that will exist for only a very short time in the containers liftcycle | 09:57 |
acoles | yes. so we can think hard about whether that field would *ever* be useful later in the lifecycle, but I tend to agree. | 09:58 |
*** cshastri has quit IRC | 09:59 | |
acoles | a companion to 'X-Container-Sysmeta-Shard-Last-%d' might do it i.e. 'X-Container-Sysmeta-Shard-Progress-%d' | 09:59 |
acoles | I'll try to get my head round how 'X-Container-Sysmeta-Shard-Last-%d' is updated and think more on that | 10:00 |
mattoliverau | oh yeah, thats true. we do only need to track progress upto.. not all since.. I'm over thinking | 10:01 |
acoles | how/when I mean | 10:01 |
mattoliverau | yeah, cool. | 10:01 |
mattoliverau | nice work finding those issues :) | 10:02 |
acoles | of course, this is moot if we are only performing complete scans on smaller containers - we'll get all ShardRanges in one cycle. But I hate to punt on thinking through stuff like this and get bitten in the future. | 10:02 |
acoles | one last thing...I'd like my broker find_shard_ranges to return a list of *ShardRanges*...but to construct them I need to build a shard name and that involves the node index. That would be the only reason to pass node info to the broker :( so can you remind me does the node index need to be in the name? | 10:06 |
*** kei_yama has quit IRC | 10:08 | |
mattoliverau | to be honest, I'm not sure having the nodeid in the name really gives us anything | 10:09 |
mattoliverau | does it really matter who was the scanner.. and atm it should always be 0 | 10:09 |
mattoliverau | maybe <container>-<md5 of timestamp + pivot> is enough | 10:10 |
acoles | 'atm' yes. but maybe timestamp is enough? | 10:10 |
acoles | lol. that!! | 10:10 |
mattoliverau | yeah, lets drop the node idx. | 10:11 |
mattoliverau | :) | 10:11 |
acoles | do we need <md5 of timestamp+lower+upper> *just in case another node finds different lower, same upper at same time ?!? | 10:12 |
acoles | or is that overkill. IDK | 10:12 |
mattoliverau | true, but if the same container gets more then it's suppose to (in range) then misplaced object will move them and there should be less to move. | 10:13 |
acoles | there is also a rather fundamental assumption in swift about distinct timestamps ;) | 10:15 |
acoles | mattoliverau: I just looked at my Canberra clock - thanks for coming back late to chat, really helped me. And today also started early for you with the meeting. | 10:16 |
*** mvk has joined #openstack-swift | 10:19 | |
mattoliverau | Nps, thanks for all the awesome work and thought you've been putting in :) | 10:31 |
*** bigdogstl has joined #openstack-swift | 10:38 | |
*** bigdogstl has quit IRC | 10:45 | |
*** mat128 has joined #openstack-swift | 11:00 | |
*** gkadam has quit IRC | 11:21 | |
*** chlong has joined #openstack-swift | 11:52 | |
*** psachin has quit IRC | 12:06 | |
*** gkadam has joined #openstack-swift | 12:06 | |
*** bigdogstl has joined #openstack-swift | 12:10 | |
*** guimaluf has quit IRC | 12:25 | |
*** MVenesio has joined #openstack-swift | 12:34 | |
*** SkyRocknRoll has quit IRC | 12:41 | |
*** catintheroof has joined #openstack-swift | 12:41 | |
*** bigdogstl has quit IRC | 12:46 | |
*** gkadam has quit IRC | 12:58 | |
*** gkadam has joined #openstack-swift | 12:59 | |
*** mat128 has quit IRC | 13:00 | |
*** _ix_ has joined #openstack-swift | 13:02 | |
*** gkadam has quit IRC | 13:02 | |
*** d0ugal has quit IRC | 13:02 | |
*** gkadam has joined #openstack-swift | 13:02 | |
*** d0ugal has joined #openstack-swift | 13:03 | |
*** _ix has quit IRC | 13:04 | |
tdasilva | good morning | 13:11 |
acoles | tdasilva: o/ | 13:13 |
*** lucasxu has joined #openstack-swift | 13:16 | |
*** _ix_ has quit IRC | 13:21 | |
openstackgerrit | Joel Wright proposed openstack/swift master: WIP: Add TLO middleware https://review.openstack.org/499260 | 13:29 |
openstackgerrit | Joel Wright proposed openstack/swift master: WIP: Add TLO middleware https://review.openstack.org/499260 | 13:36 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/python-swiftclient master: Updated from global requirements https://review.openstack.org/89250 | 13:39 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/swift master: Updated from global requirements https://review.openstack.org/88736 | 13:40 |
*** bigdogstl has joined #openstack-swift | 13:43 | |
*** mat128 has joined #openstack-swift | 13:46 | |
*** bigdogstl has quit IRC | 13:47 | |
*** lucasxu has quit IRC | 13:54 | |
*** lucasxu has joined #openstack-swift | 14:03 | |
*** saint_ has joined #openstack-swift | 14:11 | |
*** bigdogstl has joined #openstack-swift | 14:12 | |
*** bigdogstl has quit IRC | 14:16 | |
*** ukaynar has joined #openstack-swift | 14:17 | |
*** itlinux has joined #openstack-swift | 14:23 | |
*** itlinux has quit IRC | 14:23 | |
*** openstackgerrit has quit IRC | 14:33 | |
*** vint_bra has joined #openstack-swift | 14:35 | |
*** itlinux has joined #openstack-swift | 14:41 | |
*** itlinux has quit IRC | 14:43 | |
*** silor has joined #openstack-swift | 14:49 | |
*** _ix has joined #openstack-swift | 14:59 | |
*** links has quit IRC | 15:07 | |
*** bigdogstl has joined #openstack-swift | 15:12 | |
*** bigdogstl has quit IRC | 15:17 | |
*** bigdogstl has joined #openstack-swift | 15:25 | |
*** zaitcev has joined #openstack-swift | 15:25 | |
*** ChanServ sets mode: +v zaitcev | 15:25 | |
*** klrmn has joined #openstack-swift | 15:28 | |
*** bigdogstl has quit IRC | 15:30 | |
*** bigdogstl has joined #openstack-swift | 15:34 | |
*** gyee has joined #openstack-swift | 15:37 | |
*** bigdogstl has quit IRC | 15:39 | |
*** oshritf_ has joined #openstack-swift | 15:43 | |
*** openstackgerrit has joined #openstack-swift | 15:43 | |
openstackgerrit | Romain LE DISEZ proposed openstack/swift master: WIP: Configure diskfile per storage policy https://review.openstack.org/447129 | 15:43 |
*** bigdogstl has joined #openstack-swift | 15:52 | |
*** ukaynar has quit IRC | 15:53 | |
*** ukaynar has joined #openstack-swift | 15:53 | |
*** bigdogstl has quit IRC | 15:58 | |
*** saint_ has quit IRC | 16:00 | |
*** saint_ has joined #openstack-swift | 16:01 | |
openstackgerrit | Merged openstack/swift master: api-ref: Fix container PUT response codes https://review.openstack.org/502483 | 16:11 |
*** ukaynar has quit IRC | 16:11 | |
*** bigdogstl has joined #openstack-swift | 16:14 | |
*** cbartz has quit IRC | 16:14 | |
*** oshritf_ has quit IRC | 16:14 | |
*** bigdogstl has quit IRC | 16:20 | |
timburke | good morning | 16:29 |
timburke | thanks for cleaning up p 505858 for me mattoliverau! | 16:30 |
patchbot | https://review.openstack.org/#/c/505858/ - swift (feature/deep) - Bring us closer to master (MERGED) | 16:30 |
timburke | and for doing p 505976 | 16:32 |
patchbot | https://review.openstack.org/#/c/505976/ - swift (feature/deep) - Merge Shard-{Account, Container} to 1 metadata | 16:32 |
timburke | i was just thinking about taking that work on :-) | 16:32 |
*** bigdogstl has joined #openstack-swift | 16:33 | |
timburke | it makes me wonder if X-Container-Sysmeta-Shard-Timestamp is strictly necessary... or if we could just use the ts associated with X-Container-Sysmeta-Shard-Root | 16:33 |
*** chsc has joined #openstack-swift | 16:39 | |
*** chsc has joined #openstack-swift | 16:39 | |
openstackgerrit | Alistair Coles proposed openstack/swift feature/deep: sharding: make find_shard_ranges a broker method https://review.openstack.org/505763 | 16:39 |
*** geaaru has quit IRC | 16:40 | |
acoles | mattoliverau: timburke ^^ I hope that is on right track | 16:41 |
*** bigdogstl has quit IRC | 16:42 | |
*** itlinux has joined #openstack-swift | 16:45 | |
*** tesseract has quit IRC | 16:50 | |
*** rcernin has quit IRC | 16:52 | |
acoles | timburke: mattoliverau: got consumed by that ^^ today, I'll try to land some more of the feature branch patches tomorrow | 16:55 |
*** mvk has quit IRC | 17:01 | |
*** _ix has quit IRC | 17:05 | |
*** ukaynar has joined #openstack-swift | 17:07 | |
*** bigdogstl has joined #openstack-swift | 17:12 | |
*** bigdogstl has quit IRC | 17:17 | |
*** itlinux has quit IRC | 17:18 | |
*** mat128 has quit IRC | 17:32 | |
notmyname | good morning | 17:36 |
*** mvk has joined #openstack-swift | 17:41 | |
*** mat128 has joined #openstack-swift | 17:42 | |
*** mat128 has quit IRC | 17:47 | |
*** mat128 has joined #openstack-swift | 17:49 | |
*** ianychoi has quit IRC | 17:57 | |
*** Sukhdev has joined #openstack-swift | 17:57 | |
*** ianychoi has joined #openstack-swift | 17:58 | |
*** mat128 has quit IRC | 18:02 | |
*** Sukhdev has quit IRC | 18:03 | |
*** bigdogstl has joined #openstack-swift | 18:06 | |
*** bigdogstl has quit IRC | 18:14 | |
*** mat128 has joined #openstack-swift | 18:16 | |
-openstackstatus- NOTICE: Gerrit OpenIDs have been accidentally overwritten and are in the process of being restored | 18:18 | |
*** vint_bra has quit IRC | 18:19 | |
*** vint_bra has joined #openstack-swift | 18:20 | |
*** vint_bra has quit IRC | 18:20 | |
*** vint_bra has joined #openstack-swift | 18:20 | |
-openstackstatus- NOTICE: The Gerrit service on review.openstack.org is being taken offline briefly to perform database repair work but should be back up shortly | 18:24 | |
*** ChanServ changes topic to "The Gerrit service on review.openstack.org is being taken offline briefly to perform database repair work but should be back up shortly" | 18:24 | |
*** vint_bra has quit IRC | 18:26 | |
*** vint_bra has joined #openstack-swift | 18:26 | |
*** ChubYann has joined #openstack-swift | 18:30 | |
*** silor has quit IRC | 18:34 | |
*** silor has joined #openstack-swift | 18:34 | |
*** silor has quit IRC | 18:51 | |
*** _ix has joined #openstack-swift | 19:06 | |
*** bigdogstl has joined #openstack-swift | 19:11 | |
*** bigdogstl has quit IRC | 19:16 | |
*** ChanServ changes topic to "Queens PTG planning https://etherpad.openstack.org/p/swift-ptg-queens | Logs: http://eavesdrop.openstack.org/irclogs/%23openstack-swift/ | Meetings: https://wiki.openstack.org/wiki/Meetings/Swift | Review Dashboard: http://not.mn/reviews.html" | 19:22 | |
-openstackstatus- NOTICE: OpenIDs for the Gerrit service have been restored from a recent backup and the service is running again; before/after table states are being analyzed now to identify any remaining cleanup needed for changes made to accounts today | 19:22 | |
*** pcaruana has quit IRC | 19:41 | |
openstackgerrit | Merged openstack/swift feature/deep: Merge Shard-{Account, Container} to 1 metadata https://review.openstack.org/505976 | 19:45 |
*** mat128 has quit IRC | 19:54 | |
*** mvk has quit IRC | 20:05 | |
*** mvk has joined #openstack-swift | 20:17 | |
*** skudlik has quit IRC | 21:03 | |
*** lucasxu has quit IRC | 21:04 | |
*** catintheroof has quit IRC | 21:07 | |
*** bigdogstl has joined #openstack-swift | 21:13 | |
timburke | huh. our PipeMutex pretty consistently pollutes my test output when running test_utils in isolation :-/ http://paste.openstack.org/show/621660/ | 21:16 |
*** bigdogstl has quit IRC | 21:18 | |
*** Sukhdev has joined #openstack-swift | 21:28 | |
*** _ix has quit IRC | 21:30 | |
timburke | also weird: the new tests in https://review.openstack.org/#/c/506002/ run fine on their own, but fail like http://logs.openstack.org/02/506002/1/check/gate-swift-python27-ubuntu-xenial/c989f80/console.html#_2017-09-21_06_42_08_854517 if we run common/test_daemon.py:TestRunDaemon.test_run_daemon first | 21:30 |
patchbot | patch 506002 - swift (feature/deep) - Use temp db and a rename for set_sharding_state | 21:30 |
timburke | this isn't the first time that test has given me trouble: http://eavesdrop.openstack.org/irclogs/%23openstack-swift/%23openstack-swift.2017-07-13.log.html | 21:31 |
timburke | but this time makes so much less sense :-( | 21:31 |
*** nottrobin has quit IRC | 21:46 | |
*** kmARC has quit IRC | 21:46 | |
*** dr_gogeta86 has quit IRC | 21:48 | |
*** dja has quit IRC | 21:48 | |
*** kmARC has joined #openstack-swift | 21:48 | |
*** nottrobin has joined #openstack-swift | 21:49 | |
*** tanee has quit IRC | 21:49 | |
*** hugokuo has quit IRC | 21:49 | |
*** MVenesio has quit IRC | 21:49 | |
*** tanee has joined #openstack-swift | 21:50 | |
*** hugokuo has joined #openstack-swift | 21:50 | |
*** dr_gogeta86 has joined #openstack-swift | 21:53 | |
*** dr_gogeta86 has quit IRC | 21:53 | |
*** dr_gogeta86 has joined #openstack-swift | 21:53 | |
*** dja has joined #openstack-swift | 21:56 | |
*** _ix has joined #openstack-swift | 22:06 | |
*** bigdogstl has joined #openstack-swift | 22:16 | |
*** bigdogstl has quit IRC | 22:23 | |
openstackgerrit | Tim Burke proposed openstack/swift master: Stop reloading swift.common.utils in test_daemon https://review.openstack.org/506410 | 22:31 |
*** _ix has quit IRC | 22:39 | |
*** bigdogstl has joined #openstack-swift | 22:53 | |
*** ukaynar has quit IRC | 22:54 | |
*** Sukhdev has quit IRC | 22:57 | |
*** bigdogstl has quit IRC | 23:00 | |
*** chsc has quit IRC | 23:05 | |
*** bigdogstl has joined #openstack-swift | 23:06 | |
*** bigdogstl has quit IRC | 23:11 | |
timburke | so thinking about https://review.openstack.org/#/c/505321/3/swift/container/sharder.py@597 -- it doesn't look like we cache get_part_nodes in the other place we call roundrobin_datadirs -- how expensive is get_part_nodes, really? | 23:27 |
patchbot | patch 505321 - swift (feature/deep) - remove node-id, node_idx and part instance vars in... (MERGED) | 23:27 |
openstackgerrit | Tim Burke proposed openstack/swift feature/deep: Address some lingering review comments https://review.openstack.org/506367 | 23:29 |
*** kei_yama has joined #openstack-swift | 23:35 | |
*** bigdogstl has joined #openstack-swift | 23:41 | |
*** ukaynar has joined #openstack-swift | 23:46 | |
*** bigdogstl has quit IRC | 23:50 | |
*** chsc has joined #openstack-swift | 23:50 | |
*** chsc has joined #openstack-swift | 23:50 | |
briancline | anyone ever seen a ring hashing collision for two distinct objects? | 23:55 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!