| *** jamielennox|away is now known as jamielennox | 00:02 | |
| *** lucasxu has joined #openstack-swift | 00:46 | |
| *** lucasxu has quit IRC | 00:46 | |
| *** zhurong has joined #openstack-swift | 00:49 | |
| *** zhengyin has joined #openstack-swift | 00:58 | |
| *** jamielennox is now known as jamielennox|away | 01:00 | |
| *** jamielennox|away is now known as jamielennox | 01:10 | |
| kota_ | good morning! | 01:14 |
|---|---|---|
| kota_ | yey, https://review.openstack.org/439572 landed | 01:15 |
| patchbot | patch 439572 - swift - Limit number of revert tombstone SSYNC requests (MERGED) | 01:15 |
| kota_ | thanks clayg and mahatic | 01:15 |
| charz | clayg: thanks! checking | 01:16 |
| *** zhurong has quit IRC | 01:29 | |
| *** mingyu has joined #openstack-swift | 02:03 | |
| *** mingyu has quit IRC | 02:04 | |
| *** zhurong has joined #openstack-swift | 02:04 | |
| charz | Looks it passed. | 02:04 |
| *** mingyu has joined #openstack-swift | 02:12 | |
| *** mingyu has quit IRC | 02:48 | |
| *** mingyu has joined #openstack-swift | 02:50 | |
| *** mingyu has quit IRC | 02:51 | |
| *** mingyu has joined #openstack-swift | 02:53 | |
| *** mingyu has quit IRC | 03:09 | |
| *** mingyu has joined #openstack-swift | 03:35 | |
| *** gkadam has joined #openstack-swift | 03:38 | |
| mahatic | good morning | 04:15 |
| *** SkyRocknRoll_ has joined #openstack-swift | 04:15 | |
| mahatic | yay patch 439572 landed | 04:15 |
| patchbot | https://review.openstack.org/#/c/439572/ - swift - Limit number of revert tombstone SSYNC requests (MERGED) | 04:15 |
| *** SkyRocknRoll has quit IRC | 04:17 | |
| *** mingyu has quit IRC | 04:19 | |
| *** links has joined #openstack-swift | 04:20 | |
| *** zhurong has quit IRC | 04:25 | |
| *** psachin has joined #openstack-swift | 04:31 | |
| *** zhurong has joined #openstack-swift | 04:34 | |
| *** zhengyin has quit IRC | 04:35 | |
| *** aselius has quit IRC | 04:36 | |
| *** zhengyin has joined #openstack-swift | 04:40 | |
| *** gyee has quit IRC | 04:49 | |
| *** gabor_antal has quit IRC | 05:07 | |
| *** skudlik has joined #openstack-swift | 05:09 | |
| *** links has quit IRC | 05:10 | |
| *** aselius_ has joined #openstack-swift | 05:12 | |
| *** links has joined #openstack-swift | 05:25 | |
| *** SkyRocknRoll_ has quit IRC | 05:43 | |
| *** cshastri has joined #openstack-swift | 05:48 | |
| *** klrmn has quit IRC | 06:02 | |
| *** ChubYann has quit IRC | 06:05 | |
| *** zhengyin has quit IRC | 06:08 | |
| *** zhengyin has joined #openstack-swift | 06:09 | |
| *** jaosorior has joined #openstack-swift | 06:18 | |
| *** zhengyin has quit IRC | 06:27 | |
| *** zhengyin has joined #openstack-swift | 06:28 | |
| *** zhengyin has quit IRC | 06:32 | |
| *** zhengyin has joined #openstack-swift | 06:34 | |
| *** cbartz has joined #openstack-swift | 06:37 | |
| *** zhengyin has quit IRC | 06:40 | |
| openstackgerrit | Pallavi proposed openstack/swift master: Fix html_last_updated_fmt for Python3. https://review.openstack.org/472534 | 06:40 |
| *** zhengyin has joined #openstack-swift | 06:40 | |
| *** zhengyin has quit IRC | 06:47 | |
| *** zhengyin has joined #openstack-swift | 06:47 | |
| *** hseipp has joined #openstack-swift | 06:48 | |
| *** geaaru has joined #openstack-swift | 06:49 | |
| *** hseipp has quit IRC | 06:51 | |
| *** hseipp has joined #openstack-swift | 06:51 | |
| *** zhengyin has quit IRC | 06:52 | |
| *** zhengyin has joined #openstack-swift | 06:52 | |
| *** tesseract has joined #openstack-swift | 06:53 | |
| openstackgerrit | Timur Alperovich proposed openstack/python-swiftclient master: Default to binary/octet-stream. https://review.openstack.org/472541 | 06:54 |
| *** zhurong has quit IRC | 06:55 | |
| *** zhurong has joined #openstack-swift | 06:56 | |
| *** zhurong has quit IRC | 06:57 | |
| *** pcaruana has joined #openstack-swift | 07:05 | |
| *** early has quit IRC | 07:07 | |
| *** links has quit IRC | 07:09 | |
| *** rcernin has joined #openstack-swift | 07:09 | |
| *** early has joined #openstack-swift | 07:12 | |
| *** aselius_ has quit IRC | 07:21 | |
| *** links has joined #openstack-swift | 07:25 | |
| *** kei_yama has quit IRC | 07:37 | |
| *** kei_yama has joined #openstack-swift | 07:45 | |
| *** SkyRocknRoll has joined #openstack-swift | 07:58 | |
| *** oshritf has joined #openstack-swift | 08:13 | |
| *** qwerty has joined #openstack-swift | 08:21 | |
| *** qwerty has quit IRC | 08:38 | |
| *** qwerty has joined #openstack-swift | 08:41 | |
| *** oshritf has quit IRC | 08:48 | |
| *** qwerty has quit IRC | 08:55 | |
| *** Dw_Sn has joined #openstack-swift | 09:23 | |
| *** qwerty has joined #openstack-swift | 09:27 | |
| *** qwerty has quit IRC | 09:33 | |
| openstackgerrit | Hieu LE proposed openstack/swift master: OSprofiler in OpenStack Swift https://review.openstack.org/468316 | 09:41 |
| openstackgerrit | Hieu LE proposed openstack/swift master: OSprofiler in OpenStack Swift https://review.openstack.org/468316 | 09:42 |
| openstackgerrit | Lingyong Xu proposed openstack/python-swiftclient master: Drop py34 target in tox.ini https://review.openstack.org/471685 | 09:43 |
| *** links has quit IRC | 09:51 | |
| *** zhurong has joined #openstack-swift | 09:55 | |
| *** links has joined #openstack-swift | 10:03 | |
| *** zhurong has quit IRC | 10:06 | |
| openstackgerrit | Lingyong Xu proposed openstack/swift master: Fix html_last_updated_fmt for Python3 https://review.openstack.org/472621 | 10:09 |
| *** mat128 has joined #openstack-swift | 11:08 | |
| *** psachin has quit IRC | 11:13 | |
| *** zhengyin has quit IRC | 11:33 | |
| *** SkyRocknRoll has quit IRC | 11:51 | |
| *** SkyRocknRoll has joined #openstack-swift | 11:55 | |
| rledisez | clayg: agree that 10 devices per worker for 24 devices should give 8 devices per worker. seems reasonable. I don't see any reason someone would want 10+10+4, because the attribution may be "unpredictable" (well, it's not random, but you wouldn't be able do a real attribution) | 12:05 |
| rledisez | but i agree that everywhere concurrency is configured in swift, it's about how many workers/thread/… to run. so it might be confusing for the operator that this particular item is not that way, but the opposite (we calculate the concurrency) | 12:07 |
| *** kei_yama has quit IRC | 12:17 | |
| rledisez | acoles: I had a look at https://bugs.launchpad.net/swift/+bug/1652323 , it's actually quite simple (too simple to be real?). i should write some new tests, but at least existing tests didn't break. i'll have a look on monday for tests. in the mean time i'm pushing the patch (see below) | 12:29 |
| openstack | Launchpad bug 1652323 in OpenStack Object Storage (swift) "ssync syncs an expired object as a tombstone" [Medium,Confirmed] | 12:29 |
| openstackgerrit | Romain LE DISEZ proposed openstack/swift master: Allow to rebuild a fragment of an expired object https://review.openstack.org/472659 | 12:29 |
| *** jaosorior has quit IRC | 12:37 | |
| *** blair has quit IRC | 12:39 | |
| *** catintheroof has joined #openstack-swift | 12:40 | |
| *** catintheroof has quit IRC | 12:45 | |
| *** catintheroof has joined #openstack-swift | 12:46 | |
| *** links has quit IRC | 12:48 | |
| *** SkyRocknRoll has quit IRC | 12:55 | |
| *** joeljwright has joined #openstack-swift | 13:02 | |
| *** ChanServ sets mode: +v joeljwright | 13:02 | |
| *** lucasxu has joined #openstack-swift | 13:03 | |
| *** SkyRocknRoll has joined #openstack-swift | 13:03 | |
| *** blair has joined #openstack-swift | 13:10 | |
| *** lucasxu has quit IRC | 13:16 | |
| *** ujjain has quit IRC | 13:21 | |
| *** klamath has joined #openstack-swift | 13:30 | |
| *** klamath has quit IRC | 13:31 | |
| *** klamath has joined #openstack-swift | 13:32 | |
| *** klamath has quit IRC | 13:35 | |
| *** lcurtis has joined #openstack-swift | 13:45 | |
| *** chlong has joined #openstack-swift | 13:51 | |
| *** jaosorior has joined #openstack-swift | 13:56 | |
| *** lucasxu has joined #openstack-swift | 14:09 | |
| lcurtis | hello all...running into some sever performance problems on container servers | 14:10 |
| lcurtis | could someone tell me the algorithm proxy server uses for choosing a container server? | 14:11 |
| lcurtis | or maybe some tweaks for faster container response? | 14:12 |
| lcurtis | using ssds | 14:12 |
| *** lucasxu has quit IRC | 14:14 | |
| *** lucasxu has joined #openstack-swift | 14:14 | |
| *** lucasxu has quit IRC | 14:16 | |
| *** tesseract has quit IRC | 14:16 | |
| *** aselius_ has joined #openstack-swift | 14:19 | |
| *** lucasxu has joined #openstack-swift | 14:20 | |
| *** tesseract has joined #openstack-swift | 14:23 | |
| *** lucasxu has quit IRC | 14:23 | |
| *** SkyRocknRoll has quit IRC | 14:25 | |
| *** gkadam_ has joined #openstack-swift | 14:26 | |
| *** gkadam has quit IRC | 14:26 | |
| *** SkyRocknRoll has joined #openstack-swift | 14:28 | |
| *** oshritf has joined #openstack-swift | 14:29 | |
| *** cshastri has quit IRC | 14:30 | |
| *** gkadam_ has quit IRC | 14:31 | |
| *** Dinesh_Bhor has quit IRC | 14:44 | |
| *** oshritf has quit IRC | 14:52 | |
| -openstackstatus- NOTICE: The Gerrit service on review.openstack.org is being restarted now to clear an issue arising from an unanticipated SSH API connection flood | 14:58 | |
| *** oshritf has joined #openstack-swift | 15:04 | |
| *** oshritf has quit IRC | 15:07 | |
| *** rcernin has quit IRC | 15:08 | |
| *** jaosorior has quit IRC | 15:09 | |
| Dw_Sn | timburke: would you please leave me an answer regarding how can I use the swift3 middleware for pre-signed instead of tempurl ? I have the module enabled, but can't find any docs on how to sign and generate the AWS S3 look alike URL .. | 15:09 |
| *** oshritf has joined #openstack-swift | 15:15 | |
| *** oshritf has quit IRC | 15:17 | |
| *** oshritf has joined #openstack-swift | 15:17 | |
| *** oshritf has quit IRC | 15:25 | |
| *** aselius_ has quit IRC | 15:26 | |
| *** aselius has joined #openstack-swift | 15:26 | |
| *** oshritf has joined #openstack-swift | 15:28 | |
| *** oshritf has quit IRC | 15:30 | |
| clayg | lcurtis: what makes you think the issue is container server? | 15:36 |
| openstackgerrit | Gábor Antal proposed openstack/swift master: Use more specific asserts in test/unit/common https://review.openstack.org/342781 | 15:37 |
| clayg | lcurtis: generally speaking on object PUT the container server selection is deterministic by container name. | 15:37 |
| lcurtis | clayg: getting timeouts and update failures, tons of async pendings | 15:39 |
| lcurtis | definitely containers | 15:39 |
| lcurtis | async pending in the billions | 15:39 |
| *** itlinux has joined #openstack-swift | 15:40 | |
| *** chsc has joined #openstack-swift | 15:40 | |
| timburke | Dw_Sn: it works the same as is described at http://docs.aws.amazon.com/AmazonS3/latest/API/sigv4-query-string-auth.html -- the region name used in the credential needs to match the location configured in the proxy-server.conf, but aside from that, there isn't much special; <your-access-key-id> and <SecretAccessKey> come out of keystone (or whatever auth system) | 15:41 |
| timburke | fwiw, i've used boto to generate such urls before: https://github.com/tipabu/swift3-demo/blob/master/boto_demo.py#L16 | 15:42 |
| *** oshritf has joined #openstack-swift | 15:46 | |
| *** oshritf has quit IRC | 15:48 | |
| *** cbartz has quit IRC | 15:48 | |
| *** oshritf has joined #openstack-swift | 15:49 | |
| Dw_Sn | timburke: cool ! I will give it a try next week ! | 15:50 |
| Dw_Sn | thanks a lot :) | 15:50 |
| timburke | sure! sorry it took so long to unblock you | 15:51 |
| *** oshritf has quit IRC | 15:52 | |
| Dw_Sn | timburke: no worries at all, I will look into it, not sure it will be the exact same as I expected but will see :) | 15:52 |
| *** gyee has joined #openstack-swift | 15:53 | |
| Dw_Sn | also I don't think I have location in my proxy-server.conf | 15:53 |
| timburke | iirc it defaults to "US" | 15:53 |
| *** SkyRocknRoll_ has joined #openstack-swift | 15:53 | |
| Dw_Sn | timburke: cool ! thanks again, I will read about if changing a location now might mess up anything or not .. and read the link you sent :) | 15:55 |
| *** SkyRocknRoll has quit IRC | 15:55 | |
| *** Dw_Sn has quit IRC | 15:58 | |
| *** oshritf has joined #openstack-swift | 16:02 | |
| notmyname | good morning | 16:04 |
| *** oshritf has quit IRC | 16:06 | |
| *** chsc has quit IRC | 16:06 | |
| *** oshritf has joined #openstack-swift | 16:08 | |
| *** itlinux has quit IRC | 16:09 | |
| *** oshritf has quit IRC | 16:10 | |
| *** klrmn has joined #openstack-swift | 16:12 | |
| clayg | lcurtis: do you have a container server offline? Are the update timeouts/error all pointing to the same node/ip/disk/part? | 16:13 |
| *** oshritf has joined #openstack-swift | 16:14 | |
| *** klrmn has quit IRC | 16:14 | |
| lcurtis | clayg: no....numerous containers | 16:14 |
| lcurtis | all container servers under heavy load | 16:14 |
| clayg | on the container servers are the updates coming from the container updaters? | 16:15 |
| *** oshritf has quit IRC | 16:15 | |
| *** joeljwright has quit IRC | 16:16 | |
| notmyname | lcurtis: what version of swift? | 16:17 |
| lcurtis | clayg: seeing this on object servers | 16:17 |
| lcurtis | we have proxy and object on same server | 16:17 |
| lcurtis | containers separate | 16:17 |
| clayg | "this" == ???? | 16:17 |
| lcurtis | sorry...the timeouts i mentioned | 16:18 |
| lcurtis | updates and heads | 16:18 |
| lcurtis | version is 2.2.0-0ubuntu1~cloud0 | 16:18 |
| lcurtis | which is prob quite old | 16:19 |
| *** ChubYann has joined #openstack-swift | 16:20 | |
| *** oshritf has joined #openstack-swift | 16:21 | |
| clayg | ok, so the proxy's are trying to HEAD containers for the existence check cache - the object servers are trying to update the containers with either new PUT/DELETE (or the existing async pending updates via the swift-object-updater which I called the "container updater" earlier because it updates the containers) - and both (all three) processes are on the same physical server and throwing timeouts talking to container | 16:21 |
| clayg | servers. | 16:21 |
| clayg | ... which you said were under heavy load | 16:21 |
| clayg | what do things look like from the container-servers perspective? | 16:22 |
| *** oshritf has quit IRC | 16:22 | |
| clayg | can you get a sense if the majority of updates are coming from the proxy-server, object-server or object-updater? The requests should say in the user-agent line... | 16:22 |
| *** chsc has joined #openstack-swift | 16:24 | |
| lcurtis | we are trying to add more nodes into container layer | 16:25 |
| lcurtis | so replication also under heavy load on container servers | 16:25 |
| lcurtis | we may just need to wait it out | 16:25 |
| *** hseipp has quit IRC | 16:27 | |
| *** vint_bra has joined #openstack-swift | 16:32 | |
| *** tesseract has quit IRC | 16:32 | |
| *** pcaruana has quit IRC | 16:34 | |
| *** SkyRocknRoll_ has quit IRC | 16:37 | |
| *** oshritf has joined #openstack-swift | 16:38 | |
| clayg | maybe, you might slow things down overall trying to do too many things at once. Maybe stop the object-updater until the replicators finish? Still not clear if the load is on disks or cpu. Good luck! | 16:38 |
| *** oshritf has quit IRC | 16:40 | |
| *** SkyRocknRoll has joined #openstack-swift | 16:40 | |
| *** rcernin has joined #openstack-swift | 16:45 | |
| *** honga has joined #openstack-swift | 16:46 | |
| lcurtis | thanks clayg | 16:47 |
| honga | hello, is there a way I can view the contents of a dlo/slo manifest file instead of downloading the actual file? | 16:47 |
| lcurtis | load definitely on the ssds | 16:47 |
| lcurtis | high iowait | 16:47 |
| *** oshritf has joined #openstack-swift | 16:47 | |
| notmyname | lcurtis: do you have a good sense of how many object write (PUT/DELETE) operations per second each container is getting? how many objects are in each container? | 16:48 |
| notmyname | honga: add "?multipart-manifest=get" to the end of the GET request | 16:48 |
| *** oshritf has quit IRC | 16:49 | |
| honga | doh, it was right in the docs :P thanks notmyname | 16:49 |
| lcurtis | notmyname: what would be best way to check ops / sec on containers? | 16:52 |
| *** oshritf has joined #openstack-swift | 16:53 | |
| notmyname | the best way? track the logs, feed it to some analysis, group by container, split by verb, watch the graphs | 16:53 |
| notmyname | lcurtis: the fastest way? grab some logs, count and divide by the seconds of logs you grabbed | 16:54 |
| lcurtis | i am told about 10k objects per container | 16:56 |
| notmyname | lcurtis: but where I'm going with this is checking to see if you're in the 10/sec, 100/sec, or 1000/sec range, and then compare that against how many objects are in the container | 16:56 |
| lcurtis | yes...checking | 16:56 |
| lcurtis | thanks much | 16:56 |
| notmyname | 10k objects in a container isn't much | 16:59 |
| notmyname | however, if the usage pattern on the container is very ... idk ... "fluid"? ie instead of just adding stuff you're also deleting stuff | 17:00 |
| notmyname | eg you have 10k objects, but you're adding 1 million per day and deleting 1 million per day | 17:00 |
| notmyname | that could end up with larger on-disk container DBs, which IIRC would result in longer lock times for sqlite, which results in more lock errors and timeouts and therefore async pendings | 17:01 |
| notmyname | (that's based on my assumption that a non-vacuumed container DB that is large on-disk is similarly slow to acquire a lock or do a DB index update as a DB that's big because it's got a lot of objects in it) | 17:02 |
| *** klrmn has joined #openstack-swift | 17:10 | |
| openstackgerrit | Timur Alperovich proposed openstack/python-swiftclient master: Default to binary/octet-stream. https://review.openstack.org/472541 | 17:17 |
| *** vint_bra has quit IRC | 17:17 | |
| *** tonanhngo has joined #openstack-swift | 17:21 | |
| *** vint_bra has joined #openstack-swift | 17:22 | |
| timburke | notmyname: there are a couple other complications with "fluid" containers -- we've gotta keep the deleted rows around for reclaim_age (which i'm fairly certain *must* have the same characteristics as a "normal" container with a lot of objects in it) and when we *do* call reclaim, we're doing a scan over *all* deleted objects -- there's no deleted, created_at index for containers (nor a deleted, delete_timestamp index for accounts) | 17:30 |
| notmyname | yep | 17:30 |
| notmyname | all of that. plus I was thinking about our customer who has large on-disk DBs even after the reclaim age, just because the DBs aren't vacuumed | 17:31 |
| notmyname | I think my average words-per-email that I send to the openstack ML must be measured in the 1000s :-( | 17:34 |
| * notmyname is currently working on the email about patch 468105 | 17:35 | |
| patchbot | https://review.openstack.org/#/c/468105/ - swift - Require that known-bad EC schemes be deprecated | 17:35 |
| notmyname | it's going to be long | 17:35 |
| timburke | oh yeah, speaking of MLs -- notmyname, should we do a swiftclient release in the nearish future? | 17:35 |
| notmyname | let me check my helper tool... | 17:38 |
| *** mat128 has quit IRC | 17:39 | |
| notmyname | https://gist.github.com/notmyname/e2b5292bad2891ed4b67f3b80780739a | 17:39 |
| *** mat128 has joined #openstack-swift | 17:41 | |
| *** mat128 has quit IRC | 17:41 | |
| *** oshritf has quit IRC | 17:51 | |
| *** vint_bra has quit IRC | 18:11 | |
| *** vint_bra has joined #openstack-swift | 18:13 | |
| *** geaaru has quit IRC | 18:15 | |
| clayg | why did we put 'workers' in [DEFAULT] | 18:26 |
| clayg | do we have freedom!? | 18:26 |
| notmyname | http://www.alexmaximo.com/wp-content/uploads/2016/02/braveheart-freedom-400x250.jpg | 18:42 |
| tdasilva | clayg: or free doom? | 18:42 |
| notmyname | https://upload.wikimedia.org/wikipedia/en/c/c8/Doom_gibs.png ? | 18:44 |
| clayg | i think you are all mocking me | 18:44 |
| clayg | but srly, I can't name the reconstructors workers config option workers :\ | 18:45 |
| notmyname | clayg: ah. that sthinks | 18:45 |
| notmyname | sorry, I just spend a lot of time concentrating, just finished, and now need to let off steam | 18:45 |
| notmyname | and I type with a lithp? | 18:45 |
| clayg | ok | 18:46 |
| * notmyname should probably stand up and walk around instead of make comments in irc | 18:46 | |
| clayg | i could either call the option "processes" stealing from the expirer (where processes means soemthing else entirely) or I could require "use_workers = true" + "set workers = X" config syntax | 18:49 |
| * clayg shrugs | 18:50 | |
| clayg | i'll use "reconstructor_workers" for now and sort it out during review | 18:50 |
| notmyname | I like "reconstructor_workers" | 18:51 |
| *** itlinux_ has joined #openstack-swift | 19:27 | |
| tdasilva | timburke: about swift3 acls, my understanding is that swift3 has some acls support correct? what is missing is just the object-level ACL support? | 20:29 |
| *** itlinux_ has quit IRC | 20:29 | |
| timburke | tdasilva: we've even got some support for that -- it's just much more wasteful. gotta go start reading the object, get some (sys?)meta to the proxy, then decide whether or not to allow access | 20:30 |
| timburke | i honestly don't remember the full state of acls in swift3. i know we don't pass all the ceph s3 tests about them, but as i recall we passed more than i was expecting | 20:31 |
| tdasilva | timburke: right, that's what I was going to say next, I remember IBM??? was working on that piece, where you would have to essentially HEAD the object | 20:31 |
| tdasilva | timburke: ah ok, cool | 20:31 |
| tdasilva | timburke: are the ceph s3 tests results published anywhere? | 20:31 |
| timburke | might be worth grepping for acl in https://github.com/openstack/swift3/blob/master/swift3/test/functional/conf/ceph-known-failures-tempauth.yaml ... | 20:32 |
| timburke | a lot of that just isn't going to happen -- we don't have a way to look up users, for example, so you can't really do user existence checks or grants by email | 20:33 |
| timburke | other parts are tricky for other reasons -- all the "raw" stuff really means "unauthenticated" and we don't (currently) have a way to know what account to use | 20:34 |
| tdasilva | timburke: so that's a list of known failures? is there a list of what's known to pass? :P or do i just have to collate the data? | 20:35 |
| tdasilva | timburke: ahh,,that's the public container thing, right? | 20:35 |
| timburke | one of those lists is much longer than the other :-) | 20:36 |
| tdasilva | S3 having global namespace, you can just make container public and don't really have the concept of an account | 20:36 |
| timburke | i felt like it was better to carve out the exclusions | 20:36 |
| timburke | yup | 20:36 |
| *** honga has quit IRC | 20:36 | |
| timburke | i've got a thought about making a .swift3-buckets account or somesuch but haven't found the time... | 20:36 |
| timburke | and migrating existing buckets will be...interesting | 20:37 |
| tdasilva | he | 20:37 |
| *** itlinux_ has joined #openstack-swift | 20:39 | |
| *** SkyRocknRoll has quit IRC | 21:10 | |
| clayg | notmyname: oh and maybe late Friday afternoon isn't the best time to send the ec warning - could wait till Monday morning... | 21:14 |
| notmyname | clayg: yep. that's exactly what I'm doing | 21:14 |
| notmyname | also wanting to make sure rledisez thinks it's good from an ops perspective, and acoles and cschwede in EU | 21:14 |
| clayg | righto! | 21:23 |
| *** chlong has quit IRC | 21:58 | |
| *** itlinux_ has quit IRC | 22:06 | |
| *** skudlik has quit IRC | 22:11 | |
| *** rcernin has quit IRC | 22:19 | |
| lcurtis | hmmm..an i correct in stating replication happens source to destination? | 22:27 |
| lcurtis | i think issue is that i weighted down 2 container servers at once to replace ssds | 22:27 |
| lcurtis | container layer under heavy load | 22:27 |
| lcurtis | and now trying to weight these container servers back in | 22:28 |
| lcurtis | they are under no load, but rest of container servers are | 22:28 |
| lcurtis | can i do some kind of reverse rsync? | 22:28 |
| timur | timburke: cschwede I looked into the python3 + requests issue with the Content-Type being set to application/x-www-form-urlencoded and found that the issue is fixed with requests versions > 2.4. I'd like to consider either bumping the requests version in requirements.txt or doing a version check and only sending "Content-Type: " when we detect an older requests library (and python3?) | 22:31 |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Untangle ServersPerPortStrategy.bind_ports https://review.openstack.org/472851 | 22:31 |
| clayg | lcurtis: all swift replication is pushed | 22:32 |
| lcurtis | any utils to do reverse replication so can utilize cpu on nodes being weighted back in? | 22:33 |
| *** vint_bra has quit IRC | 22:35 | |
| *** ^andrea^ has joined #openstack-swift | 22:41 | |
| *** ^_andrea_^ has joined #openstack-swift | 22:41 | |
| *** ^_andrea_^ has quit IRC | 22:41 | |
| *** catintheroof has quit IRC | 22:42 | |
| *** lcurtis has quit IRC | 22:49 | |
| openstackgerrit | Timur Alperovich proposed openstack/python-swiftclient master: Do not set Content-Type to ''. https://review.openstack.org/472541 | 22:50 |
| openstackgerrit | Timur Alperovich proposed openstack/python-swiftclient master: Do not set Content-Type to '' with new requests. https://review.openstack.org/472541 | 22:51 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
