| *** diogogmt has quit IRC | 00:01 | |
| openstackgerrit | John Dickinson proposed openstack/swift: authors and changelog updates for 2.6.0 https://review.openstack.org/269911 | 00:06 |
|---|---|---|
| *** nadeem has quit IRC | 00:18 | |
| *** diazjf has joined #openstack-swift | 00:20 | |
| *** diazjf has quit IRC | 00:20 | |
| *** mrmoje has quit IRC | 00:27 | |
| *** siva_krishnan has quit IRC | 00:27 | |
| *** siva_krishnan has joined #openstack-swift | 00:28 | |
| notmyname | mattoliverau: thanks | 00:29 |
| openstackgerrit | John Dickinson proposed openstack/swift: authors and changelog updates for 2.6.0 https://review.openstack.org/269911 | 00:29 |
| mattoliverau | notmyname: anything to distract me from presentation writing :P | 00:30 |
| notmyname | yeah. I need to stay up tonight to work on that | 00:30 |
| *** vinsh has quit IRC | 00:31 | |
| *** vinsh has joined #openstack-swift | 00:32 | |
| *** siva_krishnan has quit IRC | 00:33 | |
| *** vinsh has quit IRC | 00:36 | |
| *** dmorita has quit IRC | 00:38 | |
| kota_ | good morning | 00:47 |
| *** dmorita has joined #openstack-swift | 00:49 | |
| *** m_kazuhiro has joined #openstack-swift | 00:50 | |
| *** dmorita has quit IRC | 00:54 | |
| *** dmorita has joined #openstack-swift | 00:55 | |
| *** dmorita has quit IRC | 00:55 | |
| *** dmorita has joined #openstack-swift | 00:55 | |
| *** vinsh has joined #openstack-swift | 00:59 | |
| *** siva_krishnan has joined #openstack-swift | 01:11 | |
| mattoliverau | kota_: morning | 01:21 |
| *** david-lyle has quit IRC | 01:31 | |
| *** Jeffrey4l has joined #openstack-swift | 01:31 | |
| *** bill_az has quit IRC | 01:41 | |
| *** haomaiwang has quit IRC | 01:42 | |
| *** vinsh has quit IRC | 01:44 | |
| *** vinsh has joined #openstack-swift | 01:50 | |
| kota_ | notmyname: hmm...eventlet==0.16.1 doesn't work in my envrionment but i got another issue test_get_conns stucked long time. | 01:54 |
| kota_ | notmyname: anyways, nosetests test/unit/common/test_memcached.py with eventlet>=0.17.1 worked for me. | 01:54 |
| kota_ | notmyname:^^ looking at patcch 269811 | 01:54 |
| kota_ | patch 269811 | 01:55 |
| patchbot | kota_: https://review.openstack.org/#/c/269811/ - Bump eventlet min version to 0.17.4 | 01:55 |
| *** vinsh has quit IRC | 01:57 | |
| *** vinsh has joined #openstack-swift | 01:59 | |
| *** vinsh has quit IRC | 02:00 | |
| *** siva_krishnan has quit IRC | 02:02 | |
| *** siva_krishnan has joined #openstack-swift | 02:10 | |
| notmyname | kota_: that means you like that patch to bump eventlet? | 02:13 |
| kota_ | notmyname: yeah, I like the bumping, right now, I am writing my idea for the patch. | 02:15 |
| notmyname | thanks | 02:16 |
| kota_ | done | 02:16 |
| * notmyname will be back online later tonight | 02:17 | |
| kota_ | torgomatic: thanks to trigger my patch 255096 got merged^^ I'm happy it is uneccessary to consider such a annoying failure message w/o my bad | 02:22 |
| patchbot | kota_: https://review.openstack.org/#/c/255096/ - Sleep enough for trampoline (MERGED) | 02:22 |
| *** siva_krishnan has quit IRC | 02:41 | |
| *** nakagawamsa has joined #openstack-swift | 02:42 | |
| *** PsionTheory has quit IRC | 02:47 | |
| *** mingdang1 has joined #openstack-swift | 02:48 | |
| *** sgundur has left #openstack-swift | 02:59 | |
| *** rminmin has joined #openstack-swift | 03:00 | |
| *** haomaiwa_ has joined #openstack-swift | 03:04 | |
| *** janonymous has quit IRC | 03:05 | |
| openstackgerrit | Tim Burke proposed openstack/swift: Stop logging "client disconnected" for 304 / 412 responses https://review.openstack.org/244394 | 03:05 |
| openstackgerrit | Catherine Northcott proposed openstack/swift: Let developers/operators add watchers to account audit https://review.openstack.org/269979 | 03:06 |
| openstackgerrit | Catherine Northcott proposed openstack/swift: Let developers/operators add watchers to account audit https://review.openstack.org/268830 | 03:09 |
| *** ho has joined #openstack-swift | 03:25 | |
| *** lpabon has quit IRC | 03:25 | |
| *** ho is now known as Guest60309 | 03:25 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/174343 | 03:28 |
| *** Guest60309 is now known as ho_ | 03:28 | |
| *** rminmin has quit IRC | 03:29 | |
| mingdang1 | hello,I want to debug the proxy,but I set pdb in .py file,but I run proxy-server in command,the log prompt "raise BdbQuit#012BdbQuit" | 03:34 |
| mingdang1 | the proxy-server can't debug? | 03:34 |
| mingdang1 | @kota_ excuse me,:)... | 03:35 |
| *** venkat has joined #openstack-swift | 03:36 | |
| kota_ | mingdang1: what's up? | 03:41 |
| mingdang1 | I want to debug the proxy-server to learn more detail,but faild | 03:41 |
| mingdang1 | I set pdb in .py file,but I run proxy-server in command,the log prompt "raise BdbQuit#012BdbQuit" | 03:42 |
| kota_ | ah... i'm not sure about pdb but you could see some debug message with running up debu mode swift which configured in /etc/swift/proxy-server.conf | 03:42 |
| kota_ | (note that i suggested another way) | 03:43 |
| *** gyee has quit IRC | 03:44 | |
| kota_ | mingdang1: which command did you use? | 03:44 |
| mingdang1 | I run '/usr/bin/python2 /usr/bin/swift-proxy-server /etc/swift/proxy-server.conf' | 03:45 |
| *** dmorita has quit IRC | 03:46 | |
| *** janonymous has joined #openstack-swift | 03:46 | |
| *** mragupat has joined #openstack-swift | 03:48 | |
| kota_ | seems reasonable though, did it work w/o pdb? | 03:49 |
| mingdang1 | if I don't set pdb in .py,it run ok .but when set pdb,it doesn't run normal | 03:50 |
| mingdang1 | maybe it hava childer process | 03:50 |
| mingdang1 | swift 23557 23241 5 11:47 pts/1 00:00:00 /usr/bin/python2 /usr/bin/swift-proxy-server /etc/swift/proxy-server.conf -v | 03:51 |
| mingdang1 | swift 23566 23557 0 11:47 pts/1 00:00:00 /usr/bin/python2 /usr/bin/swift-proxy-server /etc/swift/proxy-server.conf -v | 03:51 |
| mingdang1 | swift 23567 23557 0 11:47 pts/1 00:00:00 /usr/bin/python2 /usr/bin/swift-proxy-server /etc/swift/proxy-server.conf -v | 03:51 |
| kota_ | ah, yeah. | 03:51 |
| mingdang1 | how to set the process counts? | 03:51 |
| kota_ | "workers" variable in conf file | 03:52 |
| *** mragupat has quit IRC | 03:53 | |
| *** mragupat has joined #openstack-swift | 03:53 | |
| kota_ | mingdang1: sorry i have a meeting in 5 minutes in my company. | 03:56 |
| kota_ | no time to think of, for now. | 03:57 |
| mingdang1 | take you time,ok ,thanks | 03:57 |
| *** links has joined #openstack-swift | 04:06 | |
| *** dmorita has joined #openstack-swift | 04:06 | |
| *** Jeffrey4l has quit IRC | 04:08 | |
| notmyname | hello, world | 04:08 |
| *** dmorita has quit IRC | 04:10 | |
| ho_ | notmyname: good evening! | 04:13 |
| notmyname | hello ho_. how are you? | 04:13 |
| ho_ | notmyname: good :) (but i have a lot of meetings today) | 04:15 |
| notmyname | I know what that's like ;-) | 04:16 |
| ho_ | notmyname: :-) | 04:16 |
| *** mragupat has quit IRC | 04:21 | |
| mattoliverau | notmyname: o/ | 04:29 |
| notmyname | mattoliverau: I'm working on my preso. how about you? | 04:29 |
| notmyname | well, I did get distracted by something else, but I do have Keynote open | 04:30 |
| notmyname | that counts, right? | 04:30 |
| *** klrmn has quit IRC | 04:32 | |
| *** vinsh has joined #openstack-swift | 04:36 | |
| *** vinsh has quit IRC | 04:39 | |
| mahatic | notmyname: mattoliverau o/ | 04:41 |
| *** haomaiwa_ has quit IRC | 04:43 | |
| *** david-lyle has joined #openstack-swift | 04:44 | |
| *** haomaiwa_ has joined #openstack-swift | 04:46 | |
| *** Jeffrey4l has joined #openstack-swift | 04:46 | |
| *** dmorita has joined #openstack-swift | 04:47 | |
| *** Jeffrey4l has quit IRC | 04:47 | |
| *** Jeffrey4l has joined #openstack-swift | 04:47 | |
| *** dmorita has quit IRC | 04:51 | |
| *** dimasot has joined #openstack-swift | 05:13 | |
| *** ppai has joined #openstack-swift | 05:35 | |
| *** Jeffrey4l has quit IRC | 05:47 | |
| *** dmorita has joined #openstack-swift | 05:48 | |
| *** trifon has joined #openstack-swift | 05:49 | |
| *** dimasot has quit IRC | 05:52 | |
| openstackgerrit | Hisashi Osanai proposed openstack/swift: Fix time unit of Recon's replication_time for object https://review.openstack.org/270014 | 05:52 |
| *** dmorita has quit IRC | 05:53 | |
| *** joker_ has joined #openstack-swift | 05:59 | |
| *** Jeffrey4l has joined #openstack-swift | 06:07 | |
| mattoliverau | notmyname: I am! | 06:15 |
| mattoliverau | notmyname: sorry, got sidetracked.. was debugging some sharding stuff | 06:15 |
| mattoliverau | mahatic: o/ | 06:15 |
| notmyname | certainly don't apologize for that! | 06:17 |
| *** Jeffrey4l has quit IRC | 06:18 | |
| mattoliverau | getting it to clean up parent containers (ie delete them) after sharding, cause we only care about the leaves, and we have reclaim age + if objects get replicated in the container wont fully delete.... however, I was doing some tweaking and forgot the case of not deleted the root container when it shards. Lets just say, sharding doesn't work when you kill the root (master) container.. opps :P | 06:20 |
| notmyname | heh | 06:20 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Fix EC ring validation at ring reload https://review.openstack.org/268993 | 06:21 |
| *** Jeffrey4l has joined #openstack-swift | 06:21 | |
| *** Jeffrey4l has quit IRC | 06:26 | |
| *** ChubYann has quit IRC | 06:36 | |
| *** jamielennox is now known as jamielennox|away | 06:40 | |
| *** proteusguy has quit IRC | 06:42 | |
| *** takashi_ has joined #openstack-swift | 06:42 | |
| *** proteusguy has joined #openstack-swift | 06:42 | |
| *** balajir has quit IRC | 06:43 | |
| *** balajir has joined #openstack-swift | 06:44 | |
| *** Jeffrey4l has joined #openstack-swift | 06:47 | |
| *** silor has joined #openstack-swift | 06:51 | |
| *** silor1 has joined #openstack-swift | 06:57 | |
| *** haomai___ has joined #openstack-swift | 06:58 | |
| *** haomaiwa_ has quit IRC | 06:58 | |
| *** silor has quit IRC | 06:59 | |
| *** silor1 is now known as silor | 06:59 | |
| *** haomai___ has quit IRC | 07:01 | |
| *** 6A4ABQMYH has joined #openstack-swift | 07:01 | |
| *** rcernin has joined #openstack-swift | 07:02 | |
| *** Guest40809 has quit IRC | 07:07 | |
| *** km__ has joined #openstack-swift | 07:08 | |
| *** km__ is now known as Guest42578 | 07:09 | |
| *** asettle_ has joined #openstack-swift | 07:25 | |
| *** asettle_ has quit IRC | 07:27 | |
| *** asettle has quit IRC | 07:29 | |
| kota_ | acoles: i found an issue for fast-post^^ (report is https://bugs.launchpad.net/swift/+bug/1536037) | 07:54 |
| openstack | Launchpad bug 1536037 in OpenStack Object Storage (swift) "fast-post broken with object mem server" [Undecided,New] | 07:54 |
| kota_ | acoles: i think it's just a bug not so big/urgent and also not affected to fast-post progress. | 07:55 |
| *** arnox has joined #openstack-swift | 07:58 | |
| *** 6A4ABQMYH has quit IRC | 08:01 | |
| *** haomaiwa_ has joined #openstack-swift | 08:01 | |
| openstackgerrit | Jude Augustine Job proposed openstack/python-swiftclient: Adding session to switfclient. https://review.openstack.org/270045 | 08:03 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Fix EC ring validation at ring reload https://review.openstack.org/268993 | 08:03 |
| *** rledisez has joined #openstack-swift | 08:05 | |
| kota_ | acoles: i tried to take care of fast-post route on proxy-server with patch 268993, that unfortunately failed in gerrit gate :/ | 08:05 |
| patchbot | kota_: https://review.openstack.org/#/c/268993/ - Fix EC ring validation at ring reload | 08:05 |
| *** cbartz has joined #openstack-swift | 08:06 | |
| *** Jeffrey4l has quit IRC | 08:08 | |
| *** Jeffrey4l has joined #openstack-swift | 08:14 | |
| *** eranrom has joined #openstack-swift | 08:19 | |
| *** remix_tj has quit IRC | 08:21 | |
| *** remix_tj has joined #openstack-swift | 08:22 | |
| *** trifon has quit IRC | 08:30 | |
| *** trifon has joined #openstack-swift | 08:31 | |
| *** geaaru has joined #openstack-swift | 08:35 | |
| *** remix_tj has quit IRC | 08:40 | |
| *** remix_tj has joined #openstack-swift | 08:40 | |
| *** dmorita has joined #openstack-swift | 08:47 | |
| *** dmorita has quit IRC | 08:51 | |
| eranrom | Hi all, Does https://review.openstack.org/#/c/269911/3 means the next release is finalyzed? | 08:52 |
| *** ho_ has quit IRC | 08:54 | |
| *** kei_yama has quit IRC | 08:55 | |
| *** takashi_ has quit IRC | 09:00 | |
| *** haomaiwa_ has quit IRC | 09:01 | |
| *** haomaiwang has joined #openstack-swift | 09:01 | |
| *** venkat has quit IRC | 09:05 | |
| *** dimasot has joined #openstack-swift | 09:11 | |
| openstackgerrit | Béla Vancsics proposed openstack/swift: Reduce code duplication https://review.openstack.org/254266 | 09:13 |
| *** joeljwright has joined #openstack-swift | 09:16 | |
| *** ChanServ sets mode: +v joeljwright | 09:16 | |
| *** jistr has joined #openstack-swift | 09:16 | |
| *** venkat has joined #openstack-swift | 09:17 | |
| *** Jeffrey4l has quit IRC | 09:34 | |
| venkat | Hi all | 09:41 |
| venkat | how to use header Transfer-Encoding for object PUT api?? | 09:42 |
| venkat | Is this header is supported only for EC? | 09:42 |
| *** dimasot has quit IRC | 09:45 | |
| venkat | hi someone can answer | 09:52 |
| *** aix has joined #openstack-swift | 09:56 | |
| *** haomaiwang has quit IRC | 10:01 | |
| *** haomaiwang has joined #openstack-swift | 10:01 | |
| ahale | its supported for non-EC, it just means you dont send a content-length and instead use a chunked transfer in the client | 10:06 |
| *** mariusv has quit IRC | 10:13 | |
| *** hyakuhei has joined #openstack-swift | 10:15 | |
| venkat | ahale: | 10:15 |
| venkat | I am using like this | 10:15 |
| venkat | curl -i $publicURL/container/local.conf -X PUT -H "X-Auth-Token: $token" -T local.conf -H "Transfer-Encoding : chunked" | 10:15 |
| *** hyakuhei has quit IRC | 10:16 | |
| venkat | but its not working as expected | 10:16 |
| *** tortib has joined #openstack-swift | 10:17 | |
| tortib | So is this channel for the Apple language Swift? | 10:18 |
| ahale | tortib: nope | 10:22 |
| tortib | ok | 10:22 |
| tortib | have a good day ahale | 10:22 |
| *** tortib has left #openstack-swift | 10:22 | |
| ahale | you too :) | 10:22 |
| ahale | venkat: it should accept (my swift does), curl is a pain though sometimes, doing unexpected things | 10:23 |
| *** links has quit IRC | 10:28 | |
| venkat | ahale: can we pass using python-swift-client | 10:29 |
| venkat | can u give an example | 10:29 |
| *** mingdang1 has quit IRC | 10:36 | |
| kota_ | eranrom: it's in working progress but i expect it will be done really soon, notmyname told in IRC, he is willing to finish up the release wednesday (maybe in PDT) | 10:37 |
| kota_ | eranrom: for next swift release | 10:37 |
| kota_ | eranrom: you can see the conversation in the swift IRC logs | 10:38 |
| ahale | venkat: this works for me- sends a "> Transfer-Encoding: chunked" header: "curl -v -X POST -H 'Transfer-Encoding: chunked' -H 'X-Auth-Token: TOKEN' -d @filename.txt https://swift/acc/cont/obj " | 10:46 |
| *** links has joined #openstack-swift | 10:51 | |
| eranrom | kota_: Thanks! I was searching the logs. do you remeber the day by any chance? | 10:52 |
| kota_ | eranrom: 1 sec | 10:52 |
| -kota_- 16:01 (notmyname) [01:34:17] oh, for the other-side-of-the-world shift that's coming online now, likely to be a swift release in 2 days (my wednesday) | 10:53 | |
| -kota_- 16:01 (kota_) [01:49:55] notmyname: do we have remaining patches (likely blockers) for the next swift release? | 10:53 | |
| kota_ | maybe, yesterday or day before? | 10:53 |
| eranrom | kota_: thanks! | 10:53 |
| eranrom | kota_,: Found it | 10:53 |
| kota_ | eranrom:^^ | 10:53 |
| eranrom | yep. Thanks | 10:53 |
| *** gabor_antal has joined #openstack-swift | 10:57 | |
| *** haomaiwang has quit IRC | 11:01 | |
| *** haomaiwang has joined #openstack-swift | 11:01 | |
| *** rcernin has quit IRC | 11:05 | |
| *** links has quit IRC | 11:05 | |
| *** rcernin has joined #openstack-swift | 11:07 | |
| *** rledisez has quit IRC | 11:09 | |
| *** rcernin is now known as rcernin|lunch | 11:13 | |
| joker_ | hi everyone, I sent code for review related with this bug(https://bugs.launchpad.net/swift/+bug/1509429): | 11:18 |
| openstack | Launchpad bug 1509429 in OpenStack Object Storage (swift) "X-Timestamp missing from object PUT/COPY response headers" [Low,In progress] - Assigned to Mustafa Çelik (celikmustafa89) | 11:18 |
| joker_ | and this one is my patch https://review.openstack.org/#/c/268163/ | 11:19 |
| openstackgerrit | Béla Vancsics proposed openstack/swift: Reduce code duplication https://review.openstack.org/267767 | 11:20 |
| joker_ | and this one is related with the previous patch about swift api https://review.openstack.org/#/c/269680/ | 11:20 |
| joker_ | can somebody help me with code-review? | 11:21 |
| *** hyakuhei has joined #openstack-swift | 11:21 | |
| *** daemontool has joined #openstack-swift | 11:21 | |
| joker_ | I can add you as reviewer to patch: | 11:21 |
| joker_ | thanks | 11:21 |
| *** links has joined #openstack-swift | 11:32 | |
| *** venkat has quit IRC | 11:33 | |
| *** sanchitmalhotra has quit IRC | 11:35 | |
| *** links has quit IRC | 11:38 | |
| *** hyakuhei has quit IRC | 11:42 | |
| *** openstackgerrit has quit IRC | 11:43 | |
| -openstackstatus- NOTICE: review.openstack.org is being restarted to apply patches | 11:44 | |
| *** ChanServ changes topic to "review.openstack.org is being restarted to apply patches" | 11:44 | |
| *** openstackgerrit has joined #openstack-swift | 11:44 | |
| *** venkat has joined #openstack-swift | 11:45 | |
| *** m_kazuhiro has quit IRC | 11:52 | |
| *** ChanServ changes topic to "Review Dashboard: https://goo.gl/mtEv1C | Hackathon: https://www.eventbrite.com/e/swift-hackathon-bristol-sponsored-by-hpe-tickets-19994495073 | Logs: http://eavesdrop.openstack.org/irclogs/%23openstack-swift/ | Meetings: https://wiki.openstack.org/wiki/Meetings/Swift | What's going on survey: https://gist.github.com/notmyname/be49b04165928fd6662f" | 11:55 | |
| -openstackstatus- NOTICE: Restart done, review.openstack.org is available | 11:55 | |
| *** mingdang1 has joined #openstack-swift | 11:57 | |
| *** aix has quit IRC | 11:59 | |
| *** haomaiwang has quit IRC | 12:01 | |
| *** haomaiwang has joined #openstack-swift | 12:01 | |
| *** ppai has quit IRC | 12:06 | |
| *** mrmoje has joined #openstack-swift | 12:07 | |
| *** ho_away has joined #openstack-swift | 12:07 | |
| *** Guest42578 has quit IRC | 12:09 | |
| *** ppai has joined #openstack-swift | 12:20 | |
| *** Jeffrey4l has joined #openstack-swift | 12:21 | |
| *** fthiagogv has joined #openstack-swift | 12:21 | |
| openstackgerrit | Merged openstack/python-swiftclient: Get rid of FakeConn cruft https://review.openstack.org/269399 | 12:27 |
| *** aix has joined #openstack-swift | 12:33 | |
| ho_away | takashi: around? | 12:34 |
| *** proteusguy has quit IRC | 12:37 | |
| *** rledisez has joined #openstack-swift | 12:43 | |
| *** rledisez has quit IRC | 12:43 | |
| *** takashi_ has joined #openstack-swift | 12:45 | |
| *** ppai has quit IRC | 12:47 | |
| *** proteusguy has joined #openstack-swift | 12:50 | |
| *** ppai has joined #openstack-swift | 12:59 | |
| *** Jeffrey4l has quit IRC | 12:59 | |
| *** davidhadas has joined #openstack-swift | 12:59 | |
| *** haomaiwang has quit IRC | 13:01 | |
| *** haomaiwa_ has joined #openstack-swift | 13:01 | |
| *** rcernin|lunch is now known as rcernin | 13:07 | |
| *** haomaiwa_ has quit IRC | 13:07 | |
| *** jmccarthy has joined #openstack-swift | 13:08 | |
| *** rledisez has joined #openstack-swift | 13:13 | |
| *** venkat has quit IRC | 13:14 | |
| *** Jeffrey4l has joined #openstack-swift | 13:17 | |
| *** daemontool has quit IRC | 13:18 | |
| tdasilva | onovy: hi, I think I mentioned to you before, but devconf is happening in a couple of weeks: http://devconf.cz/ | 13:22 |
| *** siva_krishnan has joined #openstack-swift | 13:23 | |
| *** marcusrafael has joined #openstack-swift | 13:30 | |
| onovy | tdasilva: yep, i have it on my backlog | 13:33 |
| onovy | will be there | 13:33 |
| marcusrafael | Good morning guys | 13:33 |
| tdasilva | onovy: cool, maybe we can meet there | 13:34 |
| onovy | of course | 13:34 |
| marcusrafael | Guys, I'm trying to install a devstack with swift and I'm getting this error: Missing parameter(s): Set an authentication URL, with --os-auth-url, OS_AUTH_URL or auth.auth_url | 13:34 |
| marcusrafael | Do you have any idea what is happening? | 13:35 |
| *** haomaiwa_ has joined #openstack-swift | 13:35 | |
| *** bill_az has joined #openstack-swift | 13:37 | |
| *** ho_away has quit IRC | 13:38 | |
| *** luksaur has joined #openstack-swift | 13:46 | |
| *** iGeek2001 has joined #openstack-swift | 13:46 | |
| iGeek2001 | Hi All | 13:46 |
| *** takashi__ has joined #openstack-swift | 13:47 | |
| iGeek2001 | Would someone be able to assist with swift proxy server query | 13:47 |
| iGeek2001 | All my swift SAIO services are starting | 13:47 |
| iGeek2001 | however the proxy server fails | 13:47 |
| iGeek2001 | with the following error | 13:48 |
| iGeek2001 | ImportError: No module named catch_errors | 13:48 |
| *** mahatic_ has joined #openstack-swift | 13:49 | |
| *** cebruns_ has joined #openstack-swift | 13:50 | |
| *** mingdang1 has quit IRC | 13:50 | |
| *** hachi-ju has joined #openstack-swift | 13:50 | |
| *** hachi-ju is now known as hguemar | 13:50 | |
| *** hguemar has quit IRC | 13:50 | |
| *** baffle__ has joined #openstack-swift | 13:51 | |
| *** dfg_ has joined #openstack-swift | 13:51 | |
| *** dosaboy_ has joined #openstack-swift | 13:51 | |
| *** lifeless_ has joined #openstack-swift | 13:51 | |
| *** saltsa_ has joined #openstack-swift | 13:51 | |
| *** remix_auei has joined #openstack-swift | 13:51 | |
| *** tonyb_ has joined #openstack-swift | 13:51 | |
| *** j_king_ has joined #openstack-swift | 13:51 | |
| *** hachi-ju has joined #openstack-swift | 13:51 | |
| *** clyps___1 has joined #openstack-swift | 13:52 | |
| *** timur_ has joined #openstack-swift | 13:52 | |
| *** number80 has quit IRC | 13:52 | |
| *** hachi-ju is now known as number80 | 13:52 | |
| *** number80 has quit IRC | 13:53 | |
| *** number80 has joined #openstack-swift | 13:53 | |
| *** jlhinson_ has joined #openstack-swift | 13:53 | |
| *** jrichli_ has joined #openstack-swift | 13:54 | |
| *** coreycb` has joined #openstack-swift | 13:54 | |
| *** wbhuber_ has joined #openstack-swift | 13:54 | |
| *** raginbaj- has joined #openstack-swift | 13:55 | |
| *** ajiang_ has joined #openstack-swift | 13:55 | |
| *** hrou_ has joined #openstack-swift | 13:55 | |
| *** trifon has quit IRC | 13:56 | |
| *** takashi has quit IRC | 13:56 | |
| *** d0ugal has quit IRC | 13:56 | |
| *** saltsa has quit IRC | 13:56 | |
| *** wbhuber has quit IRC | 13:56 | |
| *** ajiang has quit IRC | 13:56 | |
| *** dosaboy has quit IRC | 13:56 | |
| *** mahatic has quit IRC | 13:56 | |
| *** raginbajin has quit IRC | 13:56 | |
| *** ahale has quit IRC | 13:56 | |
| *** jlhinson has quit IRC | 13:56 | |
| *** a1|away has quit IRC | 13:56 | |
| *** anteaya has quit IRC | 13:56 | |
| *** timur has quit IRC | 13:56 | |
| *** dfg has quit IRC | 13:56 | |
| *** coreycb has quit IRC | 13:56 | |
| *** remix_tj has quit IRC | 13:56 | |
| *** eranrom has quit IRC | 13:56 | |
| *** cebruns has quit IRC | 13:56 | |
| *** zackmdavis has quit IRC | 13:56 | |
| *** tonyb has quit IRC | 13:56 | |
| *** lifeless has quit IRC | 13:56 | |
| *** hrou has quit IRC | 13:56 | |
| *** clyps___ has quit IRC | 13:56 | |
| *** j_king has quit IRC | 13:56 | |
| *** baffle has quit IRC | 13:56 | |
| *** jrichli has quit IRC | 13:56 | |
| *** raginbaj- is now known as raginbajin | 13:56 | |
| *** ahale has joined #openstack-swift | 13:57 | |
| *** zackmdavis has joined #openstack-swift | 13:57 | |
| *** remix_auei is now known as remix_tj | 13:58 | |
| *** haomaiwa_ has quit IRC | 14:01 | |
| *** haomaiwang has joined #openstack-swift | 14:01 | |
| *** Jeffrey4l has quit IRC | 14:01 | |
| *** PsionTheory has joined #openstack-swift | 14:02 | |
| *** trifon has joined #openstack-swift | 14:02 | |
| *** d0ugal has joined #openstack-swift | 14:03 | |
| *** a1|away has joined #openstack-swift | 14:03 | |
| *** anteaya has joined #openstack-swift | 14:03 | |
| *** tonyb_ is now known as tonyb | 14:03 | |
| *** ho_away has joined #openstack-swift | 14:10 | |
| *** eranrom has joined #openstack-swift | 14:12 | |
| *** dslevin has joined #openstack-swift | 14:15 | |
| *** dslevin has quit IRC | 14:16 | |
| *** dslevin has joined #openstack-swift | 14:16 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Decrypt multipart mime doc https://review.openstack.org/248377 | 14:21 |
| *** acoles_ is now known as acoles | 14:22 | |
| openstackgerrit | Victor Stinner proposed openstack/swift: py3: Use the six module in the xprofile middleware https://review.openstack.org/233145 | 14:24 |
| *** takashi_ has quit IRC | 14:24 | |
| *** ho_away is now known as ho | 14:26 | |
| *** ho is now known as Guest88479 | 14:27 | |
| *** Guest88479 is now known as ho_away | 14:29 | |
| *** haypo has joined #openstack-swift | 14:32 | |
| notmyname | good morning | 14:32 |
| haypo | hi. i'm running "tox -e py34" on Swift, and it fails with: ERROR: unable to find suitable PyECLib type (none of ['liberasurecode_rs_vand', 'jerasure_rs_vand'] found in []) | 14:32 |
| haypo | i have PyEClib 1.2.0 | 14:32 |
| notmyname | min version is 1.0.7 | 14:33 |
| notmyname | and later versions have unbundled liberasurecode | 14:33 |
| haypo | swift unit tests still require an external library? i doesn't include its own erasure code? | 14:33 |
| notmyname | so you need to install liberasurecode separately now (instead of it being bundled with pyeclib) | 14:33 |
| haypo | notmyname: i confirm that liberasurecode is not installed on my system :) | 14:33 |
| notmyname | :-) | 14:33 |
| haypo | sorry, i forgot all detail about PyEClib :) i will install liberasurecode | 14:34 |
| *** hseipp has joined #openstack-swift | 14:34 | |
| eranrom | notmyname,: morning. Just want to make sure you saw my comment on 2.6.0 (patch 269911) | 14:35 |
| patchbot | eranrom: https://review.openstack.org/#/c/269911/ - authors and changelog updates for 2.6.0 | 14:35 |
| haypo | notmyname: if i recall correctly, it was no mmore embedded, but then embedded again | 14:37 |
| haypo | notmyname: it looks like it changed a third time :) | 14:37 |
| notmyname | eranrom: yes, and I've got it open now | 14:38 |
| notmyname | haypo: yeah. the plan is to keep it separate, and later versions now have their own simple EC alorithm (so you don't need jerasure or isa-l or anything) | 14:39 |
| eranrom | notmyname,: thanks! and apologies for the parallel efforts here :-) | 14:39 |
| *** petertr7_away is now known as petertr7 | 14:40 | |
| openstackgerrit | Peter Lisák proposed openstack/swift: Unification of manpages and swift-* --help https://review.openstack.org/251397 | 14:44 |
| *** sgundur has joined #openstack-swift | 14:46 | |
| *** ppai has quit IRC | 14:48 | |
| openstackgerrit | Victor Stinner proposed openstack/swift: Port get_hmac() and hash_path() to Python 3 https://review.openstack.org/236998 | 14:50 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Port parse_mime_headers() to Python 3 https://review.openstack.org/237027 | 14:53 |
| *** lpabon has joined #openstack-swift | 14:53 | |
| haypo | ok, you have py3 patches to review ;-) https://review.openstack.org/237027 https://review.openstack.org/236998 https://review.openstack.org/237019 | 14:55 |
| haypo | enjoy! | 14:56 |
| notmyname | :-) | 14:56 |
| * notmyname is about to push some patches | 14:59 | |
| openstackgerrit | John Dickinson proposed openstack/swift: Fix memory/socket leak in proxy on truncated SLO/DLO GET https://review.openstack.org/270233 | 14:59 |
| openstackgerrit | John Dickinson proposed openstack/swift: authors and changelog updates for 2.6.0 https://review.openstack.org/269911 | 15:00 |
| openstackgerrit | John Dickinson proposed openstack/swift: Fix memory/socket leak in proxy on truncated SLO/DLO GET https://review.openstack.org/270234 | 15:00 |
| *** haomaiwang has quit IRC | 15:01 | |
| *** haomaiwa_ has joined #openstack-swift | 15:01 | |
| *** daemontool has joined #openstack-swift | 15:02 | |
| notmyname | now only 22+ hours for those to get through the gate (assuming they go through the first time and there aren't spurious test failures) | 15:04 |
| *** blmartin has joined #openstack-swift | 15:05 | |
| *** petertr7 is now known as petertr7_away | 15:06 | |
| *** iGeek2001 has quit IRC | 15:06 | |
| *** daemontool has quit IRC | 15:07 | |
| *** daemontool has joined #openstack-swift | 15:07 | |
| *** zaitcev has joined #openstack-swift | 15:12 | |
| *** ChanServ sets mode: +v zaitcev | 15:12 | |
| *** lcurtis has joined #openstack-swift | 15:12 | |
| *** daemontool has quit IRC | 15:13 | |
| notmyname | for those of you rolling your own swift packages, I'd strongly recommend deploying this patch | 15:14 |
| notmyname | timburke: joeljwright: https://review.openstack.org/#/c/270045/1 looks interesting | 15:15 |
| notmyname | eranrom: ^ | 15:15 |
| notmyname | ie something we talked about in tokyo | 15:16 |
| joeljwright | interesting… need to understand what that's doing, but it's not a big change | 15:17 |
| notmyname | if anyone does .NET dev work, I just saw this http://blog.veritech.io/2016/01/openstack-swift-client-for-net-core.html | 15:18 |
| *** hrou_ has quit IRC | 15:19 | |
| *** jrichli_ has quit IRC | 15:19 | |
| *** jlhinson_ has quit IRC | 15:19 | |
| *** hrou has joined #openstack-swift | 15:20 | |
| eranrom | notmyname: looking... | 15:20 |
| notmyname | joeljwright: yeah, that's exactly what I thought | 15:21 |
| *** jrichli has joined #openstack-swift | 15:21 | |
| onovy | hmm, CVE... | 15:21 |
| *** jlhinson has joined #openstack-swift | 15:21 | |
| joeljwright | notmyname: first thing to work out is the spectacular test failures | 15:22 |
| notmyname | yeah, I think that's the pip 8 issue that was brought up on the ML yesterday. seemed to break the world | 15:22 |
| notmyname | or that's most likely. ie everything is breaking | 15:23 |
| notmyname | onovy: fun, right? | 15:23 |
| onovy | notmyname: yep :) | 15:23 |
| joeljwright | notmyname: tell me about it! pip 8.0.0 has been breaking our testing all day :( | 15:23 |
| *** joker_ has quit IRC | 15:26 | |
| onovy | notmyname: if i understand it correclty <= 2.2.0 is not affected? | 15:27 |
| notmyname | onovy: that should be true. and that's what is said by https://review.openstack.org/#/c/270241/1/ossa/OSSA-2016-004.yaml | 15:28 |
| onovy | perfect. no work for debian maintainer. jessie-stable = 2.2.0 :) | 15:29 |
| eranrom | notmyname,: I thought that we wanted to leverage sessions to somehow free the user from being aware of the protocol version. That is session plus a swift plugin to allow v1. anyway this change looks like semantics but I amprobably missing something. | 15:29 |
| notmyname | onovy: I'm not sure I agree with that ;-) | 15:29 |
| onovy | notmyname: arg! | 15:30 |
| notmyname | eranrom: yeah, I think you're right | 15:30 |
| * onovy don't like backporting fixies to old code | 15:30 | |
| onovy | but exactly i will need backport for 2.3.0 liberty | 15:30 |
| notmyname | onovy: you should be using 2.6.0 as soon as it's released (hopefully today, but tagged later this week because of gate slowness) | 15:30 |
| notmyname | onovy: ;-) | 15:30 |
| onovy | notmyname: i will! | 15:31 |
| onovy | it will be in debian in few minutes after your release | 15:31 |
| notmyname | yay | 15:31 |
| onovy | and few minutes later in our production | 15:31 |
| onovy | why wait? :) | 15:31 |
| *** mragupat has joined #openstack-swift | 15:32 | |
| *** cdelatte has joined #openstack-swift | 15:32 | |
| notmyname | the eventlet bump needs to land. the authors/changelog of course. and, if at all possible (and it's had a lot of reviews already) eranrom's container sync patch, although it's not small | 15:32 |
| *** diogogmt has joined #openstack-swift | 15:33 | |
| eranrom | notmyname,: Thanks! | 15:36 |
| notmyname | eranrom: so that's what I'm thinking. I do definitely want to get your container sync patch it (and so far, it's super easy to read!) | 15:37 |
| notmyname | the CVE should result in a release, too | 15:37 |
| *** delattec has joined #openstack-swift | 15:37 | |
| eranrom | notmyname,: CVE? | 15:38 |
| notmyname | the patches I pushed up 38 minutes ago | 15:38 |
| notmyname | patch 270233 | 15:39 |
| patchbot | notmyname: https://review.openstack.org/#/c/270233/ - Fix memory/socket leak in proxy on truncated SLO/D... | 15:39 |
| eranrom | notmyname,: looking... | 15:39 |
| notmyname | eranrom: you really really want to deploy that patch | 15:39 |
| *** cdelatte has quit IRC | 15:39 | |
| *** diogogmt has quit IRC | 15:40 | |
| notmyname | eranrom: the tl;dr is "client starts to download a *LO and stops partway through. server leaks a socket fd and a small amount of memory" | 15:40 |
| notmyname | the memory is actually less concerning (still important) than the socket leak. you'll run out of file descriptors before you run out of ram | 15:41 |
| eranrom | notmyname,: yep got that. I can help testing / reviewing this. | 15:41 |
| notmyname | eranrom: it's already been done when it was under embargo. and it's already been +A. just needs to go through the gate | 15:41 |
| *** jordanP has joined #openstack-swift | 15:41 | |
| notmyname | you can see the discussion on the linked bug | 15:42 |
| eranrom | notmyname,: I see. | 15:43 |
| *** cbartz has left #openstack-swift | 15:43 | |
| openstackgerrit | John Dickinson proposed openstack/swift: authors and changelog updates for 2.6.0 https://review.openstack.org/269911 | 15:45 |
| acoles | notmyname: it was a nit !) | 15:49 |
| notmyname | acoles: ;-) | 15:49 |
| notmyname | it's easy to iterate | 15:49 |
| acoles | they're easy to iterate | 15:51 |
| notmyname | lol | 15:51 |
| *** siva_krishnan1 has joined #openstack-swift | 15:57 | |
| *** siva_krishnan1 has quit IRC | 15:57 | |
| *** hyakuhei has joined #openstack-swift | 15:59 | |
| *** haomaiwa_ has quit IRC | 16:01 | |
| *** haomaiwa_ has joined #openstack-swift | 16:01 | |
| *** nadeem has joined #openstack-swift | 16:02 | |
| *** nadeem has quit IRC | 16:03 | |
| notmyname | eranrom: I've looked through the patch for the core changes but not the tests yet. seems pretty straightforward and a pretty clever solution. | 16:04 |
| *** nadeem has joined #openstack-swift | 16:04 | |
| eranrom | notmyname,: Thanks! | 16:07 |
| notmyname | so supposed we land this and it turns out to be a terrible idea. since all you're doing here is doing some symlinks, there's not even any major concern with doing something different later | 16:07 |
| eranrom | notmyname,: I really^3 appreciate this | 16:07 |
| eranrom | Agree. I can also add the following: Idea was reviewed by clayg and acoles. and when container sharding comes in, we will automatically gain spreading the sync work over all container servers free. | 16:09 |
| eranrom | s/free/for free | 16:09 |
| notmyname | assuming there's a db broker for each of the shards | 16:10 |
| eranrom | notmyname,: Well, this was my understanding at the time. Otherwise, we should adopt container sync accordingly. | 16:11 |
| acoles | notmyname: eranrom fwiw i think the biggest risk (in landing the sync patch) is that we missed a (failure) scenario which ends up with a db that should be sync'd but no symlink to it. | 16:11 |
| acoles | otherwise its a fairly simple (but neat) idea | 16:13 |
| eranrom | acoles: you mean a failure scenario where e.g. the replicator will fail to fix that? Well, I guess that this is why IBM first beta this :-) | 16:13 |
| eranrom | acoles,: BTW we are also currently building an automated system tests for the feature over two non trivial clusters | 16:14 |
| eranrom | acoles,: I will be happy to describe in the hakathon | 16:14 |
| notmyname | acoles: yeah. the good news is that the simplicity of the design is that it's pretty easy to reason about those. it's updated on PUT, POST, and DELETE. there's not too many other places to worry about | 16:14 |
| notmyname | eranrom: you're coming to bristol, right? | 16:15 |
| eranrom | notmyname: yep. Got a queue of topics as well :-) | 16:15 |
| notmyname | heh. me too ;-) | 16:15 |
| eranrom | :-) | 16:15 |
| acoles | eranrom: i was just reflecting on what another reviewer might look out for, given that i am fallible :) | 16:17 |
| zaitcev | Oh god, another big and important review | 16:18 |
| *** joeljwright has quit IRC | 16:18 | |
| notmyname | zaitcev: there's always another one, isn't there? ;-) | 16:18 |
| eranrom | acoles,: well we are humans at the end of the day, but I would guess "fallibelity" is lower thn average | 16:20 |
| *** diazjf has joined #openstack-swift | 16:22 | |
| *** ho_away has quit IRC | 16:22 | |
| notmyname | I just got an email saying hackathon hotel block group rate is extended. that's good (for those who haven't booked yet). thanks acoles | 16:23 |
| eranrom | notmyname,: ^^ which remonds me I should start getting prepared... | 16:24 |
| acoles | notmyname: well that's a coincidence cos I just sent a similar email :) | 16:24 |
| acoles | eranrom: YES | 16:24 |
| *** Jeffrey4l has joined #openstack-swift | 16:24 | |
| *** rcernin has quit IRC | 16:25 | |
| *** Jeffrey4l has quit IRC | 16:25 | |
| *** Jeffrey4l has joined #openstack-swift | 16:25 | |
| *** jordanP has quit IRC | 16:30 | |
| *** davidhadas has quit IRC | 16:35 | |
| *** awelleck has joined #openstack-swift | 16:36 | |
| *** davidhadas_ has joined #openstack-swift | 16:40 | |
| ahale | I heard about the existence of a UK Openstack Swift Hackathon but didnt see it mentioned anywhere, is it private ? | 16:47 |
| notmyname | ahale: nope. it's in the channel topic | 16:47 |
| ahale | who reads topics ! | 16:47 |
| notmyname | clearly not you! | 16:47 |
| notmyname | ;-) | 16:47 |
| ahale | clearly | 16:48 |
| notmyname | ahale: I think it would be awesome if you were to make it to bristol | 16:48 |
| mahatic_ | ha, and i just discovered this "what's going on survey". there should be a notification for channel topic changes ;) | 16:48 |
| *** trifon has quit IRC | 16:49 | |
| mahatic_ | oh it doesn't have a gist of answers | 16:50 |
| notmyname | I'm still working on putting that together. I'll highlight a couple of things in today's meeting | 16:52 |
| *** gabor_antal has quit IRC | 16:53 | |
| notmyname | ahale: you might be interested in the results onovy/peterlisak got from the ionice changes: https://docs.google.com/spreadsheets/d/1dzcMutNwRYcDiSMNOhKLM1tk8QX3uqFg9BJszJ6plr4/edit?usp=sharing | 16:53 |
| *** haypo has left #openstack-swift | 16:56 | |
| *** delattec has quit IRC | 17:00 | |
| *** haomaiwa_ has quit IRC | 17:01 | |
| openstackgerrit | Mustafa ÇELİK proposed openstack/swift: X-Timestamp missing from response headers https://review.openstack.org/268163 | 17:02 |
| *** delattec has joined #openstack-swift | 17:03 | |
| *** jistr has quit IRC | 17:04 | |
| *** klrmn has joined #openstack-swift | 17:04 | |
| *** haomaiwa_ has joined #openstack-swift | 17:05 | |
| *** chsc has joined #openstack-swift | 17:05 | |
| *** eranrom has quit IRC | 17:05 | |
| *** gyee has joined #openstack-swift | 17:06 | |
| *** arnox has quit IRC | 17:08 | |
| notmyname | bah, eranrom left | 17:08 |
| openstackgerrit | Mustafa ÇELİK proposed openstack/swift: Extra X-Timestamp are removed from error responses https://review.openstack.org/268163 | 17:08 |
| *** klrmn has quit IRC | 17:11 | |
| *** ho_away has joined #openstack-swift | 17:23 | |
| *** dmorita has joined #openstack-swift | 17:26 | |
| *** dmorita has quit IRC | 17:27 | |
| *** ho_away has quit IRC | 17:27 | |
| *** dmorita has joined #openstack-swift | 17:31 | |
| *** daemontool has joined #openstack-swift | 17:34 | |
| openstackgerrit | John Dickinson proposed openstack/swift: authors and changelog updates for 2.6.0 https://review.openstack.org/269911 | 17:35 |
| notmyname | updated changelog to include the contaienr sync improvements ^ | 17:35 |
| notmyname | now it's time to commute to the office and do a bunch of meetings | 17:36 |
| *** mragupat has quit IRC | 17:37 | |
| *** Jeffrey4l has quit IRC | 17:40 | |
| *** daemontool has quit IRC | 17:44 | |
| *** jmccarthy has quit IRC | 17:46 | |
| *** rledisez has quit IRC | 17:51 | |
| *** nadeem has quit IRC | 17:54 | |
| *** mragupat has joined #openstack-swift | 17:58 | |
| *** mragupat has quit IRC | 17:59 | |
| *** haomaiwa_ has quit IRC | 18:01 | |
| *** haomaiwang has joined #openstack-swift | 18:01 | |
| *** klrmn has joined #openstack-swift | 18:07 | |
| *** lifeless_ is now known as lifeless | 18:12 | |
| *** natarej_ has joined #openstack-swift | 18:13 | |
| *** takashi has joined #openstack-swift | 18:13 | |
| *** diazjf has quit IRC | 18:14 | |
| *** timur has joined #openstack-swift | 18:16 | |
| *** Zyric_ has joined #openstack-swift | 18:16 | |
| *** kragniz_ has joined #openstack-swift | 18:17 | |
| *** peluse_ has joined #openstack-swift | 18:18 | |
| *** wasmum- has joined #openstack-swift | 18:19 | |
| *** peterlisak_ has joined #openstack-swift | 18:20 | |
| *** eranrom has joined #openstack-swift | 18:20 | |
| *** diazjf has joined #openstack-swift | 18:21 | |
| *** ahale has quit IRC | 18:22 | |
| *** timur_ has quit IRC | 18:22 | |
| *** takashi__ has quit IRC | 18:22 | |
| *** marcusrafael has quit IRC | 18:22 | |
| *** luksaur has quit IRC | 18:22 | |
| *** peluse has quit IRC | 18:22 | |
| *** natarej has quit IRC | 18:22 | |
| *** peterlisak has quit IRC | 18:22 | |
| *** wasmum has quit IRC | 18:22 | |
| *** sudorandom has quit IRC | 18:22 | |
| *** Zyric has quit IRC | 18:22 | |
| *** portante has quit IRC | 18:22 | |
| *** pchng has quit IRC | 18:22 | |
| *** kragniz has quit IRC | 18:22 | |
| *** Zyric_ is now known as Zyric | 18:22 | |
| *** peterlisak_ is now known as peterlisak | 18:22 | |
| *** gyee has quit IRC | 18:22 | |
| *** gyee has joined #openstack-swift | 18:23 | |
| *** davidhadas_ has quit IRC | 18:23 | |
| *** hyakuhei has quit IRC | 18:23 | |
| *** ahale has joined #openstack-swift | 18:23 | |
| *** portante has joined #openstack-swift | 18:23 | |
| *** lpabon has quit IRC | 18:24 | |
| *** ho_away has joined #openstack-swift | 18:24 | |
| *** sudorandom has joined #openstack-swift | 18:24 | |
| *** hseipp has quit IRC | 18:25 | |
| *** delattec has quit IRC | 18:26 | |
| *** delattec has joined #openstack-swift | 18:27 | |
| *** luksaur has joined #openstack-swift | 18:28 | |
| *** marcusrafael has joined #openstack-swift | 18:29 | |
| *** ho_away has quit IRC | 18:30 | |
| *** dmorita has quit IRC | 18:31 | |
| *** dmorita has joined #openstack-swift | 18:33 | |
| *** sudorandom has quit IRC | 18:35 | |
| *** sudorandom has joined #openstack-swift | 18:37 | |
| *** pchng has joined #openstack-swift | 18:39 | |
| clayg | yay eranrom container sync! | 18:40 |
| eranrom | clayg,: Thanks Clay! Its only the beginnig though :-) | 18:43 |
| notmyname | eranrom: note the inline comment on my review. there is one small follow-up that I think should happen | 18:46 |
| *** sudorandom has quit IRC | 18:48 | |
| *** sudorandom has joined #openstack-swift | 18:48 | |
| *** flaper87 has quit IRC | 18:54 | |
| *** flaper87 has joined #openstack-swift | 18:54 | |
| eranrom | notmyname,:looking | 18:56 |
| *** cebruns_ has quit IRC | 18:56 | |
| *** cebruns_ has joined #openstack-swift | 18:56 | |
| *** hrou has quit IRC | 18:57 | |
| *** hrou has joined #openstack-swift | 18:57 | |
| *** jrichli has quit IRC | 18:57 | |
| *** jrichli has joined #openstack-swift | 18:57 | |
| *** jlhinson has quit IRC | 18:57 | |
| *** jlhinson has joined #openstack-swift | 18:57 | |
| *** awelleck has quit IRC | 18:58 | |
| *** awelleck has joined #openstack-swift | 18:58 | |
| *** sudorandom has quit IRC | 18:59 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Fix bug in mem_diskfile write_metadata method https://review.openstack.org/270422 | 18:59 |
| *** wasmum- has quit IRC | 18:59 | |
| *** wasmum- has joined #openstack-swift | 18:59 | |
| *** portante has quit IRC | 18:59 | |
| *** portante has joined #openstack-swift | 18:59 | |
| acoles | kota_: well that was fun :/ ^^ | 18:59 |
| *** pchng has quit IRC | 18:59 | |
| *** pchng has joined #openstack-swift | 18:59 | |
| acoles | kota_: maybe try applying that patch to your test to see if it fixes things | 19:00 |
| acoles | eranrom: so, now container-sync is all done, would you like to review fast-post?? :P | 19:00 |
| *** haomaiwang has quit IRC | 19:01 | |
| *** 17WABLM3R has joined #openstack-swift | 19:01 | |
| *** dmorita has quit IRC | 19:02 | |
| *** mragupat has joined #openstack-swift | 19:02 | |
| *** mragupat has quit IRC | 19:02 | |
| *** mragupat has joined #openstack-swift | 19:02 | |
| *** mragupat has quit IRC | 19:03 | |
| eranrom | acoles,: Absolutely. in fact I thought there is also come container sync coding to do there. | 19:03 |
| eranrom | acoles: and it is totally not all done :-) | 19:03 |
| acoles | mattoliverau: I added dashboard links to my gerrit menu like you discovered, but they seem to open a new browser tab every time - is that what you see? | 19:03 |
| *** mragupat has joined #openstack-swift | 19:03 | |
| acoles | eranrom: true! | 19:04 |
| eranrom | acoles: do you want to point me to a specific patch to start with? | 19:04 |
| *** dmorita has joined #openstack-swift | 19:04 | |
| *** dmorita has quit IRC | 19:05 | |
| acoles | eranrom: i was sort of kidding, but patch 135380 | 19:05 |
| patchbot | acoles: https://review.openstack.org/#/c/135380/ - Update container on fast-POST | 19:05 |
| acoles | eranrom: could be worth reading the spec first https://specs.openstack.org/openstack/swift-specs/specs/in_progress/fastpostupdates.html | 19:06 |
| eranrom | acoles,: I was not :-) I would also like to do the containr sync conding stuff. but first things first... | 19:06 |
| acoles | eranrom: caveat - we may have deviated from the spec in practice | 19:06 |
| eranrom | acoles,: alright. | 19:07 |
| acoles | eranrom: thanks! | 19:07 |
| acoles | back later for meeting | 19:07 |
| *** dmorita_ has joined #openstack-swift | 19:07 | |
| *** acoles is now known as acoles_ | 19:08 | |
| *** acoles_ is now known as acoles | 19:09 | |
| *** acoles is now known as acoles_ | 19:09 | |
| *** geaaru has quit IRC | 19:09 | |
| *** ChubYann has joined #openstack-swift | 19:12 | |
| *** aix has quit IRC | 19:16 | |
| *** zhill has joined #openstack-swift | 19:19 | |
| *** dmorita_ has quit IRC | 19:22 | |
| *** dmorita has joined #openstack-swift | 19:22 | |
| *** lpabon has joined #openstack-swift | 19:23 | |
| *** lpabon has quit IRC | 19:25 | |
| *** ho_away has joined #openstack-swift | 19:26 | |
| *** dmorita has quit IRC | 19:27 | |
| *** diogogmt has joined #openstack-swift | 19:28 | |
| *** ho_away has quit IRC | 19:30 | |
| *** mrmoje has quit IRC | 19:33 | |
| *** mrmoje has joined #openstack-swift | 19:33 | |
| *** diazjf has quit IRC | 19:35 | |
| *** diazjf has joined #openstack-swift | 19:47 | |
| *** dmorita has joined #openstack-swift | 19:49 | |
| *** silor has quit IRC | 19:50 | |
| jrichli | yay! container sync improvements. | 19:50 |
| *** nadeem has joined #openstack-swift | 19:55 | |
| *** nadeem has quit IRC | 19:56 | |
| *** nadeem has joined #openstack-swift | 19:56 | |
| *** 17WABLM3R has quit IRC | 20:01 | |
| *** haomaiwang has joined #openstack-swift | 20:01 | |
| Zyric | Good morning | 20:03 |
| *** mragupat_ has joined #openstack-swift | 20:08 | |
| eranrom | jrichli,: Thanks! :-) | 20:08 |
| *** mragupat has quit IRC | 20:11 | |
| *** sudorandom has joined #openstack-swift | 20:18 | |
| *** changbl has quit IRC | 20:19 | |
| *** mrmoje has quit IRC | 20:21 | |
| *** dmorita has quit IRC | 20:22 | |
| *** dmorita has joined #openstack-swift | 20:23 | |
| *** dmorita has quit IRC | 20:27 | |
| *** sudorandom has quit IRC | 20:28 | |
| *** diazjf1 has joined #openstack-swift | 20:35 | |
| *** diazjf has quit IRC | 20:37 | |
| *** awelleck has quit IRC | 20:40 | |
| mattoliverau | morning all | 20:40 |
| mattoliverau | The wife has a doctors appointment this morning, so I'll do the right thing and get her to drive while I attend the meeting on my phone.. if I'm a bit slow or in and out thats why ;) | 20:41 |
| eranrom | mattoliverau,: morning. Are you sure this is the right thing? | 20:41 |
| mattoliverau | eranrom: her suggestion, so I trained her right ;) | 20:42 |
| eranrom | mattoliverau,: :-) In my case I am the one trained right I guess... | 20:42 |
| *** sudorandom has joined #openstack-swift | 20:42 | |
| eranrom | mattoliverau,: depends who you ask of-corse :-) | 20:43 |
| *** cbartz has joined #openstack-swift | 20:43 | |
| mattoliverau | eranrom: lol! | 20:44 |
| *** ho_away has joined #openstack-swift | 20:49 | |
| *** cbartz has quit IRC | 20:49 | |
| *** cbartz has joined #openstack-swift | 20:50 | |
| notmyname | mattoliverau: lol | 20:51 |
| *** sudorandom has quit IRC | 20:53 | |
| *** cbartz has quit IRC | 20:54 | |
| *** changbl has joined #openstack-swift | 20:54 | |
| *** awelleck has joined #openstack-swift | 20:55 | |
| *** cbartz has joined #openstack-swift | 20:56 | |
| *** acoles_ is now known as acoles | 20:57 | |
| mattoliverau | acoles: mine don't open in new windows. Once I'm in front my laptop I'll take a closer look | 20:58 |
| mattoliverau | ^^ dashboard links | 20:58 |
| kota_ | good morning | 20:59 |
| notmyname | hello | 20:59 |
| ho_away | good morning! | 20:59 |
| onovy | morning | 20:59 |
| notmyname | meeting time in #openstack-meeting | 20:59 |
| acoles | mattoliverau: thanks. maybe i did something wrong | 21:00 |
| *** haomaiwang has quit IRC | 21:01 | |
| *** joeljwright has joined #openstack-swift | 21:01 | |
| *** ChanServ sets mode: +v joeljwright | 21:01 | |
| *** m_kazuhiro has joined #openstack-swift | 21:13 | |
| *** finisherr has joined #openstack-swift | 21:14 | |
| finisherr | Just in case anyone is interested, the strange encoding issue I was having was coming from the boto client's s3 bridge | 21:14 |
| finisherr | requets with curl did not encode colons to %3a | 21:14 |
| finisherr | Now, I have another issue on my hands. I have a staging cluster that was built. Getting 404s on puts (uploads) | 21:15 |
| finisherr | anyone know off hand might cause that outside of maybe permissions ont he account or container directories? | 21:15 |
| cschwede | zaitcev: please have a look at patch 269811 if you want to have a look at the eventlet version bump before it get’s merged | 21:17 |
| patchbot | cschwede: https://review.openstack.org/#/c/269811/ - Bump eventlet min version to 0.17.4 | 21:17 |
| zaitcev | cschwede: F23 ships with eventlet-0.17.4. What is your concern? | 21:18 |
| *** m_kazuhi_ has joined #openstack-swift | 21:19 | |
| cschwede | zaitcev: no concerns from my side, i just wanted to ping you about it and check if you have any concerns ;) | 21:19 |
| *** m_kazuhiro has quit IRC | 21:19 | |
| *** nadeem has quit IRC | 21:19 | |
| *** diazjf1 has quit IRC | 21:19 | |
| zaitcev | cschwede: Thanks, I saw that patch and I don't have any issues with it. | 21:20 |
| cschwede | zaitcev: alright, thx! +A’ed, should be merged in 1-2 days… | 21:21 |
| *** diazjf has joined #openstack-swift | 21:22 | |
| *** delattec has quit IRC | 21:28 | |
| *** m_kazuhiro has joined #openstack-swift | 21:33 | |
| notmyname | finisherr: good to know (where the bug is) | 21:33 |
| finisherr | notmyname: right? | 21:33 |
| hrou | At the account level can you get a list of which containers belong to which storage policy (I see its in the account db container table) but is there an external API way of getting at it or its not returned ? | 21:34 |
| *** m_kazuhi_ has quit IRC | 21:34 | |
| *** m_kazuhi_ has joined #openstack-swift | 21:36 | |
| *** m_kazuhi_ has quit IRC | 21:37 | |
| clayg | hrou: oh bother, does it not come back in the json response? | 21:37 |
| *** m_kazuhiro has quit IRC | 21:37 | |
| hrou | clayg, I didn't think I was seeing it but maybe I was losing it ;) | 21:38 |
| clayg | hrou: no i don't see where the select in account.backend is calling for it - so I don't see how it could make it's way into the response | 21:40 |
| hrou | clayg, yep: {"count": 2, "bytes": 15, "name": "cont6"} | 21:41 |
| clayg | hrou: seems like an annoying-ish oversight - I *think* the way the account currently handles storage_policy_index is to just update the row/stats for which ever (container, policy) update came in last - so it should be trivial to add to the response | 21:41 |
| hrou | clayg, seems like an easy fix ? :) | 21:41 |
| clayg | hrou: not sure if there's a api version issue with adding a new field | 21:41 |
| clayg | quite possible when we were doing the storage policy work we tried to avoid the whole api question/debate by not adding it and no one thought to loop back around and ask if the question/debate should be raised? | 21:42 |
| clayg | there's always hope swift might get a new api version some day | 21:42 |
| notmyname | lol | 21:42 |
| eranrom | what if we return this only if we have it specifically asked for in the query string? | 21:42 |
| hrou | clayg: there's always one day ;) | 21:42 |
| clayg | eranrom: qs param might be a good way to go | 21:44 |
| hrou | clayg, awesome thanks for confirming I wasn't crazy (given there are backdoors at getting at it there's always workarounds), let me dig into the code see if it is indeed an API problem | 21:44 |
| clayg | eranrom: or at least a way to dodge the issue | 21:44 |
| eranrom | clayg,that is until bumping the api version :-) | 21:45 |
| clayg | eranrom: if you keep coming up with reasonable ways to extend new features into the api without breaking backwards compat we'll *never* get a new api version | 21:45 |
| hrou | lol | 21:46 |
| notmyname | is adding a key a breaking change that we can't do? | 21:46 |
| eranrom | clayg,Any way to delete comments in IRC? | 21:46 |
| clayg | notmyname: not sure | 21:46 |
| clayg | lol! | 21:46 |
| * notmyname goes back to the meeting | 21:46 | |
| hrou | darn it eranrom; just stop being so clever; That happens a lot outside of swift too ; ) | 21:46 |
| *** sudorandom has joined #openstack-swift | 21:47 | |
| clayg | notmyname: I'd guess that json clients would be fine - based on all the ways I could imagine someone might be parsing the json - maybe some new client doesn't work with some old swift cluster if they don't realize it was an *addition* to the account listing response | 21:47 |
| *** cdelatte has joined #openstack-swift | 21:47 | |
| *** dmorita has joined #openstack-swift | 21:47 | |
| *** haomaiwa_ has joined #openstack-swift | 21:47 | |
| clayg | notmyname: xml - i have no idea - there's probably schemas involved and the entire world will melt in a ball of fire | 21:47 |
| notmyname | as is typical for xml | 21:48 |
| eranrom | for the sake of large api versions I will not pose an idea how to overcome the potential xml issue. | 21:50 |
| ahale | you could maybe argue it could fit into the Accept request header | 21:50 |
| *** zhill has quit IRC | 21:53 | |
| *** changbl has quit IRC | 21:55 | |
| *** cdelatte has quit IRC | 21:56 | |
| torgomatic | once we changed the xml output by using a real XML library instead of string concatenation... IIRC, the output was identical except for whitespace | 21:56 |
| torgomatic | we had to revert it because it broke clients | 21:56 |
| torgomatic | "parsing" | 21:56 |
| *** cdelatte has joined #openstack-swift | 21:58 | |
| *** bjkeller has joined #openstack-swift | 22:01 | |
| *** haomaiwa_ has quit IRC | 22:01 | |
| *** haomaiwang has joined #openstack-swift | 22:01 | |
| notmyname | I don't like the idea of adding a storage policy in the output based on a query parameter | 22:02 |
| *** cdelatte has quit IRC | 22:02 | |
| *** cbartz has left #openstack-swift | 22:03 | |
| acoles | clayg: hi, got a few mins? | 22:03 |
| onovy | ho_away: are you here? | 22:04 |
| ho_away | onovy: yup | 22:05 |
| onovy | i want to talk about half of * servers downed | 22:05 |
| onovy | we are using 4 replicas :) | 22:05 |
| eranrom | notmyname,: Now I can sleep well knowing I did not contribute to leaving the API version as is :-) Good night and thanks again! | 22:05 |
| onovy | so if i understand it correctly, post/put/delete to account/container will fail? | 22:05 |
| onovy | when 2 of primary replica is offline? | 22:05 |
| *** cdelatte has joined #openstack-swift | 22:06 | |
| kota_ | eranrom: good night. | 22:06 |
| acoles | eranrom: congrats on the sync patch btw :D | 22:06 |
| ho_away | onovy: as for the post account, if account servers downed, the api will be failed | 22:06 |
| hrou | eranrom, night! | 22:06 |
| eranrom | thanks all! good night | 22:07 |
| *** eranrom has left #openstack-swift | 22:07 | |
| ho_away | onovy: post container is same as post container. | 22:07 |
| onovy | ho_away: but only write operations, post/put/delete, right? | 22:07 |
| onovy | OR only posts are affected? | 22:07 |
| ho_away | onovy: yes. | 22:08 |
| onovy | yes first or second one? :) | 22:08 |
| ho_away | onovy: as for the put/delete container, if account servers dowend the api will be failed | 22:08 |
| ho_away | onovy: first one. | 22:08 |
| ho_away | onovy: replication saves the situation but api failed | 22:09 |
| onovy | hmm, thats critical for us | 22:10 |
| ho_away | onovy: in 4 replicas case it is necessary to down two servers at one time, it's not easy :-) | 22:11 |
| onovy | we have georeplicated cluster. 2+2 | 22:11 |
| onovy | and we will disconnect whole datacenter | 22:11 |
| *** kragniz_ is now known as kragniz | 22:12 | |
| ho_away | onovy: different region in rings? | 22:12 |
| onovy | yes | 22:12 |
| onovy | 2 regions, 2 zones in every region. r=4 | 22:12 |
| onovy | 16 a+c+o servers | 22:13 |
| onovy | so 4 servers / zone | 22:13 |
| acoles | hrou: clayg notmyname the discussion re adding storage policy to listing sounds similar to kota_'s patch to add last modified time | 22:13 |
| ho_away | onovy: i see. if connection b/w regions is lost, it might happen | 22:13 |
| clayg | acoles: sorry i missed you :'( | 22:14 |
| acoles | clayg: i'm still here ! | 22:14 |
| onovy | ho_away: read operations will work, right? | 22:14 |
| ho_away | onovy: right! i tested with 2 replicas env. | 22:14 |
| ho_away | onovy: and object related operations will work too with default config. | 22:15 |
| clayg | acoles: oh neat! you shouldn't be asleep tho. you need your beauty rest before you host the hack-a-thon! | 22:15 |
| clayg | er... should be asleep - w/e | 22:15 |
| acoles | hrou: clayg oops I meant to add a link there, https://review.openstack.org/#/c/198634/14 | 22:15 |
| onovy | so writting to existing containers will work | 22:15 |
| acoles | clayg: you mean i don;t have enough beauty already? ;) | 22:16 |
| acoles | clayg: so here's my question... | 22:16 |
| clayg | heh | 22:16 |
| ho_away | onovy: i think so. | 22:16 |
| *** joeljwright is now known as 17WABLOCX | 22:16 | |
| *** joeljwright has joined #openstack-swift | 22:16 | |
| acoles | clayg: while fixing .meta files never getting reclaimed I discovered that .ts files *may* never get reclaimed | 22:16 |
| onovy | ho_away: ok. thanks for information. i will test it | 22:17 |
| ho_away | onovy: you are welcome! please share your results :-) | 22:18 |
| acoles | clayg: AFAICT reclaim only happens if the suffix is re-hashed, and that only happens if something calls an invalidate_hash for the suffix, so if nothing changes in the suffi then a .ts will never be reclained | 22:18 |
| clayg | acoles: maybe you discovered a *new* way in which ts files may never get reclaimed? because we already have a bug open for the inactive suffixes don't trigger reclaim thing | 22:18 |
| acoles | clayg: aha! | 22:18 |
| clayg | acoles: that sounds liek the old way | 22:18 |
| acoles | clayg: so its a known thing | 22:18 |
| *** joeljwright1 has joined #openstack-swift | 22:18 | |
| onovy | ho_away: we will know 27th - scheduled downtime for datacenter | 22:19 |
| * acoles goes to look for bug | 22:19 | |
| *** 17WABLOCX has quit IRC | 22:19 | |
| clayg | acoles: maybe - lp bug #1301728 | 22:19 |
| openstack | Launchpad bug 1301728 in OpenStack Object Storage (swift) "tombstone (.ts) object will never be deleted if the hash_suffix exists in the hashes.pkl " [Undecided,In progress] https://launchpad.net/bugs/1301728 - Assigned to Brian Cline (briancline) | 22:19 |
| hrou | acoles, thanks for that ! | 22:20 |
| acoles | clayg: yeah that looks like it. thought I had gone insane! | 22:20 |
| onovy | btw: where is Swift recommendation to have even replicas count? | 22:20 |
| acoles | clayg: the do_listdir behavior looks wacky too but thats a different issue i think | 22:21 |
| *** joeljwright has quit IRC | 22:21 | |
| clayg | onovy: I think three is a pretty popular replica count - and my understanding of number systems wouldn't classify that as even? | 22:21 |
| clayg | onovy: that is to say - i'm not sure I"m aware of such a recommendation? | 22:21 |
| onovy | clayg: https://review.openstack.org/#/c/266190/ | 22:21 |
| onovy | Swift recommends number of replicas is odd number but some deployers | 22:21 |
| onovy | use even number because of combination with SDS products. | 22:21 |
| acoles | clayg: that sounds odd :) | 22:21 |
| onovy | don't have even, sry :) | 22:21 |
| *** nadeem has joined #openstack-swift | 22:22 | |
| onovy | so: where is Swift recommendation to have __odd__ replicas count? :) | 22:23 |
| clayg | onovy: we have some implementations that seem to prefer it - there's some debate if quorum on 4 is 3 or 2 | 22:23 |
| clayg | well not really *debate* - the answer according to code/implementation/tests is *3* | 22:23 |
| clayg | but the consequences when you have a replcia count of 2 is decidly annoying :'( | 22:23 |
| onovy | so event replicas count is fine? | 22:23 |
| onovy | because we have 4 and need to have 4 :) | 22:24 |
| clayg | yeah 4 is one better than 3 | 22:24 |
| onovy | not really everytime | 22:24 |
| clayg | 2/3 => 3/4 - that's some strong data protection right there | 22:24 |
| onovy | look to that review | 22:24 |
| clayg | which one? ho has like a dozen | 22:25 |
| onovy | https://review.openstack.org/#/c/266190/ for example | 22:25 |
| clayg | that man is like timburke it's impossible to keep up | 22:25 |
| onovy | :))) | 22:25 |
| acoles | clayg: there's useful discussion on that bug report, thanks for the link. we've seen these lingering tombstones in the wild too. | 22:25 |
| onovy | exactly i'm really woried now | 22:25 |
| *** blmartin has quit IRC | 22:28 | |
| *** zhill has joined #openstack-swift | 22:29 | |
| clayg | acoles: yeah I *thought* we had a plan to move forward with them but not make replication sad - but it got lost | 22:30 |
| clayg | onovy: yeah I'm pretty sure I agree with cschwede that treating the 404's as 204 is gunna have a bad time | 22:31 |
| onovy | from production point | 22:31 |
| onovy | 2+2 replicas (in two regions) | 22:31 |
| acoles | clayg: maybe I'll look at a fix. I'd also thought about having the auditor do it but it would be nice if it all got taken care of in replicator. | 22:31 |
| clayg | onovy: I think hrou's main point was that almost everything sucks when you have two replicas | 22:31 |
| onovy | one region offline | 22:31 |
| clayg | acoles: well i'm down either way | 22:31 |
| clayg | onovy: I think metadata handling in geo clusters is sort of an orthoganal issue tho :\ | 22:32 |
| onovy | clayg: but from this point it sucks to have 4 replicas in two region too | 22:32 |
| acoles | clayg: oh man, you're going to have to remind again, is 'down' positive or negative in your nation's language? | 22:32 |
| clayg | acoles: it can go either way! | 22:32 |
| notmyname | acoles: it's english. don't you know that? | 22:33 |
| clayg | acoles: in this context I meant that it was acceptable to me for either the auditor or the replicator to handle reclimation of tombstones in inactive suffixes | 22:33 |
| onovy | clayg: so simple question: is BAD idea to have geo cluster with 2 replica in each region? | 22:34 |
| acoles | clayg: lol. that's the way, spell it out nice and clear for the slow members of the team ;) | 22:34 |
| clayg | onovy: I think it just sucks to have remote regions in the direct data path for metadata updates | 22:34 |
| *** joeljwright1 has quit IRC | 22:34 | |
| clayg | acoles: hardly - language is hard - esspecially when some people don't type native 'merican! | 22:34 |
| *** joeljwright has joined #openstack-swift | 22:34 | |
| *** ChanServ sets mode: +v joeljwright | 22:34 | |
| *** zhill has quit IRC | 22:35 | |
| clayg | onovy: I think that's fine for object data | 22:35 |
| onovy | clayg: simple question, not too simple answer :) | 22:35 |
| *** joeljwright has quit IRC | 22:35 | |
| *** changbl has joined #openstack-swift | 22:36 | |
| clayg | onovy: but I think metadata layer is less obvious - it was one of the pain points to come up in japan from operators on geo clusters | 22:37 |
| onovy | if you say "metadata" you mean X-Meta headers right? | 22:38 |
| notmyname | no, the account and container | 22:39 |
| clayg | onovy: no - i meant account and container storage | 22:39 |
| onovy | hmm | 22:39 |
| clayg | notmyname: you've got my saio all busted running crappy old eventlet - dork | 22:39 |
| notmyname | I fixed it though. patch 269811 | 22:40 |
| patchbot | notmyname: https://review.openstack.org/#/c/269811/ - Bump eventlet min version to 0.17.4 | 22:40 |
| onovy | but if you say "object data" is fine and container server is not, how can i combine it? | 22:40 |
| notmyname | clayg: or, "you're not running latest nightly code? wow." | 22:40 |
| onovy | i can't add any object without container server... :) | 22:40 |
| clayg | notmyname: :) | 22:41 |
| onovy | our usecase is simple: we have 2 DC for redundancy. Inside DC must be redundancy too. So 4 replicas is "ideal". 2 in every DC | 22:41 |
| ho_away | onovy: clayg: i'm back. i heard if apis response with 50X, users thought the cloud downed so deployers would like to avoid the situation. | 22:42 |
| clayg | yeah i get it - and the account & container layer is going to get on your nerves | 22:42 |
| onovy | we are putting images/videos/garbage inside Swift | 22:42 |
| clayg | ho_away: yup, i get it | 22:42 |
| onovy | clayg: ok, how to avoid it? | 22:42 |
| clayg | jesus - I DON'T KNOW :P | 22:42 |
| clayg | write better geo cluster support for the metadata layer (whatever that looks like) | 22:43 |
| notmyname | onovy: yeah. and you're not the only one who's done that exact setup. but with the containers, you've got 3 or 4 replicas there, and there's no concept of policies. so the replicas are spread across the ring. which means that even with a single-region policy, you might be updating a remote contianer replica | 22:43 |
| onovy | hey :) so how other ones do this? | 22:43 |
| * clayg has ideas - but dick all for getting it done | 22:43 | |
| onovy | notmyname: yep, i'm fine with this. latency is not a problem | 22:43 |
| *** BAKfr has joined #openstack-swift | 22:43 | |
| clayg | onovy: I think "others" mostly gripe (along with us) about the existing implementation is not ideal when we get together at design summits? | 22:44 |
| notmyname | torgomatic has an interesting idea along these lines that involves a ring of rings | 22:44 |
| onovy | but if one region go offline i want to use 2 replicas inside working location and second 2 from handoff | 22:44 |
| onovy | so all 4 replicas will go from one location | 22:44 |
| clayg | notmyname: orthoganal I think | 22:44 |
| notmyname | if latency isn't a problem, then what's the problem? :-) | 22:45 |
| onovy | notmyname: availability | 22:45 |
| *** acoles is now known as acoles_ | 22:46 | |
| onovy | https://review.openstack.org/#/c/266190/ // this review says: when one of your region will be offline (which is half of replicas) you can write | 22:46 |
| onovy | can't | 22:46 |
| *** jamielennox|away is now known as jamielennox | 22:47 | |
| clayg | onovy: I think that's just restating what you've already pointed out | 22:47 |
| notmyname | yeah. I think ho_away is actually trying to quantify and solve the problem | 22:47 |
| onovy | clayg: yep. but john asked what's the problem | 22:47 |
| *** BAKfr has quit IRC | 22:47 | |
| notmyname | as if I know what's even going on ;-) | 22:48 |
| onovy | so conclusion: it's not bad setup if latency is not a problem and when ho fixies this, i'm fine | 22:48 |
| onovy | right? :) | 22:48 |
| onovy | and i will be fine event if whole region is offline | 22:49 |
| onovy | *even | 22:49 |
| notmyname | seems like the current question at hand is for us (all of us) to say if http://paste.openstack.org/show/484089/ is what's intended. and if not, then we've got the patches to fix it | 22:49 |
| *** asettle has joined #openstack-swift | 22:51 | |
| clayg | notmyname: cschwede has looked at those patches more closely than I have - but his -1's seem to read "this changes the behavior, but the consequnses are not desirable" | 22:52 |
| clayg | i could make *every* request reutrn 2XX | 22:52 |
| clayg | it wouldn't even that big of middleware | 22:52 |
| *** diogogmt has quit IRC | 22:52 | |
| clayg | ok, well good - shut that discussion right down - wtg clayg | 22:55 |
| notmyname | no, I just havent' looked at the patches. and I just got an interesting email reply from someone who's been doing some code analysis on swift | 22:57 |
| notmyname | and that distracted me | 22:57 |
| *** m_kazuhiro has joined #openstack-swift | 22:58 | |
| ho_away | clayg: sorry i'm preparing to go office, can we share the impact of problem? | 22:58 |
| *** mrmoje has joined #openstack-swift | 22:59 | |
| clayg | ho_away: honestly I think the impact of the problem is understood, but cschwede's review is sound - the approach/implementation in those patches is problamatic | 22:59 |
| onovy | sry must go sleep now. it's midnight here | 22:59 |
| onovy | gn | 23:00 |
| clayg | I also think the specific situation with a 2-region-4-replica ring is more interesting than 2 replica metadata rings - and we should be careful trying to trivialize them as equivilant | 23:00 |
| ho_away | clays: i understand. it's same as object's behavior. is there any histrical reason why account and container don't take care of it. | 23:00 |
| *** haomaiwang has quit IRC | 23:01 | |
| *** haomaiwa_ has joined #openstack-swift | 23:01 | |
| ho_away | clayg: i don't find doc but a potential deployer said there is a doc for two replica example in swiftstack. | 23:02 |
| ho_away | onovy: good night! | 23:02 |
| clayg | ho_away: that's one way to say it - the storage implementation is different - and they've evolved at different rates - I would say the object server handling for geo clusters is much more sofisticated than the metadata controllers - but the same strategies won't apply one-to-one | 23:02 |
| *** bhakta has quit IRC | 23:03 | |
| clayg | ho_away: that might be true, but even best practices of today may not lead to ideal results - always room to improve! | 23:03 |
| *** saltsa_ has quit IRC | 23:04 | |
| *** joeljwright has joined #openstack-swift | 23:04 | |
| *** ChanServ sets mode: +v joeljwright | 23:04 | |
| joeljwright | timburke: you available? | 23:04 |
| timburke | joeljwright: yep | 23:05 |
| *** saltsa has joined #openstack-swift | 23:05 | |
| joeljwright | Just wanted to revisit patch 184957 - it seems to have seen some action while I was away enjoying myself | 23:05 |
| patchbot | joeljwright: https://review.openstack.org/#/c/184957/ - Add --decode-content option | 23:05 |
| *** bhakta has joined #openstack-swift | 23:05 | |
| *** mragupat_ has quit IRC | 23:05 | |
| joeljwright | last time we spoke about it you weren't convinced it was something we should merge… | 23:06 |
| *** zhiyan has quit IRC | 23:06 | |
| ho_away | clayg: thanks for the explanation. do you have any idea to solve the situation? | 23:06 |
| *** dmorita has quit IRC | 23:07 | |
| clayg | ho_away: idk, your probably is specifically 2-region-4-replica yah? | 23:08 |
| clayg | *problem | 23:08 |
| *** zhiyan has joined #openstack-swift | 23:08 | |
| timburke | joeljwright: well, i've tried to make sure it was always in a state where i'd be content to see it merged, but i've always been a bit skeptical that it's in-scope for swiftclient. i'd still like to know what the community thinks; notmyname seemed optimistic about it, as i recall, but more voices are better | 23:08 |
| ho_away | clayg: and also 2-region-2-replica | 23:08 |
| ho_away | clayg: the problem comes from 2-region-2-replica with SDS env. | 23:09 |
| clayg | I'm not sure I understand the use-case for less than 3 replicas in the metadata layer | 23:09 |
| clayg | the risk cost curve doesn't make sense to me | 23:10 |
| joeljwright | timburke: maybe something to stuff under people's noses at the hackathon and extract an opinion by force? | 23:10 |
| joeljwright | :) | 23:10 |
| timburke | sounds like a plan | 23:11 |
| clayg | timburke: you should totally prepare a fire-side chat on "content/transfer encoding for dummies" | 23:12 |
| ho_away | clayg: yeah, they would like to have same num of replica in one region because of performance (want to have same performance) | 23:12 |
| joeljwright | ok, I have to disappear - have a good day peeps | 23:13 |
| timburke | clayg: i definitely know more about it now than i ever really cared to :P | 23:13 |
| *** blair has quit IRC | 23:13 | |
| *** joeljwright has quit IRC | 23:13 | |
| timburke | clayg: though it turns out transfer-encoding isn't in the purview of WSGI filters/applications. not that i've let that stop me; i just break the spec | 23:14 |
| clayg | timburke: but the wsgi server could still support/handle it yeah? | 23:15 |
| timburke | clayg: yeah. it'd mean some eventlet patches | 23:16 |
| timburke | (if we wanted to do it properly) | 23:16 |
| clayg | timburke: i doubt that's the case ;) | 23:16 |
| *** blair has joined #openstack-swift | 23:17 | |
| *** km__ has joined #openstack-swift | 23:17 | |
| *** km__ is now known as Guest79571 | 23:18 | |
| *** bjkeller has left #openstack-swift | 23:19 | |
| *** nadeem has quit IRC | 23:20 | |
| *** awelleck has quit IRC | 23:21 | |
| timburke | clayg: in that case... https://gist.github.com/tipabu/7c8ed7a713a7f51f20c2 (there's a wsgi_input hack around L286) | 23:21 |
| openstackgerrit | John Dickinson proposed openstack/swift: remove unneeded duplicate dict keys https://review.openstack.org/270523 | 23:21 |
| notmyname | totally critical patch there ^ /s | 23:21 |
| *** diazjf has quit IRC | 23:22 | |
| *** ho_away has quit IRC | 23:23 | |
| timburke | notmyname: makes me think of my recent patch 269399 and patch 269891 | 23:23 |
| patchbot | timburke: https://review.openstack.org/#/c/269399/ - Get rid of FakeConn cruft (MERGED) | 23:24 |
| patchbot | timburke: https://review.openstack.org/#/c/269891/ - Prevent test runs from cluttering current directory | 23:24 |
| clayg | look at you two trying to make things better | 23:24 |
| clayg | so cute | 23:24 |
| *** kei_yama has joined #openstack-swift | 23:30 | |
| *** cdelatte has quit IRC | 23:32 | |
| openstackgerrit | Catherine Northcott proposed openstack/swift: Let developers/operators add watchers to account audit https://review.openstack.org/268830 | 23:42 |
| *** dmorita has joined #openstack-swift | 23:46 | |
| *** BAKfr has joined #openstack-swift | 23:47 | |
| *** chsc has quit IRC | 23:47 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!