| *** urth has quit IRC | 00:01 | |
| clayg | TIL: we have an "examples" TLD | 00:04 |
|---|---|---|
| *** chlong has quit IRC | 00:05 | |
| *** urth has joined #openstack-swift | 00:05 | |
| *** garthb_ has quit IRC | 00:13 | |
| *** nadeem has joined #openstack-swift | 00:31 | |
| *** itlinux has quit IRC | 00:36 | |
| *** clu_ has quit IRC | 00:43 | |
| *** pgbridge has quit IRC | 00:45 | |
| *** janonymous has quit IRC | 00:47 | |
| *** Guest40083 has quit IRC | 00:49 | |
| *** sc has joined #openstack-swift | 00:50 | |
| *** tdasilva has quit IRC | 00:50 | |
| *** sc is now known as Guest28457 | 00:50 | |
| *** tdasilva has joined #openstack-swift | 00:52 | |
| *** suyash has quit IRC | 00:57 | |
| *** gyee has quit IRC | 01:03 | |
| *** lyrrad has quit IRC | 01:11 | |
| *** Jeffrey4l has joined #openstack-swift | 01:29 | |
| *** chlong has joined #openstack-swift | 01:30 | |
| *** papercup has left #openstack-swift | 01:42 | |
| *** MVenesio has joined #openstack-swift | 01:45 | |
| *** MVenesio has quit IRC | 01:50 | |
| *** ukaynar has quit IRC | 01:59 | |
| *** nadeem has quit IRC | 02:17 | |
| *** vinsh has joined #openstack-swift | 02:22 | |
| *** daemontool_ has quit IRC | 03:04 | |
| *** itlinux has joined #openstack-swift | 03:18 | |
| openstackgerrit | Brian Ober proposed openstack/swift: Storage Account Object Count Quota Support https://review.openstack.org/312603 | 03:26 |
| *** dmorita_ has joined #openstack-swift | 03:32 | |
| *** amit213 has quit IRC | 03:33 | |
| *** amit213 has joined #openstack-swift | 03:33 | |
| *** dmorita has quit IRC | 03:34 | |
| *** nadeem has joined #openstack-swift | 03:37 | |
| *** asettle has quit IRC | 03:42 | |
| *** asettle has joined #openstack-swift | 03:42 | |
| *** MVenesio has joined #openstack-swift | 03:46 | |
| *** asettle has quit IRC | 03:47 | |
| *** MVenesio has quit IRC | 03:51 | |
| *** esker has quit IRC | 03:55 | |
| *** ppai has joined #openstack-swift | 03:57 | |
| *** links has joined #openstack-swift | 04:02 | |
| *** mingdang1 has joined #openstack-swift | 04:18 | |
| *** ppai has quit IRC | 04:24 | |
| *** csmart is now known as grammer_nazi | 04:38 | |
| *** grammer_nazi is now known as csmart | 04:38 | |
| *** asettle has joined #openstack-swift | 04:47 | |
| openstackgerrit | John Dickinson proposed openstack/swift: Rework the contributor docs https://review.openstack.org/301871 | 05:02 |
| *** mingdang1 has quit IRC | 05:14 | |
| *** janonymous has joined #openstack-swift | 05:14 | |
| *** SkyRocknRoll has quit IRC | 05:22 | |
| *** ppai has joined #openstack-swift | 05:27 | |
| *** zaitcev has quit IRC | 05:34 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:36 | |
| *** MVenesio has joined #openstack-swift | 05:47 | |
| *** MVenesio has quit IRC | 05:51 | |
| *** janonymous has quit IRC | 05:59 | |
| *** klrmn has quit IRC | 06:12 | |
| *** mingdang1 has joined #openstack-swift | 06:14 | |
| ahale | clayg: hah | 06:16 |
| *** mingdang1 has quit IRC | 06:19 | |
| *** zul has quit IRC | 06:21 | |
| *** asettle has quit IRC | 06:28 | |
| *** janonymous has joined #openstack-swift | 06:28 | |
| *** jmccarthy has quit IRC | 06:35 | |
| openstackgerrit | Matthew Oliver proposed openstack/swift: Set quorum_size of 50% for even replicas and majority_size https://review.openstack.org/313245 | 06:46 |
| mattoliverau | ^^ This change may still require for discussion, but am raising it so I don't forget | 06:46 |
| mattoliverau | about it in the sharding context. If it requires a WIP, then that's fine too. | 06:46 |
| mattoliverau | torgomatic: ^^ I think that covers what was discussed at summit. Or have I missed anything? | 06:47 |
| *** ppai has quit IRC | 06:52 | |
| *** janonymous has quit IRC | 07:05 | |
| *** ppai has joined #openstack-swift | 07:06 | |
| *** permalac has joined #openstack-swift | 07:06 | |
| *** jmccarthy has joined #openstack-swift | 07:09 | |
| *** itlinux has quit IRC | 07:14 | |
| *** ankur has joined #openstack-swift | 07:22 | |
| ankur | hello | 07:23 |
| ankur | the command : swift-dispersion-populate throws exception on RHEL 1.7, it says "BAD REQUEST" | 07:25 |
| ankur | below is the conf file i.e dispersion.conf | 07:26 |
| ankur | conf file for swift-dispersion command https://www.irccloud.com/pastebin/Zp9va9Ke/dispersion.conf | 07:26 |
| ankur | can anyone please suggest anything ? | 07:27 |
| *** dukhlov has joined #openstack-swift | 07:27 | |
| clayg | ankur: idk, looks legi - you're using tempauth or keystone? | 07:32 |
| *** ntt has joined #openstack-swift | 07:34 | |
| clayg | *legit | 07:34 |
| clayg | is it a swiftclient.exceptions.ClientException -> 400 Bad Request? | 07:35 |
| ntt | Hi, I have a problem with rsync and swift (kilo) on centos 7. When I restart the storage node, rsyncd doesn't start and I have an error message: "bind() failed: Cannot assign requested address (address-family 2) unable to bind any inbound sockets on port 873". But if I start manually rsyncd with "systemctl start rsyncd" all works well. Someone can help me please? | 07:35 |
| clayg | ankur: it seems to conditionally import swiftclient get_auth when available - so you might try installing python-swiftclient package via pip | 07:35 |
| clayg | ah the joys of systemd | 07:36 |
| clayg | ntt: AF2 is ipv6 yeah? | 07:36 |
| *** dukhlov has quit IRC | 07:37 | |
| *** openstackstatus has quit IRC | 07:38 | |
| ntt | clayg: AF2?? | 07:40 |
| ntt | mmm | 07:40 |
| *** openstackstatus has joined #openstack-swift | 07:40 | |
| *** ChanServ sets mode: +v openstackstatus | 07:40 | |
| clayg | address-family 2 | 07:41 |
| ntt | clayg: ok, thank you. But I don't need ipv6 on my storage server | 07:41 |
| ntt | ipv6 is enabled by default on rsync? Actually I completely disabled ipv6 on network interfaces... so I think that's the problem | 07:41 |
| ankur | yes it is a swiftclient exeption | 07:43 |
| ankur | swiftclient.exceptions.ClientException: Auth GET failed: http://192.168.52.131:8080/auth/v1.0 400 Bad Request [first 60 chars of response] <html><h1>Bad Request</h1><p>The server could not comply wit | 07:43 |
| *** acoles_ is now known as acoles | 07:45 | |
| ankur | clayg : actually i have installed python-swiftclient via setup.py | 07:45 |
| clayg | ankur: yeah, and what auth service? | 07:47 |
| *** MVenesio has joined #openstack-swift | 07:48 | |
| clayg | ntt: either that or just like networking isn't running yet when it tries to start rsyncd - systemd is all into that sort of timing issue/error tho - i'm sure someone on the internet would have written down how systemd works somewhere... | 07:48 |
| ankur | clayg : i think its keystone , | 07:50 |
| ankur | clayg : is there a way to confirm it ? | 07:51 |
| clayg | ankur: well if the auth url you normally use is v2.0 or v3.0 - then it's *probably* keystone | 07:52 |
| *** MVenesio has quit IRC | 07:52 | |
| *** daemontool has joined #openstack-swift | 07:53 | |
| clayg | ankur: doesn't the dispersion.conf-example have some details about different auth endpoints - projects and domains and all that jazz - can you you just crib from that and fill in w/e you *normally* use to talk to your cluster? | 07:53 |
| acoles | good morning | 07:53 |
| *** chlong has quit IRC | 07:54 | |
| ankur | clayg : no, the dispersion.conf-example doesn't have details about different auth endpoints | 07:56 |
| acoles | ankur: https://github.com/openstack/swift/blob/master/etc/dispersion.conf-sample | 07:58 |
| clayg | ankur: https://github.com/openstack/swift/blob/master/etc/dispersion.conf-sample | 07:58 |
| clayg | hrph | 07:58 |
| acoles | lol | 07:58 |
| clayg | ankur: neway - maybe you're running an old version | 07:59 |
| clayg | good morning acoles - i'll let you take over from here ;) | 07:59 |
| *** ppai has quit IRC | 08:00 | |
| *** mkrcmari__ has joined #openstack-swift | 08:05 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: crypto - add functional test to verify container listing detail https://review.openstack.org/307121 | 08:05 |
| *** mmcardle has joined #openstack-swift | 08:05 | |
| *** mvk_ has quit IRC | 08:08 | |
| *** dmk0202 has joined #openstack-swift | 08:08 | |
| *** ppai has joined #openstack-swift | 08:12 | |
| openstackgerrit | Merged openstack/swift: Rework the contributor docs https://review.openstack.org/301871 | 08:16 |
| *** dukhlov has joined #openstack-swift | 08:28 | |
| openstackgerrit | Merged openstack/swift: Use smaller quorum size in proxy for even numbers of replicas https://review.openstack.org/310608 | 08:29 |
| *** jistr has joined #openstack-swift | 08:32 | |
| *** sorrison_laptop has joined #openstack-swift | 08:33 | |
| *** janonymous has joined #openstack-swift | 08:33 | |
| *** dmorita_ has quit IRC | 08:34 | |
| acoles | clayg: <begging> this has +2 from torgomatic, if you have chance could you review? patch 268251 </begging> | 08:35 |
| patchbot | acoles: https://review.openstack.org/#/c/268251/ - swift - Reclaim isolated .meta files | 08:35 |
| *** dmorita has joined #openstack-swift | 08:35 | |
| acoles | clayg: I would really hope this one does not require 5 hours! | 08:35 |
| *** dukhlov has quit IRC | 08:36 | |
| *** pcaruana has joined #openstack-swift | 08:36 | |
| *** sorrison_laptop has quit IRC | 08:37 | |
| *** sorrison_laptop has joined #openstack-swift | 08:37 | |
| acoles | any core - this should be an easy (albeit tedious) review - patch 290391 | 08:38 |
| patchbot | acoles: https://review.openstack.org/#/c/290391/ - swift - Rename hash_cleanup_listdir tests | 08:38 |
| *** dukhlov has joined #openstack-swift | 08:39 | |
| *** mkrcmari__ has quit IRC | 08:43 | |
| *** dukhlov has quit IRC | 08:45 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: crypto - align cryptography version requirement with global https://review.openstack.org/313281 | 08:52 |
| *** jordanP has joined #openstack-swift | 08:55 | |
| *** tesseract has joined #openstack-swift | 08:56 | |
| *** tesseract is now known as Guest42930 | 08:56 | |
| *** sorrison_laptop has quit IRC | 09:01 | |
| *** nadeem has quit IRC | 09:10 | |
| *** joeljwright has joined #openstack-swift | 09:15 | |
| *** ChanServ sets mode: +v joeljwright | 09:15 | |
| *** mkrcmari__ has joined #openstack-swift | 09:28 | |
| *** dukhlov has joined #openstack-swift | 09:28 | |
| *** asettle has joined #openstack-swift | 09:35 | |
| *** asettle has quit IRC | 09:39 | |
| openstackgerrit | Merged openstack/swift: crypto - only set body crypto meta if body is read and encrypted https://review.openstack.org/305794 | 09:43 |
| *** dukhlov has quit IRC | 09:43 | |
| *** dukhlov has joined #openstack-swift | 09:48 | |
| *** dukhlov has quit IRC | 09:48 | |
| *** MVenesio has joined #openstack-swift | 09:49 | |
| *** MVenesio has quit IRC | 09:53 | |
| *** ekarlso has quit IRC | 10:08 | |
| *** mingdang1 has joined #openstack-swift | 10:21 | |
| *** psachin has joined #openstack-swift | 10:23 | |
| *** chuck__ has joined #openstack-swift | 10:25 | |
| *** mingdang1 has quit IRC | 10:26 | |
| *** mingdang1 has joined #openstack-swift | 10:39 | |
| *** Shashikant86 has joined #openstack-swift | 10:40 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 10:48 | |
| *** sanchitmalhotra has quit IRC | 10:49 | |
| *** dmk0202 has quit IRC | 10:57 | |
| *** dmk0202 has joined #openstack-swift | 10:57 | |
| *** asettle has joined #openstack-swift | 10:59 | |
| *** asettle has quit IRC | 10:59 | |
| *** chuck__ has quit IRC | 11:06 | |
| *** jmccarthy has quit IRC | 11:15 | |
| *** jmccarthy has joined #openstack-swift | 11:17 | |
| *** mingdang1 has quit IRC | 11:23 | |
| *** chuck__ has joined #openstack-swift | 11:28 | |
| *** Shashikant86 has quit IRC | 11:28 | |
| *** MVenesio has joined #openstack-swift | 11:49 | |
| *** MVenesio has quit IRC | 11:54 | |
| *** MVenesio has joined #openstack-swift | 11:58 | |
| *** mingdang1 has joined #openstack-swift | 11:58 | |
| acoles | jrichli: mahatic fyi I am starting thsi task https://trello.com/c/Dq2dZvbP/89-change-to-use-random-object-key-that-is-wrapped-with-the-container-key | 12:00 |
| *** ChanServ sets mode: +v tdasilva | 12:00 | |
| *** asettle has joined #openstack-swift | 12:07 | |
| *** Jeffrey4l has quit IRC | 12:17 | |
| *** ekarlso has joined #openstack-swift | 12:18 | |
| *** Shashikant86 has joined #openstack-swift | 12:21 | |
| *** Shashikant86 has quit IRC | 12:23 | |
| *** Jeffrey4l has joined #openstack-swift | 12:28 | |
| tdasilva | good morning | 12:30 |
| *** NM has joined #openstack-swift | 12:32 | |
| acoles | tdasilva: good morning | 12:42 |
| tdasilva | acoles: hello o/ | 12:43 |
| tdasilva | acoles: boaty lives on: http://www.bbc.com/news/uk-36225652 | 12:43 |
| ntt | clayg and other people interested: The problem with rsync and centos7 is related to the configuration parameter "address". If I don't use "address", rsyncd starts on boot and it works on all interfaces | 12:44 |
| hrou | notmyname - thanks for the explanation, that makes sense ! WIP removed ! | 12:44 |
| ntt | I'd like to bind rsyncd just on the interface connected to the replication network | 12:44 |
| acoles | tdasilva: yes! | 12:45 |
| *** SkyRocknRoll has quit IRC | 12:46 | |
| *** asettle has quit IRC | 12:49 | |
| *** alejandrito has joined #openstack-swift | 12:52 | |
| *** cguadall__ has quit IRC | 13:00 | |
| *** ppai has quit IRC | 13:06 | |
| *** psachin has quit IRC | 13:06 | |
| *** ppai has joined #openstack-swift | 13:19 | |
| *** klamath has joined #openstack-swift | 13:24 | |
| *** pcaruana has quit IRC | 13:25 | |
| *** ankur has quit IRC | 13:27 | |
| *** BigWillie has joined #openstack-swift | 13:34 | |
| pdardeau | good morning | 13:46 |
| pdardeau | notmyname: awesome work on the new contributor docs! | 13:47 |
| *** chuck__ is now known as zul | 13:57 | |
| *** ametts has joined #openstack-swift | 14:00 | |
| *** alejandrito has quit IRC | 14:05 | |
| *** catintheroof has joined #openstack-swift | 14:05 | |
| *** pauloewerton has joined #openstack-swift | 14:05 | |
| *** ppai has quit IRC | 14:07 | |
| *** _JZ_ has joined #openstack-swift | 14:10 | |
| *** catintheroof has quit IRC | 14:15 | |
| -openstackstatus- NOTICE: Zuul is currently recovering from a large number of changes, it will take a few hours until your job is processed. Please have patience and enjoy a great weekend! | 14:16 | |
| *** links has quit IRC | 14:17 | |
| sgundur- | notmyname: morning, I changed to maintain uniform port range (but some random ports in 6000 range r still in use) | 14:22 |
| sgundur- | notmyname: and I mistook last comment on torgomatic's patch was suggesting at making these changes. | 14:22 |
| *** catintheroof has joined #openstack-swift | 14:27 | |
| catintheroof | guys, one quick question, does swift from kilo and on, supports to upload large files without having to create the manifest manually ? | 14:29 |
| glange | "swift" the client? | 14:29 |
| catintheroof | glange, nope, the api | 14:30 |
| glange | there are two types of large objects, Dynamic Large Objects and Static Large Objects | 14:31 |
| glange | Static needs a manifest with a list of all the objects that make up the larger object | 14:31 |
| glange | it's the one you should probably be using mostly because the other one depends on container listings being correct | 14:31 |
| notmyname | good morning | 14:32 |
| glange | catintheroof: http://docs.openstack.org/developer/swift/overview_large_objects.html | 14:32 |
| glange | that describes teh two types | 14:32 |
| glange | I'm not sure exactly where those features got added | 14:32 |
| notmyname | catintheroof: the swift api doesn't split up large files for you automatically. the CLI/SDK can do that, but the server doesn't auto split large objects on upload | 14:33 |
| glange | catintheroof: both types need a manifest but for dynamic large objects, the objects aren't actually listed in it | 14:33 |
| catintheroof | notmyname, glange thanks so much guys, swift answers as allways | 14:34 |
| glange | haha "swift" answers | 14:34 |
| notmyname | catintheroof: you weren't in austin were you? I saw max (briefly), but I didn't see you | 14:34 |
| *** mingdang1 has quit IRC | 14:35 | |
| catintheroof | notmyname, haha! i wasnt!! would love to, but we need to split! hope to see you @ barcelona but i think Mr. T is gonna be there | 14:36 |
| jrichli | tdasilva acoles: that boaty thing makes me lol every time I think about it! | 14:45 |
| acoles | me too | 14:45 |
| tdasilva | acoles, jrichli: watched your talk this morning, great job! | 14:46 |
| *** itlinux has joined #openstack-swift | 14:46 | |
| jrichli | thanks! acoles was awesome | 14:46 |
| -openstackstatus- NOTICE: Zuul has been restarted. As a results, we only preserved patches in the gate queue. Be sure to recheck your patches in gerrit if needed. | 14:47 | |
| *** itlinux has quit IRC | 14:51 | |
| *** ukaynar has joined #openstack-swift | 14:58 | |
| *** lcurtis has joined #openstack-swift | 15:00 | |
| *** Guest42930 has quit IRC | 15:05 | |
| *** david-lyle has joined #openstack-swift | 15:06 | |
| *** arch-nemesis has joined #openstack-swift | 15:07 | |
| *** klrmn has joined #openstack-swift | 15:08 | |
| *** ukaynar has quit IRC | 15:08 | |
| *** ukaynar has joined #openstack-swift | 15:09 | |
| *** itlinux has joined #openstack-swift | 15:22 | |
| mahatic | acoles: oh okay. I was about to grab it, but i'm sure you'll do it quicker | 15:27 |
| pdardeau | hey mahatic | 15:28 |
| mahatic | pdardeau: hello | 15:28 |
| mahatic | acoles: would that task also involve this one: https://trello.com/c/SqcHL7YZ/90-change-the-etag-currently-encrypted-with-object-key-to-instead-use-a-listing-container-key | 15:28 |
| pdardeau | mahatic: how was your post-summit tx travels? | 15:29 |
| mahatic | pdardeau: exhausting! landing home on tue is a bad idea, no time to recover from jet lag, but now i'm better, thanks. your bus ride was good? | 15:30 |
| pdardeau | mahatic: nice! my travels were good too | 15:31 |
| *** jordanP has quit IRC | 15:31 | |
| acoles | mahatic: no i left that one alone because i saw your name on the trello card | 15:32 |
| acoles | mahatic: but its likely there will some key wrapping related helper methods that you'll benefit from | 15:32 |
| acoles | mahatic: i hope to push a first cut to gerrit today | 15:32 |
| mahatic | acoles: oh great! | 15:32 |
| *** dmk0202 has quit IRC | 15:33 | |
| *** dmk0202 has joined #openstack-swift | 15:34 | |
| mahatic | pdardeau: river walk in san an was nice | 15:35 |
| *** MVenesio has quit IRC | 15:35 | |
| *** catintheroof has quit IRC | 15:36 | |
| pdardeau | mahatic: glad you liked it. that's our big attraction (aside from the rodeos, of course) | 15:36 |
| *** dmk0202 has quit IRC | 15:39 | |
| *** Shashikant86 has joined #openstack-swift | 15:47 | |
| *** lyrrad has joined #openstack-swift | 15:50 | |
| *** lakshmiS has joined #openstack-swift | 15:56 | |
| *** nadeem has joined #openstack-swift | 16:04 | |
| *** nadeem has quit IRC | 16:05 | |
| *** saltsa has quit IRC | 16:08 | |
| *** saltsa has joined #openstack-swift | 16:08 | |
| *** nadeem has joined #openstack-swift | 16:09 | |
| *** dmellado is now known as dmellado|off | 16:10 | |
| *** klrmn has quit IRC | 16:17 | |
| *** Shashikant86 has quit IRC | 16:19 | |
| *** treyd has quit IRC | 16:25 | |
| *** timur has quit IRC | 16:25 | |
| *** dabukalam has quit IRC | 16:25 | |
| *** dabukalam has joined #openstack-swift | 16:25 | |
| *** catintheroof has joined #openstack-swift | 16:27 | |
| *** timur has joined #openstack-swift | 16:27 | |
| *** jmccarthy has quit IRC | 16:28 | |
| *** mtreinish has quit IRC | 16:28 | |
| *** bsdkurt has quit IRC | 16:29 | |
| *** jidar has quit IRC | 16:29 | |
| *** treyd has joined #openstack-swift | 16:33 | |
| *** mtreinish has joined #openstack-swift | 16:35 | |
| *** jidar has joined #openstack-swift | 16:40 | |
| *** dmk0202 has joined #openstack-swift | 16:40 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: crypto - wrap random body key https://review.openstack.org/313640 | 16:42 |
| acoles | mahatic: you may want to base on that ^^ | 16:42 |
| *** bsdkurt has joined #openstack-swift | 16:42 | |
| notmyname | dfg_: did you have a chance to make any graphs? | 16:44 |
| *** acoles is now known as acoles_ | 16:45 | |
| clayg | I'm observing this weird behavior that if I send a DELETE with an x-timestamp (gatekeeper shunt timestamp off) *before* ~2016-04-30 00:06:40 the .data file get's deleted, the tombstone row gets placed in the db, but no tombstone file is left on disk? | 16:48 |
| *** garthb has joined #openstack-swift | 16:50 | |
| clayg | like 1461000000.00000 (2016-04-18 10:20:00) => no tombstone - but 1462000000.00000 (2016-04-30 00:06:40) => tombstone? | 16:50 |
| notmyname | something weird with reclaim age maybe? | 16:51 |
| *** gyee has joined #openstack-swift | 16:52 | |
| notmyname | timburke probably knows more than me about this | 16:53 |
| clayg | maybe it's the patch i'm testing? | 16:53 |
| *** dmk0202 has quit IRC | 16:55 | |
| clayg | oh, yeah i guess this may always be the case - the cleanup hashdir has to run after ever put delete - and the reclaim reaps the tombstone | 16:55 |
| clayg | so x-timestamp with delete's older than reclaim are *very* likely to leave dark data in the cluster - who knew! | 16:56 |
| notmyname | but of course, the reclaim again /should/ be longer than both the replication cycle time and the expiry cycle time. | 16:56 |
| clayg | notmyname: don't forget container-sync! | 16:56 |
| clayg | notmyname: new rule - set reclaim age to infinity - why are you deleting things anyway?! | 16:57 |
| *** dmk0202 has joined #openstack-swift | 16:57 | |
| notmyname | I've said it for a long time: implementing the DELETE verb was probably a mistake. just turn it off, and you shouldn't delete things anyway ;-) | 16:57 |
| redbo | It would be a lot easier to not implement GET | 16:58 |
| notmyname | torgomatic did that for us at the beginning of last month ;-) | 16:58 |
| notmyname | it's a little scary, but I can imagine having these sorts of timeouts all set to "auto" and they have some back pressure from the cycle times of different things in the cluster. | 16:59 |
| notmyname | on the one hand, "magic". on the other, lots of footguns without it | 16:59 |
| notmyname | I mean, we actually *know* that reclaim age must be longer than the replication cycle time. and we log the replication cycle time. so why do we make ops track all that and set reclaim age to something "safe" (and "safe" probably changes as HW fails or load changes!) | 17:01 |
| notmyname | on the other hand, there's lots of opportunity for cascading spirals of horror and pain with too much automation | 17:01 |
| *** dmk0202 has quit IRC | 17:01 | |
| ahale | ops need jobs too :/ | 17:02 |
| clayg | lol | 17:02 |
| ahale | btw clayg , been poking around in rings today with a new script, vaguely interesting.. | 17:03 |
| ahale | over time, it seems to favour moving some parts far more than others, im still pretty confused at what im doing though | 17:04 |
| ahale | gotta try with some current ringbuilder ring.gz's and see what happens | 17:06 |
| *** jistr has quit IRC | 17:07 | |
| *** tamizh_geek has quit IRC | 17:09 | |
| clayg | ahale: ok, good luck | 17:10 |
| clayg | ahale: it's a fairly imprecise tool - but the idea was you could keep tweeking pretend_min_part_hours_passed <ratio> and revert back to the old builder if you're not happy with the rebalance | 17:11 |
| clayg | obs not for the faint of heart | 17:11 |
| ahale | yeah :) i havent had a chance to poke around in it beyond reading the change. but it seems pretty sensible | 17:12 |
| *** klrmn has joined #openstack-swift | 17:15 | |
| ahale | like, looking over 70 ring changes with script im making, with 2^20 parts.. 3k parts never changed any primary dev at all, while one had its 1st primary changed 18 times | 17:15 |
| ahale | seems kinda weird :) | 17:16 |
| ahale | well not really, cos its over a long time so min_part_hours is probably ignored | 17:16 |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Rename hash_cleanup_listdir tests https://review.openstack.org/290391 | 17:20 |
| *** silor has joined #openstack-swift | 17:21 | |
| clayg | timburke: good call on cleaning up the comments - stick a +A on there! | 17:21 |
| timburke | done and done | 17:22 |
| *** mmcardle has quit IRC | 17:25 | |
| *** Jeffrey4l has quit IRC | 17:28 | |
| *** klamath has quit IRC | 17:31 | |
| *** silor has quit IRC | 17:36 | |
| *** tamizh_geek has joined #openstack-swift | 17:37 | |
| *** zaitcev has joined #openstack-swift | 17:46 | |
| *** ChanServ sets mode: +v zaitcev | 17:46 | |
| *** nadeem has quit IRC | 17:56 | |
| *** sgundur1 has quit IRC | 17:57 | |
| *** ntata has quit IRC | 17:57 | |
| *** yarkot1 has quit IRC | 17:57 | |
| *** pdardeau| has quit IRC | 17:57 | |
| *** mmotiani_ has quit IRC | 17:57 | |
| *** mmotiani_ has joined #openstack-swift | 17:58 | |
| *** ntata has joined #openstack-swift | 17:59 | |
| *** yarkot1 has joined #openstack-swift | 18:07 | |
| *** suyash has joined #openstack-swift | 18:08 | |
| *** dukhlov has joined #openstack-swift | 18:11 | |
| *** dukhlov has quit IRC | 18:13 | |
| *** BigWillie has quit IRC | 18:24 | |
| *** BigWillie has joined #openstack-swift | 18:26 | |
| *** klamath has joined #openstack-swift | 18:26 | |
| *** mmcardle has joined #openstack-swift | 18:29 | |
| *** dukhlov has joined #openstack-swift | 18:30 | |
| notmyname | so if https://review.openstack.org/#/c/312718/ lands, then I am very very much not interested in a tempest plugin to run our tests. this resolution seems to remove t100% of the incentive to write a plugin | 18:31 |
| patchbot | notmyname: patch 312718 - governance - add resolution explaining which tests we think def... | 18:31 |
| tdasilva | notmyname: tldr :-) | 18:37 |
| tdasilva | what's the summary? | 18:37 |
| notmyname | tldr is that defcore tests should be hosted and managed in the tempest repo, not in another project loaded via plugins | 18:38 |
| tdasilva | ok, they want a centralized repo | 18:39 |
| notmyname | right. owned and managed by the tempest team | 18:39 |
| notmyname | my hope is that the openstack object storage project would be able to manage at least some of the tests that are used to validate that something is in fact openstack object storage, and the mechanism to do that was the tempest plugin for our functests | 18:40 |
| *** catintheroof has quit IRC | 18:41 | |
| notmyname | so if tests in our repo loaded via a tempest plugin aren't allowed to become defocre tests, then I don't see any reason to add the tempest plugin | 18:41 |
| *** mmcardle has quit IRC | 18:44 | |
| torgomatic | this patch https://review.openstack.org/#/c/309524/ says to log the txid on INFO-level log messages. anyone know why it was skipped before? | 18:46 |
| patchbot | torgomatic: patch 309524 - swift - SwiftLogFormatter will log transaction IDs on INFO... | 18:46 |
| *** mkrcmari__ has quit IRC | 18:46 | |
| notmyname | torgomatic: INFO level stuff are access logs. demonstrably those have a txn id in them already. so probably it's already defined in the self._fmt and this check prevents double logging | 18:48 |
| notmyname | (just a guess) | 18:48 |
| torgomatic | notmyname: well, that patch doesn't seem to affect my access logs, so maybe that check was made useless at some point | 18:51 |
| dfg_ | notmyname: you have an email i can share this google doc spread sheet with? (its the hbird/swift #s) | 18:51 |
| notmyname | dfg_: me@not.mn | 18:52 |
| notmyname | dfg_: oh wait | 18:52 |
| torgomatic | certainly, with that patch, my access log lines still contain the txid once and only once | 18:52 |
| notmyname | dfg_: use john@swiftstack.com please. works better with google | 18:52 |
| notmyname | dfg_: thanks. got the invite | 18:52 |
| notmyname | dfg_: perfect, thanks | 18:54 |
| dfg_ | notmyname: cool | 18:54 |
| pdardeau | notmyname: ntata is about to propose patch for tempest plugin | 18:56 |
| notmyname | pdardeau: yeah, that's why I wanted to mention something :-) | 18:58 |
| timburke | torgomatic: notmyname: that seems to jive with clayg's assessment in the linked bug report | 18:59 |
| pdardeau | notmyname: the work is done, i believe she's just tidying up a readme.rst for it. | 18:59 |
| tdasilva | timburke: following up on your comments for copy middleware. At one point we had talked about adding something like a swift.slo_manifest_already_validated flag | 19:14 |
| tdasilva | timburke: but then we changed to use that 'format=raw' query on the GET | 19:15 |
| timburke | tdasilva: yeah...do we still want to do that? it seems like it's not a problem currently on master (because COPY is effectively right of SLO), but we could have un-copyable manifests following the change | 19:16 |
| timburke | the issue isn't so much the format of the manifest, but rather the data backing the manifest (and the constraints we put on that data) | 19:17 |
| notmyname | overview of dropbox's storage system https://blogs.dropbox.com/tech/2016/05/inside-the-magic-pocket/ | 19:17 |
| tdasilva | timburke: right, do you think 0-byte segments are likely or more of a corner case? | 19:18 |
| timburke | tdasilva: i know we have customers complaining about the fact that they can't do it anymore :/ | 19:19 |
| timburke | which led to https://review.openstack.org/#/c/301937/ | 19:19 |
| patchbot | timburke: patch 301937 - swift3 - Allow multipart uploads to have an empty final part (MERGED) | 19:19 |
| timburke | (so maybe that should have been "had" instead of "have") | 19:19 |
| timburke | i don't know of anyone using the swift API that's complained... | 19:20 |
| swiftfan_ | is Magic Pocket going to become part of Openstack? | 19:20 |
| clayg | lol? | 19:21 |
| tdasilva | timburke: so, what drove the decision to disallow 0-byte segments in slo? i missed that... | 19:24 |
| timburke | ask torgomatic; happened as part of https://review.openstack.org/#/c/252096/ | 19:25 |
| patchbot | timburke: patch 252096 - swift - Allow smaller segments in static large objects (MERGED) | 19:25 |
| clayg | tdasilva: we should allow *negative* byte segments in a slo | 19:25 |
| *** rcernin has joined #openstack-swift | 19:26 | |
| timburke | clayg: yes! don't start sending the next segment until the client sends *us* bytes! | 19:26 |
| clayg | see timburke get its ^ !!! | 19:26 |
| timburke | clayg: we have a strict "one in, one out" policy... | 19:27 |
| *** joeljwright has quit IRC | 19:28 | |
| *** mkrcmari__ has joined #openstack-swift | 19:31 | |
| openstackgerrit | David Goetz proposed openstack/swift: go: hashes.invalid https://review.openstack.org/286833 | 19:32 |
| *** dukhlov has quit IRC | 19:35 | |
| *** ukaynar has quit IRC | 19:38 | |
| openstackgerrit | Nandini Tata proposed openstack/swift: add Tempest Plugin for functional tests https://review.openstack.org/313676 | 19:51 |
| ntata | notmyname, Please take a look at this patch 313676 | 19:55 |
| patchbot | ntata: https://review.openstack.org/#/c/313676/ - swift - add Tempest Plugin for functional tests | 19:55 |
| *** pauloewerton has quit IRC | 19:59 | |
| *** nadeem has joined #openstack-swift | 20:04 | |
| *** dmk0202 has joined #openstack-swift | 20:05 | |
| *** mvk_ has joined #openstack-swift | 20:08 | |
| *** mkrcmari__ has quit IRC | 20:12 | |
| briancline | anyone seeing bug #1568650 manifest in production at scale? (or at least similar looking behavior, even when object-server isn't dying) | 20:12 |
| openstack | bug 1568650 in OpenStack Object Storage (swift) "Connection between client and proxy service does not closes" [High,Confirmed] https://launchpad.net/bugs/1568650 - Assigned to drax (devesh-gupta) | 20:12 |
| *** garthb_ has joined #openstack-swift | 20:27 | |
| *** garthb has quit IRC | 20:29 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift: Refactor server side copy as middleware https://review.openstack.org/156923 | 20:29 |
| *** BigWillie has quit IRC | 20:30 | |
| *** openstackgerrit has quit IRC | 20:32 | |
| *** openstackgerrit has joined #openstack-swift | 20:33 | |
| openstackgerrit | Merged openstack/swift: Reclaim isolated .meta files https://review.openstack.org/268251 | 20:50 |
| redbo | briancline: maybe. We see individual proxy workers start hoarding connections and memory sometimes, but we haven't narrowed down the cause yet. | 20:51 |
| *** dmk0202 has quit IRC | 20:58 | |
| *** ametts has quit IRC | 21:24 | |
| *** lakshmiS has quit IRC | 21:24 | |
| *** dmorita has quit IRC | 21:32 | |
| *** dmorita has joined #openstack-swift | 21:32 | |
| *** NM has quit IRC | 21:32 | |
| openstackgerrit | Nandini Tata proposed openstack/swift: add Tempest Plugin for functional tests https://review.openstack.org/313676 | 21:35 |
| *** dmorita has quit IRC | 21:45 | |
| *** _JZ_ has quit IRC | 21:48 | |
| *** ozialien10 has joined #openstack-swift | 21:51 | |
| *** dmorita has joined #openstack-swift | 21:54 | |
| openstackgerrit | Samuel Merritt proposed openstack/swift: Move all monkey patching to one function https://review.openstack.org/306612 | 21:55 |
| *** dmorita has quit IRC | 21:59 | |
| briancline | redbo: interesting. have seen those symptoms as well. in some cases it seems possible to trigger by doing an incomplete chunked write, but it feels like there are other causes as well | 22:01 |
| openstackgerrit | Merged openstack/swift: go: hashes.invalid https://review.openstack.org/286833 | 22:01 |
| briancline | I saw that torgomatic has an eventlet patch around that and someone else got in a swift patch to make that be reported a bit nicer, as it was a ValueError coming from inside eventlet | 22:02 |
| *** dmorita has joined #openstack-swift | 22:03 | |
| torgomatic | oh yeah, the bad chunk line... now it's a subclass of ValueError | 22:03 |
| briancline | yeah, ChunkReadError or something I think? | 22:04 |
| *** nadeem has quit IRC | 22:07 | |
| torgomatic | something like that | 22:07 |
| *** nadeem has joined #openstack-swift | 22:08 | |
| briancline | redbo: are you guys on 2.7 where you see those symptoms? | 22:10 |
| redbo | hahaha no | 22:11 |
| *** mingdang1 has joined #openstack-swift | 22:12 | |
| briancline | lol | 22:12 |
| torgomatic | oh man, we made a mistake at some point | 22:12 |
| briancline | we're a bit far behind as well, and have backported some of the monkeypatch fixes which i think helped mostly with the keystoneauth middleware | 22:12 |
| torgomatic | we should have gone from Swift 2.2 to 5.0 or something, because now I see 2.7 and I don't know if it's Swift or Python | 22:12 |
| redbo | We're more like 2.2 + a bunch of cherry picking and backporting. | 22:13 |
| briancline | aha.. we're on 2.2.2 with some cherry picks also | 22:14 |
| redbo | Well, 2.2.something, I don't know what anymore. It's a frankenstein monster. I'm working on how to deploy a clean 2.7 right now, but I don't know how we're going to do it. | 22:16 |
| notmyname | redbo: andrewd right? I heard a lot about that awesome daemon last week | 22:20 |
| redbo | I mean, it seems like it should be an ops problem, right? | 22:21 |
| briancline | would love to see andrewd be open-sourced | 22:26 |
| openstackgerrit | Merged openstack/swift: crypto - align cryptography version requirement with global https://review.openstack.org/313281 | 22:33 |
| *** david-lyle has quit IRC | 22:35 | |
| *** david-lyle has joined #openstack-swift | 22:37 | |
| *** nadeem has quit IRC | 22:42 | |
| *** lcurtis has quit IRC | 22:50 | |
| *** CaioBrentano has quit IRC | 23:04 | |
| *** daemontool has quit IRC | 23:12 | |
| *** sheel has quit IRC | 23:25 | |
| *** Jeffrey4l has joined #openstack-swift | 23:26 | |
| *** NM has joined #openstack-swift | 23:35 | |
| *** garthb_ has quit IRC | 23:38 | |
| *** NM has quit IRC | 23:40 | |
| *** arch-nemesis has quit IRC | 23:42 | |
| *** gyee has quit IRC | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!