| *** erlon has quit IRC | 00:01 | |
| openstackgerrit | Merged openstack/swift: Bulk upload: treat user xattrs as object metadata https://review.openstack.org/176161 | 00:03 |
|---|---|---|
| *** vinsh has quit IRC | 00:07 | |
| *** zhill has joined #openstack-swift | 00:07 | |
| *** ho has joined #openstack-swift | 00:09 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 00:12 |
| *** zhill has quit IRC | 00:12 | |
| *** evanjfraser has joined #openstack-swift | 00:21 | |
| *** asettle has quit IRC | 00:37 | |
| redbo | the first commit ever on swift was Thu Jul 2 10:37:16 2009 +0000 | 00:45 |
| redbo | wait that's a lie | 00:45 |
| notmyname | :-) | 00:45 |
| redbo | Sun Jun 28 08:42:44 2009 -0500 | 00:45 |
| notmyname | redbo: I went from the date in EXIF metadata on something from the party where we pushed conrad's big red button | 00:46 |
| redbo | sounds about right. https://www.facebook.com/photo.php?fbid=1455357628366&set=t.76002508&type=3&src=https%3A%2F%2Ffbcdn-sphotos-d-a.akamaihd.net%2Fhphotos-ak-xfa1%2Fv%2Ft1.0-9%2F28506_1455357628366_8364229_n.jpg%3Foh%3D25cf7202d58d869f237ed37563aab852%26oe%3D55C4D35F%26__gda__%3D1439084159_daf788b640b8f5675442c547bd237ce2&size=720%2C480 | 00:49 |
| notmyname | redbo: EXIF on http://d.not.mn/swift_launch_party.jpg | 00:52 |
| redbo | everyone looks so young | 00:52 |
| notmyname | no kidding !! | 00:52 |
| redbo | except will, he looks the same | 00:52 |
| notmyname | time for me to head home. have a good night | 00:56 |
| *** jlk has joined #openstack-swift | 01:27 | |
| *** jlk has left #openstack-swift | 01:28 | |
| *** tamizh_geek has joined #openstack-swift | 01:54 | |
| *** jlk has joined #openstack-swift | 01:59 | |
| *** WBHuber has joined #openstack-swift | 02:11 | |
| openstackgerrit | Janie Richling proposed openstack/swift: WIP - working on the encryption feature. https://review.openstack.org/157907 | 02:19 |
| *** panbalag has quit IRC | 02:21 | |
| *** jrichli has quit IRC | 02:42 | |
| *** early has quit IRC | 02:48 | |
| *** vinsh has joined #openstack-swift | 02:48 | |
| *** early has joined #openstack-swift | 03:01 | |
| *** WBHuber has quit IRC | 03:06 | |
| *** silor has joined #openstack-swift | 03:08 | |
| *** links has joined #openstack-swift | 03:31 | |
| *** tamizh_geek has quit IRC | 03:33 | |
| *** vinsh has quit IRC | 03:52 | |
| *** zaitcev has quit IRC | 03:53 | |
| *** vinsh has joined #openstack-swift | 04:18 | |
| *** vinsh has quit IRC | 04:25 | |
| *** bkopilov has joined #openstack-swift | 04:32 | |
| *** alejandrito has joined #openstack-swift | 04:41 | |
| *** alejandrito is now known as Guest58346 | 04:42 | |
| *** Guest58346 is now known as _alejandrito_ | 04:42 | |
| *** swat30 has quit IRC | 04:48 | |
| *** swat30 has joined #openstack-swift | 04:54 | |
| *** tamizh_geek has joined #openstack-swift | 05:25 | |
| *** _alejandrito_ has quit IRC | 05:37 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:41 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:41 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Imported Translations from Transifex https://review.openstack.org/179635 | 06:08 |
| *** openstackgerrit has quit IRC | 06:23 | |
| *** openstackgerrit has joined #openstack-swift | 06:23 | |
| *** ppai has joined #openstack-swift | 06:39 | |
| *** proteusguy has quit IRC | 06:41 | |
| *** proteusguy has joined #openstack-swift | 06:59 | |
| *** sandywalsh has quit IRC | 07:11 | |
| *** proteusguy has quit IRC | 07:12 | |
| *** sandywalsh has joined #openstack-swift | 07:12 | |
| *** thurloat has quit IRC | 07:13 | |
| *** thurloat has joined #openstack-swift | 07:13 | |
| *** vinsh has joined #openstack-swift | 07:25 | |
| *** vinsh has quit IRC | 07:29 | |
| *** thurloat has quit IRC | 07:32 | |
| *** thurloat has joined #openstack-swift | 07:33 | |
| *** yuan has joined #openstack-swift | 07:37 | |
| *** chlong has quit IRC | 07:44 | |
| *** geaaru has joined #openstack-swift | 07:45 | |
| mattoliverau | OK quiet swift land, I'm off to make dinner, night all | 07:48 |
| *** proteusguy has joined #openstack-swift | 07:55 | |
| *** tamizh_geek has quit IRC | 07:59 | |
| *** haomaiw__ has joined #openstack-swift | 08:08 | |
| *** SkyRocknRoll has quit IRC | 08:08 | |
| *** jordanP has joined #openstack-swift | 08:10 | |
| *** haomaiwang has quit IRC | 08:10 | |
| *** SkyRocknRoll has joined #openstack-swift | 08:21 | |
| *** slo_ has quit IRC | 08:49 | |
| *** jistr has joined #openstack-swift | 08:56 | |
| *** aix has quit IRC | 08:59 | |
| *** early has quit IRC | 09:12 | |
| openstackgerrit | Hisashi Osanai proposed openstack/swift: WIP: Enable Role-based access control using oslo.policy in Swift https://review.openstack.org/149930 | 09:17 |
| *** tamizh_geek has joined #openstack-swift | 09:18 | |
| *** ho has quit IRC | 09:19 | |
| *** aix has joined #openstack-swift | 09:26 | |
| *** joeljwright has joined #openstack-swift | 09:37 | |
| *** silor has quit IRC | 09:48 | |
| *** early has joined #openstack-swift | 09:51 | |
| *** bkopilov has quit IRC | 10:01 | |
| *** early has quit IRC | 10:03 | |
| *** panbalag has joined #openstack-swift | 10:04 | |
| *** acoles_away is now known as acoles | 10:11 | |
| *** aix has quit IRC | 10:12 | |
| *** early has joined #openstack-swift | 10:14 | |
| *** bkopilov has joined #openstack-swift | 10:17 | |
| *** tobe has joined #openstack-swift | 10:18 | |
| jordanP | hi. It seems the setup() method of the ObjectController is called twice per object-server worker. There's something I don't understand in this commit : https://github.com/openstack/swift/commit/a49fd3d8def23bfb49e605fea372c82d1ec6cdff#diff-391ba67a14720da1635b51be427fa134R290 It seems the "application" is instancied twice, once to call the app's modify_wsgi_pipeline() method and the second time to return the app to the caller. I | 10:22 |
| jordanP | 'd like someone to confirm or not that there's something weird. Thanks ! | 10:22 |
| *** aix has joined #openstack-swift | 10:25 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Cleanup and extend end to end ssync tests https://review.openstack.org/179379 | 10:33 |
| openstackgerrit | Emmanuel Cazenave proposed openstack/swift: X-Auth-Token should be a byestring. https://review.openstack.org/180098 | 10:35 |
| acoles | jordanP: thats correct, loadapp in wsgi.py loads the pipeline context, constructs the app and may give app opportunity to modify the pipeline context, and then the pipeline is created. | 10:40 |
| acoles | jordanP: a side effect is that an instance of the app is constrcuted twice | 10:40 |
| acoles | jordanP: its actually used in the proxy server https://github.com/openstack/swift/blob/master/swift/proxy/server.py#L578-578 | 10:40 |
| jordanP | acoles, right. Is there any way around it ? Like returning the application (the object created L288) after it has been given the opportunity to change its pipeline | 10:44 |
| *** ppai has quit IRC | 10:44 | |
| jordanP | I don't know if it makes sense, the thing is I have a diskFile implementaiton that has some side effect in its setup() method | 10:44 |
| jordanP | and the setup() being called twice is a bit problematic for me | 10:45 |
| *** silor has joined #openstack-swift | 10:45 | |
| *** early has quit IRC | 10:47 | |
| openstackgerrit | Merged openstack/swift: Imported Translations from Transifex https://review.openstack.org/179635 | 10:50 |
| *** early has joined #openstack-swift | 10:51 | |
| acoles | jordanP: hmmm, i just scanned the code and it all looks pretty knotted up with paste deploy, i didn't see a simple workaround | 10:55 |
| jordanP | acoles, okay thanks anyway ! I can try to work around it | 10:56 |
| acoles | jordanP: ok, sorry can't help more. if it proves hard to work around then maybe file a bug report/nice to fix. | 10:59 |
| jordanP | yep | 11:00 |
| *** lpabon has joined #openstack-swift | 11:03 | |
| *** early has quit IRC | 11:03 | |
| *** fanyaohong has joined #openstack-swift | 11:08 | |
| *** ppai has joined #openstack-swift | 11:08 | |
| *** sandywalsh has quit IRC | 11:12 | |
| *** fanyaohong has quit IRC | 11:13 | |
| *** fanyaohong has joined #openstack-swift | 11:13 | |
| *** sandywalsh has joined #openstack-swift | 11:14 | |
| *** early has joined #openstack-swift | 11:18 | |
| *** dencaval has joined #openstack-swift | 11:22 | |
| *** winggundamth has joined #openstack-swift | 11:26 | |
| *** thurloat is now known as thurloat_isgone | 11:26 | |
| *** NM1 has joined #openstack-swift | 11:26 | |
| winggundamth | hi. I follow http://docs.openstack.org/kilo/install-guide/install/apt/content/swift-install-controller-node.html with keystone but after I try doing swift list I got "Invalid user token - deferring reject downstream" error | 11:27 |
| *** nexusz99 has joined #openstack-swift | 11:28 | |
| *** nexusz99 has quit IRC | 11:28 | |
| *** NM has quit IRC | 11:28 | |
| winggundamth | I can do keystone user-list but not swift-list. This is full log http://paste.openstack.org/show/214954/ | 11:29 |
| *** ndk has quit IRC | 11:32 | |
| *** portante has quit IRC | 11:32 | |
| *** early has quit IRC | 11:36 | |
| *** early has joined #openstack-swift | 11:37 | |
| *** portante has joined #openstack-swift | 11:42 | |
| openstackgerrit | Prashanth Pai proposed openstack/swift: Make object creation more atomic in Linux https://review.openstack.org/162243 | 11:57 |
| *** thurloat_isgone is now known as thurloat | 12:10 | |
| *** tamizh_geek has quit IRC | 12:11 | |
| openstackgerrit | Emmanuel Cazenave proposed openstack/swift: X-Auth-Token should be a bytestring. https://review.openstack.org/180098 | 12:15 |
| *** SkyRocknRoll has quit IRC | 12:17 | |
| ppai | one unit test in master branch has been consistently failing in my fedora vm. Can anyone familiar with EC please have a look at it: http://paste.openstack.org/show/214947/ | 12:21 |
| *** rdaly2 has joined #openstack-swift | 12:23 | |
| *** bkopilov has quit IRC | 12:31 | |
| *** gyee has joined #openstack-swift | 12:39 | |
| *** early has quit IRC | 12:40 | |
| *** early has joined #openstack-swift | 12:45 | |
| *** mfedosin has quit IRC | 12:47 | |
| openstackgerrit | Prashanth Pai proposed openstack/swift: versioned writes middleware https://review.openstack.org/134347 | 12:51 |
| *** links has quit IRC | 12:52 | |
| *** cazino has joined #openstack-swift | 12:53 | |
| *** lpabon has quit IRC | 12:55 | |
| *** admin6 has joined #openstack-swift | 13:03 | |
| *** ppai has quit IRC | 13:08 | |
| *** winggundamth has quit IRC | 13:16 | |
| *** fanyaohong has quit IRC | 13:19 | |
| *** gyee has quit IRC | 13:22 | |
| *** WBHuber has joined #openstack-swift | 13:25 | |
| *** gyee has joined #openstack-swift | 13:27 | |
| *** madgriff has joined #openstack-swift | 13:28 | |
| *** ppai has joined #openstack-swift | 13:31 | |
| *** madgriff has quit IRC | 13:39 | |
| *** tobe has quit IRC | 13:40 | |
| *** Nadeem has joined #openstack-swift | 13:50 | |
| *** early has quit IRC | 13:57 | |
| *** early has joined #openstack-swift | 14:06 | |
| *** esker has joined #openstack-swift | 14:06 | |
| *** ppai has quit IRC | 14:11 | |
| *** jrichli has joined #openstack-swift | 14:13 | |
| *** annegentle has joined #openstack-swift | 14:29 | |
| *** Fin1te has joined #openstack-swift | 14:44 | |
| Fin1te | Hello everyone, I have a question about tuning Swift. Is there a good source of documentation for any kernel parameters and other such things? I suppose the operations guide is a good start, but looking for something beyond that. | 14:47 |
| NM1 | Fin1te: Have you taken a look at the Oreilly book about swift? | 14:49 |
| *** vinsh has joined #openstack-swift | 14:50 | |
| Fin1te | NM1: I haven't yet, but I certainly will. Thanks for the suggestion. Any websites in the mean time while I get my hands on a copy? | 14:50 |
| NM1 | Fin1te: I like to read the Swiftstack Blog. Is it pretty obvious? I hope not :) | 14:52 |
| *** Nadeem has quit IRC | 14:54 | |
| *** cdelatte has quit IRC | 14:54 | |
| *** delattec has quit IRC | 14:54 | |
| glange | Fin1te: stick around and you might get some answers once people wake up | 14:56 |
| glange | also, you might send an email to this list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators | 14:56 |
| glange | there are a lot devs in channel that don't necessaryily have a bunch of experience running a production cluster | 14:56 |
| Fin1te | NM1 I'll take a look through the blogs, thanks. glange, I'll most def stick around :) I'll give time for Ops folks to wake up and join in. I'll check into the mailing list too, thanks. | 14:57 |
| *** delattec has joined #openstack-swift | 15:02 | |
| *** cdelatte has joined #openstack-swift | 15:02 | |
| *** bsdkurt has quit IRC | 15:03 | |
| *** madgriff has joined #openstack-swift | 15:09 | |
| *** tamizh_geek has joined #openstack-swift | 15:12 | |
| *** gyee has quit IRC | 15:13 | |
| *** gyee has joined #openstack-swift | 15:14 | |
| *** tamizh_geek has quit IRC | 15:16 | |
| *** ndk has joined #openstack-swift | 15:18 | |
| *** mahatic has joined #openstack-swift | 15:21 | |
| *** macredcape has joined #openstack-swift | 15:28 | |
| *** mahatic has quit IRC | 15:33 | |
| *** bsdkurt has joined #openstack-swift | 15:44 | |
| *** admin6 has quit IRC | 15:49 | |
| *** mahatic has joined #openstack-swift | 15:51 | |
| *** cdelatte has quit IRC | 15:56 | |
| *** delattec has quit IRC | 15:56 | |
| *** cdelatte has joined #openstack-swift | 15:57 | |
| *** cdelatte has quit IRC | 15:58 | |
| *** cdelatte has joined #openstack-swift | 15:58 | |
| *** chlong has joined #openstack-swift | 16:00 | |
| notmyname | good morning | 16:04 |
| *** zhill has joined #openstack-swift | 16:05 | |
| dencaval | morning! | 16:06 |
| *** proteusguy has quit IRC | 16:12 | |
| *** chlong has quit IRC | 16:18 | |
| peluse | good morning | 16:21 |
| *** Fin1te has quit IRC | 16:21 | |
| *** Fin1te has joined #openstack-swift | 16:22 | |
| *** annegentle has quit IRC | 16:23 | |
| *** proteusguy has joined #openstack-swift | 16:25 | |
| *** annegentle has joined #openstack-swift | 16:25 | |
| *** jistr has quit IRC | 16:26 | |
| *** Fin1te has quit IRC | 16:26 | |
| *** annegent_ has joined #openstack-swift | 16:27 | |
| *** annegentle has quit IRC | 16:28 | |
| peluse | notmyname, see this? http://pulseblog.emc.com/2015/05/05/meet-project-coprhd-emc-delves-deeper-into-open-source/ | 16:29 |
| *** fthiagogv has joined #openstack-swift | 16:32 | |
| *** annegent_ has quit IRC | 16:33 | |
| *** chlong has joined #openstack-swift | 16:34 | |
| *** thurloat is now known as thurloat_isgone | 16:39 | |
| *** chlong has quit IRC | 16:39 | |
| notmyname | peluse: no. looking now | 16:43 |
| *** jordanP has quit IRC | 16:47 | |
| notmyname | that's interesting | 16:48 |
| notmyname | without this patch, is there an error? what's happening there? | 16:55 |
| notmyname | acoles: peluse: ^ | 16:55 |
| peluse | what patch? | 16:58 |
| notmyname | oh oops ;-) | 16:59 |
| notmyname | read my mind! | 17:00 |
| peluse | or I could have just said "yes, big trouble" :) | 17:00 |
| notmyname | https://review.openstack.org/#/c/178528/ | 17:00 |
| notmyname | peluse: mostly i'm curious about ppai's question. does it need to be backported? | 17:02 |
| notmyname | is this something that is causing errors? what's the pre-patch behavior? | 17:03 |
| peluse | Since it only affects .durable and I believe would be intermittent (?) chances are probably really good, expecially based on emperical testing, that we're fine. Must fix for production though I'd think | 17:03 |
| *** WBHuber has quit IRC | 17:03 | |
| notmyname | of course | 17:03 |
| peluse | chances of missing all the .durables I think would be really tiny | 17:03 |
| *** WBHuber has joined #openstack-swift | 17:03 | |
| peluse | but, hey, someone else might see it differently. Would still check w/ acoles and/or rest of the gang | 17:04 |
| notmyname | 'missing'? isn't his where the durable is written? so what other code path would write it? | 17:04 |
| -openstackstatus- NOTICE: zuul has been restarted to troubleshoot an issue, gerrit events between 15:00-17:00 utc were lost and changes updated or approved during that time will need to be rechecked or have their approval votes readded to trigger testing | 17:05 | |
| notmyname | or are you saying that generally the durable is going to be flushed even if the fsync fails | 17:05 |
| peluse | notmyname, I mean if the fsync didn't work and something else also went haywire and the .durable didn't get written | 17:05 |
| peluse | yes, that's waht I'm saying - unless there's another scdrewup also but honestly I don't know that for 100% sure | 17:05 |
| notmyname | right. so the kernel/FS will flush it at some point, even if the fsync fails | 17:05 |
| peluse | yup | 17:05 |
| peluse | unless something happens before the next flush for whatever reason (reset, crash, etc) | 17:06 |
| notmyname | but the whole point of the fsync is to protect the durable during those sorts of "other things" (power loss etc) | 17:06 |
| peluse | agree | 17:06 |
| notmyname | otherwise we wouldn't do an fsync at all | 17:06 |
| peluse | but we only need 2 .durables | 17:06 |
| notmyname | ah, ok, I see what you're saying | 17:06 |
| notmyname | ok, makes sense | 17:06 |
| peluse | so for a "beta" I'm thinking the chances of not getting at least 2 written are prettty slim | 17:06 |
| peluse | well, technically we only need 1 | 17:07 |
| acoles | peluse: notmyname i just added a comment on the patch | 17:08 |
| peluse | well 2 would need to make it to the sorage node but only one would ahve to survive a botched fsync + something else in order for it to be prop'd via the ecrecon | 17:08 |
| notmyname | it's really bad, but this is beta and definitely needs to be patched (and already has been). but IMO doesn't sound like we should backport. this isnt' the one and only one thing that would keep you from using EC in prod, and you shouldn't do that anyway with this beta code | 17:08 |
| notmyname | acoles: thanks | 17:08 |
| notmyname | got it | 17:09 |
| * peluse off to the gym... back later | 17:09 | |
| acoles | notmyname: peluse so actually we aren't fsyncing the dir after the durable is written which should also be fixed :? but its beta! | 17:10 |
| acoles | :/ i meant | 17:10 |
| peluse | ya, agree | 17:11 |
| acoles | i'm not sure we need backport though, its a beta. | 17:11 |
| *** gyee has quit IRC | 17:14 | |
| *** patrickeast has joined #openstack-swift | 17:16 | |
| *** rookie_ has joined #openstack-swift | 17:19 | |
| *** zhill has quit IRC | 17:19 | |
| rookie_ | Hi experts, I just pushed a new ring after omitting a storage node. Now that the new ring is in place and before the replication cycle completes, how does swift know where the objects being requested are since they are not yet in the new ring partition locations? | 17:21 |
| *** gyee has joined #openstack-swift | 17:21 | |
| clayg | rookie_: only one replica of a partition will move in a rebalance | 17:21 |
| rookie_ | clayg: ah - thanks! was really confused on it :) | 17:22 |
| clayg | rookie_: unless > 1 replica was on the server you popped because like you only had two servers or something | 17:22 |
| clayg | rookie_: yeah so if the "new" primary 404's it's normally pretty quick - and the proxy will just the next one down the line and down the line into hand offs until it finds it (or returns 404 because like... it really can't find it) | 17:23 |
| rookie_ | clayg: thanks - should be an issue (and wasn't in testing for pulling down some randomly selected object) as we have 8 storage nodes | 17:24 |
| *** early has quit IRC | 17:26 | |
| *** early has joined #openstack-swift | 17:26 | |
| *** AbyssOne has quit IRC | 17:31 | |
| *** bkopilov has joined #openstack-swift | 17:32 | |
| *** AbyssOne has joined #openstack-swift | 17:40 | |
| *** thurloat_isgone is now known as thurloat | 17:43 | |
| *** bkopilov has quit IRC | 17:45 | |
| *** bkopilov has joined #openstack-swift | 17:45 | |
| clduser_ | hello, can someone help me configure a ceilometer with keystone v3 for swift storage metering? I'm using Kilo on Ubuntu 14.04 and I'm trying to use the storage pollster but it doesn't work because of the keystoneclient v2_0 import (for requesting tenant list). I know it is not exactly a swift realted issue but could not find solution in related topics/rooms. | 18:04 |
| *** patrickeast has quit IRC | 18:11 | |
| *** barra204 has joined #openstack-swift | 18:12 | |
| *** bkopilov has quit IRC | 18:14 | |
| *** bkopilov has joined #openstack-swift | 18:15 | |
| *** zhill has joined #openstack-swift | 18:20 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: WIP Add POST capability to ssync for .meta files https://review.openstack.org/138498 | 18:20 |
| *** bkopilov has quit IRC | 18:21 | |
| *** acoles is now known as acoles_away | 18:23 | |
| *** zhill has quit IRC | 18:25 | |
| *** Fin1te has joined #openstack-swift | 18:25 | |
| *** barra204 has quit IRC | 18:27 | |
| *** logan2 has quit IRC | 18:28 | |
| *** logan2 has joined #openstack-swift | 18:31 | |
| *** thurloat is now known as thurloat_isgone | 18:39 | |
| *** j_king has quit IRC | 18:40 | |
| *** j_king has joined #openstack-swift | 18:40 | |
| *** barra204 has joined #openstack-swift | 18:41 | |
| *** geaaru has quit IRC | 18:43 | |
| *** thurloat_isgone is now known as thurloat | 18:46 | |
| *** fthiagogv has quit IRC | 18:46 | |
| *** zaitcev has joined #openstack-swift | 18:50 | |
| *** ChanServ sets mode: +v zaitcev | 18:50 | |
| *** annegentle has joined #openstack-swift | 18:59 | |
| *** silor1 has joined #openstack-swift | 19:00 | |
| *** tacotuesday has joined #openstack-swift | 19:01 | |
| *** silor has quit IRC | 19:02 | |
| *** rdaly2 has quit IRC | 19:03 | |
| *** madgriff has quit IRC | 19:04 | |
| *** jistr has joined #openstack-swift | 19:08 | |
| *** gyee has quit IRC | 19:13 | |
| *** EmilienM is now known as EmilienM|afk | 19:15 | |
| *** gyee has joined #openstack-swift | 19:15 | |
| clayg | clduser_: oh man, you probably lost most of us at ceilometer and the rest of us at keystone v3 :\ | 19:18 |
| openstackgerrit | Matt Riedemann proposed openstack/swift: Revert "Bulk upload: treat user xattrs as object metadata" https://review.openstack.org/180276 | 19:21 |
| *** zhill has joined #openstack-swift | 19:21 | |
| *** zhill has quit IRC | 19:25 | |
| *** annegentle has quit IRC | 19:28 | |
| *** minwoob has quit IRC | 19:48 | |
| *** jistr has quit IRC | 19:49 | |
| *** thurloat is now known as thurloat_isgone | 20:00 | |
| *** silor1 has quit IRC | 20:01 | |
| *** annegentle has joined #openstack-swift | 20:05 | |
| *** breitz has quit IRC | 20:16 | |
| *** breitz has joined #openstack-swift | 20:16 | |
| *** bkopilov has joined #openstack-swift | 20:21 | |
| *** zhill has joined #openstack-swift | 20:21 | |
| *** jkremer has joined #openstack-swift | 20:24 | |
| WBHuber | How do I make it possible for fragment payload to time out in line 308 in reconstructor.py other than setting node_timeout to zero? | 20:25 |
| *** zhill has quit IRC | 20:26 | |
| *** bkopilov has quit IRC | 20:26 | |
| *** tab__ has joined #openstack-swift | 20:27 | |
| *** gyee has quit IRC | 20:28 | |
| *** bsdkurt1 has joined #openstack-swift | 20:29 | |
| *** bsdkurt has quit IRC | 20:30 | |
| *** barra204 has quit IRC | 20:37 | |
| *** aix has quit IRC | 20:38 | |
| *** minwoob has joined #openstack-swift | 20:38 | |
| *** Trozz has quit IRC | 20:39 | |
| *** Trozz has joined #openstack-swift | 20:39 | |
| *** tamizh_geek has joined #openstack-swift | 20:40 | |
| *** bill_az has joined #openstack-swift | 20:52 | |
| bill_az | swifterdarrel: (and others) has anyone looked at using ssbench with keystone v3? | 20:52 |
| bill_az | if not, I will - but wanted to see if anything had been done here already | 20:53 |
| notmyname | bill_az: no, no one has looked at that | 20:53 |
| bill_az | notmyname: ok - thanks | 20:54 |
| *** gyee has joined #openstack-swift | 20:54 | |
| openstackgerrit | Michael Barton proposed openstack/swift: Don't quarantine on missing data files https://review.openstack.org/180317 | 20:56 |
| *** rookie_ has quit IRC | 20:57 | |
| *** jp has joined #openstack-swift | 20:57 | |
| *** Fin1te has quit IRC | 21:00 | |
| *** tacotuesday has quit IRC | 21:01 | |
| tab__ | notmyname: yesterday we had some conversation regarding changing bind_ip within account/object/container-server.conf files - i commented back bind_ip addresse and get wierd reponses. Container was created at the end , but i could not upload object.... | 21:02 |
| jp | Am I correct that ssync only works with object replication? It seems that account and container replication still relies on db_replicator which uses rsync. | 21:03 |
| tab__ | notmyname: if you are there to check i have now logs : http://pastebin.com/YeNFumMc | 21:04 |
| jp | or is the ssync_sender code actually working with partitions as a whole and not just objects? | 21:04 |
| jp | and I see that if I delete an account or container while I don't have rsync installed, the account-replicator and container-replicator logs complain when doing the rsync process call about "file not found" | 21:07 |
| jp | but even if I use sync_method=ssync | 21:08 |
| jp | I must be missing something here because either I don't have ssync configured correctly (supposed to be just a one line config change in object-server.conf right?) or I'm misinterpreting how the code works | 21:09 |
| jp | any guidance is greatly appreciated | 21:09 |
| tab__ | notmyname: also these are my builder files: http://pastebin.com/KSGLtSFM | 21:10 |
| *** annegentle has quit IRC | 21:11 | |
| *** annegentle has joined #openstack-swift | 21:12 | |
| jp | does anyone even know the best person to ask? I see the original ssync code was created by gholt | 21:12 |
| notmyname | jp: ssync is only for objects | 21:19 |
| *** zhill has joined #openstack-swift | 21:22 | |
| *** zhill has quit IRC | 21:27 | |
| zaitcev | tab__: I'm sorry if I presume too much, but if you aren't sure how to make ports in the config matcihng the ring, I would not recommend splitting service and replication ports. In fact, never do it, period. You may need to split IPs, but leave ports alone. | 21:27 |
| *** gyee has quit IRC | 21:27 | |
| *** tacotuesday has joined #openstack-swift | 21:28 | |
| jp | wow thanks notmyname...that is what I thought from experiencing the setup and looking at the code. I just wanted to get some folks more versed ssync to confirm. | 21:29 |
| zaitcev | jp: the outcome is, you must still use rsync, so don't try to run without it (unless you have nodes that only have object servers and nothing else). | 21:29 |
| jp | thanks again...will do | 21:30 |
| tab__ | zaitcev: thx for reponse. This is my ring-build script: http://pastebin.com/Qhq80z7R . I wanted to seperate storage and replication network (following some guides). If you are kind to check on script, i do not know what am i doing wrong? This is a setup on VM (2 storage nodes, each with 4 VDI disks) | 21:35 |
| *** zhill has joined #openstack-swift | 21:36 | |
| *** jrichli has quit IRC | 21:37 | |
| *** NM1 has quit IRC | 21:37 | |
| zaitcev | tab__: The script looks fine, and the resulting ring looks fine, but I'm trying to point out that ring must match the configs. Surely if you specify r1z1-172.100.10.51:6000R172.100.9.51:6003/d1, you must run 2 object servers, one listening on 6000 and another listening on 6003, right? Are you doing that? | 21:38 |
| zaitcev | tab__: Best just not do that, IMHO. Otherwise you end hurting yourself. | 21:39 |
| zaitcev | change them all to something like r1z1-172.100.10.51:6200R172.100.9.51:6200/d1 and make sure you have just 1 good config on that node, which has 1) bind_ip commented out or set to 0.0.0.0 (because you have 172.100.10.51 and 172.100.9.51), 2) bind_port=6200. | 21:42 |
| tab__ | zaitcev: thx for comment. I will recheck my configuration | 21:42 |
| zaitcev | tab__: I presume it's not a production cluster, because I cannot imagine why anyone would want a separate replication network with just 2 nodes... | 21:43 |
| tab__ | it's not a production, but i am doing some proove-of-concept.... still learning | 21:44 |
| zaitcev | okay | 21:44 |
| peluse | WBHuber, just saw your question but about to hit the road, will be back on line a bit later this eve and will get back to you if nobody else does before me | 21:46 |
| zaitcev | In that case splitting, say, 172.100.10.51:6200 and 172.100.9.51:6200 sounds fine. Split IP, not ports. I try to avoid the split because then one must be sure that each node has 2 IPs assigned correctly and never has them swapped by accident... Well, using static DHCP helps with that. Still, seems like the trouble you don't want. | 21:46 |
| tab__ | so in your case of recommendation, if bind_ip is commented , it is ok. | 21:46 |
| WBHuber | peluse: sounds good and thanks. | 21:46 |
| tab__ | and i will also do reconfiguration of ports as you adviced | 21:47 |
| zaitcev | tab__: I usually check with netstat -atn | grep 6000 all the same. Well, in your case you also want 6003. Which I hope you discontinue. | 21:47 |
| zaitcev | Reading if #bing_ip is commented tells you want you mean the services listen on, but netstat tells you what services actually listen on. Big difference :-) | 21:49 |
| tab__ | zaitcev: ok thx for now! i'll talk back to you if i have any more questions, since i see you are experienced on this subject :) | 21:53 |
| notmyname | clayg: torgomatic: http://www.meetup.com/sfpython/events/222323217/ | 21:54 |
| *** annegent_ has joined #openstack-swift | 21:55 | |
| *** asettle has joined #openstack-swift | 21:58 | |
| *** esker has quit IRC | 21:58 | |
| *** annegentle has quit IRC | 22:00 | |
| *** zhill_ has joined #openstack-swift | 22:09 | |
| mattoliverau | Morning | 22:14 |
| *** WBHuber has quit IRC | 22:16 | |
| tab__ | zaitcev: still there? so this would probably also good guide for seperate replication network to learn from: http://docs.openstack.org/developer/swift/replication_network.html#for-a-multiple-server-replication ? Or are there also some other good tutorials online that you possible know for? | 22:19 |
| zaitcev | tab__: It looks like an allright doc, but please observe that it still uses the same port (the default, whatever it is). | 22:20 |
| zaitcev | tab__: Still... It's not how you're supposed to do it. | 22:21 |
| tab__ | zaitcev: ok thank, i will take that into a count. Good bye for now, i will further study in morning,it is late here in europe :) | 22:22 |
| zaitcev | tab__: It is generally expected that you start with a normal cluster, and then, when your switches all melt down and everyone runs around like a headless chicken, someone says, well shucks we're screwed! might as well try a separate replication network! | 22:22 |
| tab__ | aha ok :) | 22:23 |
| tab__ | thx | 22:23 |
| zaitcev | Swift is full of escape hatches like that. You aren't trying to use the overload variable yet, aren't you | 22:23 |
| *** bkopilov has joined #openstack-swift | 22:25 | |
| *** mahatic_ has joined #openstack-swift | 22:29 | |
| *** mahatic has quit IRC | 22:30 | |
| *** tab__ has quit IRC | 22:33 | |
| *** EmilienM|afk is now known as EmilienM | 22:52 | |
| *** wbhuber has joined #openstack-swift | 22:57 | |
| *** tdasilva has quit IRC | 22:58 | |
| *** annegent_ has quit IRC | 23:02 | |
| *** madgriff_ has joined #openstack-swift | 23:04 | |
| *** wbhuber has quit IRC | 23:05 | |
| *** tamizh_geek has quit IRC | 23:05 | |
| *** tacotuesday has quit IRC | 23:07 | |
| *** NM has joined #openstack-swift | 23:08 | |
| *** madgriff_ has quit IRC | 23:10 | |
| *** tdasilva has joined #openstack-swift | 23:18 | |
| notmyname | looks like I forgot to schedule a session | 23:19 |
| notmyname | ie left one slot unscheduled | 23:19 |
| notmyname | which one goes in that slot? | 23:20 |
| *** vinsh has quit IRC | 23:23 | |
| *** harlowja has joined #openstack-swift | 23:24 | |
| harlowja | have u guys seen the following | 23:24 |
| harlowja | http://logs.openstack.org/51/176951/45/check/gate-tempest-dsvm-neutron-src-taskflow/9c17ff2/logs/screen-s-object.txt.gz | 23:24 |
| harlowja | " No such file or directory: '/usr/local/lib/python2.7/dist-packages/xattr/__pycache__/_cffi__x282d9483x6ee6f75.c'" ? | 23:24 |
| notmyname | harlowja: I think clayg has been talking about that for a few hours in #openstack-dev | 23:30 |
| harlowja | ah | 23:30 |
| harlowja | cools beans | 23:30 |
| harlowja | i opened https://bugs.launchpad.net/swift/+bug/1452068 anyway | 23:30 |
| openstack | Launchpad bug 1452068 in OpenStack Object Storage (swift) "Swift object server cffi xattr failure" [Undecided,New] | 23:30 |
| harlowja | close at will (if its a dup/other) | 23:30 |
| notmyname | it's not a bug in swift aiui | 23:30 |
| *** jp has quit IRC | 23:32 | |
| *** jkremer has quit IRC | 23:33 | |
| *** mahatic_ has quit IRC | 23:34 | |
| notmyname | torgomatic: mattoliverau: clayg: redbo: kota_: ho: acoles_away: there's nothing on the agenda for tomorrow's meeting. I propose we skip it. do you have anything to bring up during it? | 23:35 |
| *** chlong has joined #openstack-swift | 23:37 | |
| *** tdasilva has quit IRC | 23:40 | |
| *** annegentle has joined #openstack-swift | 23:47 | |
| *** tdasilva has joined #openstack-swift | 23:47 | |
| *** annegentle has quit IRC | 23:47 | |
| mattoliverau | notmyname: Nope, I'm fine for skipping, i suspect we're all in summit prep mode :) Speaking of which I'm hoping to have a new rewritten sharding SPEC up by my Friday at the latest. | 23:52 |
| notmyname | nice | 23:53 |
| notmyname | now I gotta see if kota and ho can know about no meeting before they get up at 4am local time | 23:54 |
| notmyname | ok, no meeting tomorrow | 23:54 |
| notmyname | hmm.. the only this is to figure out the last summit working session | 23:55 |
| notmyname | I'm not sure which one to use | 23:56 |
| mattoliverau | notmyname: want to mark down where the scheduling is at on the etherpad so we know what's left and help make the decision :) | 23:58 |
| *** jrichli has joined #openstack-swift | 23:58 | |
| notmyname | do you see the "TBD session here? | 23:59 |
| notmyname | that's where it will go | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
