Thursday, 2015-12-17

*** pgbridge has quit IRC00:02
*** ChubYann has quit IRC00:09
*** lpabon has joined #openstack-swift00:16
*** Jeffrey4l has joined #openstack-swift00:20
*** ChubYann has joined #openstack-swift00:22
openstackgerritMatthew Oliver proposed openstack/swift: Pass HTTP_REFERER down to subrequests
mattoliverau^ this version adds a SLO functional test as well.00:24
kota_mattoliverau: \o/00:25
*** kei_yama has quit IRC00:32
*** kei_yama has joined #openstack-swift00:33
*** lpabon has quit IRC00:38
*** Jeffrey4l has quit IRC00:46
*** zhill has quit IRC00:49
*** rcernin has quit IRC00:50
openstackgerritTim Burke proposed openstack/swift: Expose token expiration time in tempauth auth response
*** diogogmt has quit IRC00:56
*** badari_ has joined #openstack-swift01:05
*** chlong has joined #openstack-swift01:06
*** badari has quit IRC01:08
*** gyee has quit IRC01:11
*** zhill has joined #openstack-swift01:14
*** dmorita has quit IRC01:15
openstackgerritTimur Alperovich proposed openstack/swift: Fix IPv6 handling in MemcacheConnPool.
*** zhiyan has quit IRC01:25
*** zhiyan has joined #openstack-swift01:28
*** zhill has quit IRC01:30
*** dmorita has joined #openstack-swift01:39
*** kei_yama has quit IRC01:47
*** kei_yama has joined #openstack-swift01:48
*** lyrrad has quit IRC01:57
*** chlong has quit IRC02:10
*** klrmn has quit IRC02:13
*** daddyjoseph97 has quit IRC02:19
*** chlong has joined #openstack-swift02:22
*** Jeffrey4l has joined #openstack-swift02:28
*** zhill has joined #openstack-swift02:28
*** zhill has quit IRC02:33
openstackgerritHisashi Osanai proposed openstack/swift: Add functional test for access control (RBAC) with Keystone
*** diogogmt has joined #openstack-swift02:45
*** klrmn has joined #openstack-swift03:06
*** Jeffrey4l has quit IRC03:08
*** jamielennox is now known as jamielennox|away03:26
*** links has joined #openstack-swift03:28
*** then3rd has quit IRC03:30
*** then3rd has joined #openstack-swift03:31
*** jamielennox|away is now known as jamielennox03:31
*** klrmn has quit IRC03:36
then3rdAnyone around? Was hoping to get a quick confirmation ahead of time. I'm upgrading swift proxy from icehouse to juno (but not keystone yet). Do I need to worry changing the proxy config to use identity_uri (vs auth_host) or auth_uri v2.0?03:38
then3rdI found this, but I'm not sure if I understand.
*** venkat has joined #openstack-swift03:40
jamielennoxthen3rd: that option goes to auth_token middleware so it depends what version of that you are running, the identity_uri is ust the auth_protocol auth_host auth_port etc put together rather than have them be individual options03:45
*** nadeem has joined #openstack-swift03:45
jamielennoxif you continue to use auth_host etc then you'll get a deprecation warning, but even in master those options are still supported03:45
*** nadeem has quit IRC03:46
jamielennoxso it won't screw anything up if you don't change yet03:46
*** klrmn has joined #openstack-swift03:46
*** logan- has joined #openstack-swift03:47
then3rdCool, exactly the answer I was hoping to hear. I found this, and it doesn't have any deprecations.
then3rdThanks for the quick reply!03:49
*** mac_ified has joined #openstack-swift03:53
jamielennoxthen3rd: it's not really something swift controls, but comes from the keystonemiddleware so it may or may not be documented there03:53
jamielennoxbut yea, it's not going to be a required change yet03:54
then3rdI may as well while I'm at it. Goal is to get my stack up to date.03:57
openstackgerritHisashi Osanai proposed openstack/swift: Add functional test for access control (container ACL) with Keystone
openstackgerritHisashi Osanai proposed openstack/swift: Add functional test for access control (RBAC INFO) with Keystone
then3rdFun times.. I'm doing my first rolling upgrade without the experience of building it from scratch. o_O04:01
openstackgerritHisashi Osanai proposed openstack/swift: Add functional test for access control (COPY) with Keystone
openstackgerritHisashi Osanai proposed openstack/swift: Performance improvement for RBAC related functional tests
howow! it's easy to rebase with new gerrit :)04:01
kota_ho: some bottuns are there?04:10
kota_ho: did you try them?04:10
hokota_: yeah, there is "rebase" bottun04:11
kota_ho: rebase is one of institutive bottuns but...04:11
openstackgerritjanonymous proposed openstack/swift: [WIP] testr to run unittests
kota_i don't expect what happen with Follow-Up bottun.04:12
hokota_: hahaha. that's true.04:12
notmynamegood evening04:19
*** proteusguy_ has quit IRC04:21
kota_notmyname: good evening :)04:21
honotmyname: good evening!04:22
notmynamewhat's going on?04:22
*** mahatic has joined #openstack-swift04:22
kota_conversation for new gerrit api04:23
notmynameI was stuck in meetings all day today, so I haven't had a chance to try it out yet04:23
*** zhill has joined #openstack-swift04:23
notmynamehow is it?04:23
notmynamemahatic: hi! I was just thinking that I hadn't seen you around in a few days04:23
*** zhill has quit IRC04:23
kota_notmyname: and please add your +2 +A for my smallest patch 258233 :-)04:24
patchbotkota_: - Fix reconciler test to calc lastmodified as UTC04:24
mahaticgood morning04:24
kota_ho likes the gerrit change to make it easy to rebase off current master04:25
mahaticnotmyname: hi! I was on vacation for the past 3 days (left a message on IRC). Back now, lots to catch up04:25
kota_i think something better but something worse...04:25
notmynamemahatic: I hope you had a good vacation04:25
notmynameho: kota_: what's the follow-up button do?04:25
kota_notmyname: idk,04:26
kota_right now.04:26
kota_mahatic: welcome back :-)04:26
kota_let's try the follow-up bottun with my patch (i don't like to hurt anyone...)04:27
mahaticnotmyname: had a great one. I'm now 'officially' engaged! :) boyfriend turned fiancee04:28
mahatickota_: thanks!04:28
kota_...after clicking follow-up bottun, new text box opened...04:28
notmynamemahatic: wow! congrats!04:28
mahaticnotmyname: thank you!04:28
kota_mahatic: so congrats!04:29
mattoliveraumahatic: congrats!!!04:29
notmynamekota_: how do I review? what's the button to click?04:29
homahatic: Congrats!04:29
mattoliveraunotmyname: reply04:29
notmynameah ok04:29
mattoliverauat the top, it makes alot of sense :P04:29
pelusemahatic: heck yeah!  great news for sure!!!04:29
mattoliverauscroll to the bottom to read, then back to the top to reply04:29
*** suresh_ has joined #openstack-swift04:29
mahatickota_: mattoliverau: ho: peluse thank you!04:29
mattoliveraupeluse: hey man o/04:30
pelusemattoliverau: yo!04:30
kota_got it about follow-up bottun.04:30
kota_it will create a new commit depends on the patch in gerrit.04:31
suresh_I am trying to add meta-quota bytes to my  demo swift account but i am getting  error like Access was denied to this resource.04:31
suresh_please someone help04:31
kota_and if write down some text to the box, it will be pasted into commit message in the new commit.04:32
kota_but i think we need pull the empty commit to add some follow-up...maybe?04:32
suresh_I am trying to add meta-quota bytes to my  demo swift account but i am getting  error like Access was denied to this resource. please help me04:32
mattoliveraukota_: yeah I guess, then push it back up04:33
kota_mattoliverau: is that useful? idk04:33
*** proteusguy_ has joined #openstack-swift04:33
mattoliveraukota_: no idea, might help when making chains or follow ups.. but nah I don't think it really is04:33
kota_off a bit for another meeting..04:34
mattoliverausuresh_: hi, what are you trying to do, I don't understand? you trying to set a quota or setting some meta data?04:34
mattoliverausuresh_: ok I assume your trying to add/set account quota (bytes)04:36
notmynamemattoliverau: kota_: yup, I agree. not useful04:36
mattoliveraujanonymous: morning04:36
janonymousmahatic: congrats !!04:36
mahaticjanonymous: morning, and thank you!04:37
mattoliveraunotmyname: there is a built in editor, so could make sense, but I haven't used it outside of updating a commit message04:37
suresh_mattoliverau: yes i am trying to addd account quota bytes04:37
then3rdnotmyname: good evening!04:37
mattoliverausuresh_: and your getting access denied, how are you updating using swift client or curl?04:38
suresh_mattoliverau:using swift client04:38
mattoliverausuresh_: are you authenticated, ie can you: swift stat?04:39
suresh_mattolivera: yeah i can04:41
*** tsg has joined #openstack-swift04:41
janonymousthere is no option to see depends or needed by in gerrit ?04:42
suresh_mattolivera: i am getting output for swift stat04:43
mattoliverausuresh_: sorry, just testing locally first, I haven't used quotas much04:44
*** ppai has joined #openstack-swift04:45
suresh_mattolivera: kk04:45
suresh_ I am trying to add meta-quota bytes to my  demo swift account but i am getting  error like Access was denied to this resource04:45
mattoliverausuresh_: are you using a reseller account, cause writing quota metadata is limited to reseller admins04:45
suresh_please someone help04:46
*** badari has joined #openstack-swift04:47
*** badari_ has quit IRC04:50
mattoliverausuresh_: a normal user can't set account quotas, only reseller admins04:50
mattoliverausuresh_: is the account your using a reseller admin, I assume its now, because you'll be getting the error you speak if your not.04:50
suresh_mattoliverau: i am using demo user account04:54
mattoliverausuresh_: So only reseller admins can set account quota's. An admin on an account can set Container quotas however. So you might be able to do the latter.04:54
suresh_mattoliverau: how to give permissions for demo as reseller admin04:55
mattoliverausuresh_: I don't know what your using, devstack? Saio?04:55
suresh_i am using openstack04:56
mattoliverausuresh_: ok, but whose?04:56
suresh_mattoliverau: i am using openstack saio04:56
*** zhill has joined #openstack-swift04:56
mattoliverausuresh_: you are using openstack swift all in one?04:56
*** takashi has joined #openstack-swift04:57
suresh_no two node04:58
suresh_mattoliverau: no two node04:58
*** eranrom has quit IRC04:58
mattoliverausuresh_: I don't know what your using, but if its using keystone authentication then your user needs to be added to a reseller role defined in your swift/proxy-server.conf04:58
mahatichalf the scrollback is on tetstr patch, wasnt so difficult ;)05:02
notmynamemahatic: and now it's done, so you don't have to worry about it05:04
suresh_mattoliverau: yeah how to add that for my user05:04
mahaticnotmyname: ah! nice!05:04
mattoliveraumahatic: lol05:04
openstackgerritpei ran wu proposed openstack/python-swiftclient: :Add validation of -S option for SLO upload
mattoliverausuresh_: so you are using keystone?05:05
suresh_mattoliverau: yes i am using keystone05:05
mattoliverausuresh_: then you need to check what the reseller role has been mapped to (look in swift's proxy server config). then goto keystone and add that role to your user.05:06
*** mtreinish has quit IRC05:08
*** garthb has joined #openstack-swift05:12
suresh_mattoliverau: how to add05:12
suresh_mattoliverau : i did'nt have reseller role05:12
mattoliverausuresh_: if a reseller_admin_role hasn't been defined in the proxy config, then swift will default to a role named ResellerAdmin05:15
mattoliverausuresh_: so either use reseller_admin_role and name a role in keystone you've created and want to use, or just go create a role in keystone called ResellerAdmin and give it to your user.05:16
*** mtreinish has joined #openstack-swift05:17
mattoliverausuresh_: NOTE: that a reseller_admin has access to _all_ accounts. And thus a super administrator. One of these are required to set account quotas, otherwise people could turn it off on there own accounts.05:18
*** SkyRocknRoll has joined #openstack-swift05:20
*** diogogmt has quit IRC05:21
*** tsg has quit IRC05:27
suresh_mattoliverau: tnq05:28
openstackgerritMerged openstack/swift: Fix reconciler test to calc lastmodified as UTC
*** diogogmt has joined #openstack-swift05:33
*** yatin has joined #openstack-swift05:35
*** takashi has quit IRC05:36
*** badari has quit IRC05:44
*** takashi has joined #openstack-swift05:45
kota_notmyname: thx for reviewing patch 258233 that landed successfully :)05:47
patchbotkota_: - Fix reconciler test to calc lastmodified as UTC (MERGED)05:47
notmynameyou're welcoem05:47
*** suresh_ has quit IRC05:48
*** SkyRocknRoll has quit IRC05:49
then3rdWell, first storage node upgraded with no problems. :)05:50
kota_then3rd: congrats :)05:51
mahaticacoles_: hi, re:content-length for ranged gets  in EC, I don't think it needs to recalculate the content-length. It does so once with the request range (in _prepare_for_ranges in Encryption can use that too, except we need the decrypted content-type length and hence number of parts.05:51
*** SkyRocknRoll has joined #openstack-swift06:03
openstackgerritjanonymous proposed openstack/swift: Eventlet green threads not released back to pool
then3rdHaha.. I couldn't figure out why my proxy logs were't updating. Forgot I setup syslog to logstash.06:22
*** dmorita has quit IRC06:23
*** dmorita has joined #openstack-swift06:24
*** dmorita has quit IRC06:27
*** dmorita has joined #openstack-swift06:27
*** haomaiwang has joined #openstack-swift06:27
*** klrmn has quit IRC06:33
*** klrmn has joined #openstack-swift06:34
*** takashi has quit IRC06:34
*** haomaiwang has quit IRC06:37
*** 5EXAAIFVW has joined #openstack-swift06:38
*** dmorita has quit IRC06:40
*** Jeffrey4l has joined #openstack-swift06:40
*** dmorita has joined #openstack-swift06:43
*** trifon has joined #openstack-swift06:43
then3rdAnsible will make this a lot more fun next time. cssh only helps destroy things faster. ;)06:44
*** dmorita has quit IRC06:47
*** takashi has joined #openstack-swift06:54
*** 5EXAAIFVW has quit IRC07:01
*** haomaiwang has joined #openstack-swift07:01
*** eranrom has joined #openstack-swift07:02
*** nakagawamsa has joined #openstack-swift07:17
*** haomaiwang has quit IRC07:26
*** haomaiwang has joined #openstack-swift07:29
*** garthb has quit IRC07:31
*** chlong has quit IRC07:36
*** zhill has quit IRC07:40
*** eranrom has quit IRC07:54
*** arnox has joined #openstack-swift07:58
*** haomaiwang has quit IRC08:01
*** haomaiwang has joined #openstack-swift08:01
*** takashi has quit IRC08:12
*** wer has quit IRC08:15
*** silor has joined #openstack-swift08:19
*** geaaru has joined #openstack-swift08:24
*** wer has joined #openstack-swift08:26
*** rcernin has joined #openstack-swift08:26
*** klrmn has quit IRC08:26
*** venkat has quit IRC08:32
*** ianbrown has quit IRC08:36
*** ianbrown has joined #openstack-swift08:37
*** takashi has joined #openstack-swift08:40
*** rcernin has quit IRC08:40
*** rcernin has joined #openstack-swift08:41
*** venkat has joined #openstack-swift08:46
*** mac_ified has quit IRC08:55
then3rdWooh. 2.2.0. Next step liberty. Then I'll do keystone.
*** haomaiwang has quit IRC09:01
*** haomaiwang has joined #openstack-swift09:01
*** briancurtin has quit IRC09:04
*** jmccarthy has joined #openstack-swift09:06
*** jistr has joined #openstack-swift09:09
hoi couldn't reproduce the problem of patch 229532. i will continue to work with it...09:10
patchbotho: - swift-init: SIGKILL daemon after kill_wait period.09:10
*** briancurtin has joined #openstack-swift09:12
*** ho has quit IRC09:13
*** briancurtin has quit IRC09:17
*** takashi has quit IRC09:18
*** zigo has quit IRC09:25
*** zigo_ has joined #openstack-swift09:26
*** Jeffrey4l has quit IRC09:34
openstackgerritjanonymous proposed openstack/swift: Eventlet green threads not released back to pool
peterlisakho, hi, what problem do you mean? reload behaves like restart? or just can't make process to hang up?09:36
*** eikke has joined #openstack-swift09:38
*** venkat has quit IRC09:39
*** haomaiwang has quit IRC09:40
*** takashi has joined #openstack-swift09:40
*** briancurtin has joined #openstack-swift09:44
*** openstackgerrit has quit IRC09:47
*** haomaiwang has joined #openstack-swift09:47
*** openstackgerrit has joined #openstack-swift09:48
*** acoles_ is now known as acoles09:52
*** venkat has joined #openstack-swift09:52
acolesmahatic: congratulations! (just read scrollback) great news09:57
*** aix has joined #openstack-swift09:57
*** haomaiwang has quit IRC09:58
acolestimburke: yeah i assume the size bars are based on absolute size. i was being playful :)09:59
*** haomaiwa_ has joined #openstack-swift10:02
*** rledisez has joined #openstack-swift10:04
*** haomaiw__ has joined #openstack-swift10:06
*** haomaiwa_ has quit IRC10:07
openstackgerritCharles Hsu proposed openstack/python-swiftclient: Add another content-type for directory.
*** shakamunyi has quit IRC10:20
*** shakamunyi has joined #openstack-swift10:21
*** superflyy has quit IRC10:21
*** barra204 has joined #openstack-swift10:21
*** haomaiw__ has quit IRC10:24
*** haomaiwang has joined #openstack-swift10:28
openstackgerritOndřej Nový proposed openstack/swift: Show UTC time in swift-recon.
*** haomaiwang has quit IRC10:31
*** haomaiwa_ has joined #openstack-swift10:33
*** venkat has quit IRC10:34
*** diogogmt has quit IRC10:35
mahaticacoles: thank you! :)10:37
*** takashi has quit IRC10:45
*** venkat has joined #openstack-swift10:49
*** haomaiwa_ has quit IRC11:01
*** haomaiwang has joined #openstack-swift11:01
*** yatin has quit IRC11:17
*** yatin has joined #openstack-swift11:18
*** silor1 has joined #openstack-swift11:29
*** silor has quit IRC11:29
*** silor1 is now known as silor11:29
*** yatin has quit IRC11:33
*** sanchitmalhotra has quit IRC11:36
mahaticnew gerrit UI is kinda off-putting :\11:42
*** briancurtin has quit IRC11:53
*** rcernin has quit IRC12:00
*** haomaiwang has quit IRC12:01
*** haomaiwang has joined #openstack-swift12:01
*** km has quit IRC12:07
*** jmccarthy has quit IRC12:08
*** SkyRocknRoll has quit IRC12:11
openstackgerritAlistair Coles proposed openstack/swift: Fix incorrect kwarg in auth middleware example
*** rcernin has joined #openstack-swift12:15
*** tsg has joined #openstack-swift12:21
*** suresh has joined #openstack-swift12:24
*** silor has quit IRC12:25
sureshhi How can i set quota-bytes to another user from another user12:25
sureshplease any one hlep12:26
*** briancurtin has joined #openstack-swift12:29
*** suresh has quit IRC12:30
flaper87notmyname: mattoliverau et al: I read the scrollback and I very much agree with the proposal. I'll now take it to the glance team's meeting today and move forward with that option if there are no objections12:30
flaper87"that option" being #3 (joint effort)12:31
flaper87thanks guys for discussing it and helping out with the swift driver12:31
*** eranrom has joined #openstack-swift12:32
eranromI am looking for a coles. Seen one?12:32
*** delattec has quit IRC12:34
*** ppai has quit IRC12:41
*** janonymous_ has joined #openstack-swift12:43
*** haomaiwang has quit IRC12:43
*** tsg has quit IRC12:44
acoleseranrom: hi12:47
*** tsg has joined #openstack-swift12:47
eranromHi. Wanted to cosult with you on the following:12:47
eranromapparently container sync has an 'ancient' bug that was dicovered in reccent excavation12:48
eranromit was not dicovered but rather discovered12:48
eranromAll operations from the sync daemon to the remote proxy are done using SimpleClient12:49
eranromSimpleClient uses urllib which throws urllib exceptions. The code, however, catches ClientException12:49
eranromThis e.g. causes the following problem: when 'replicating' a delete and the proxy response is urllib.HTTPError(code=404), the daemon code misses that and considers this as a failure12:51
eranromyielding a situation where the delete will get replicated forever, and the relevant sync point does not advance12:51
eranromSo, the fix is trivial, but I am not sure if I should include this in a new patch or not.12:52
acoleseranrom: sounds like a separate bug to fix separately from your performance improvement patch12:53
acoleseranrom: did I see a bug report for this today? or did i imagine that?12:53
eranromacoles: Well not from me.12:54
eranromacoles: I did mention this yesterday in the IRC12:54
eranromacoles: ok, will do. Thanks!12:54
acolesah, maybe that was where i saw it12:55
eranromacloes: :-)12:55
eranromacoles: Well anyway I will check and open a bug if necessary.12:55
acoleseranrom: btw, you just reminded me to mention something i saw when playing with container sync for your patch12:55
eranromacoles: all ears12:56
acoleseranrom: i set up a src/dest container pair, uploaded an object and ran sync...12:56
acoleseranrom: but object not sync'd :( then I realised I had not set sync-key on dest (a common mistake of mine!)12:56
*** kei_yama has quit IRC12:57
acoleseranrom: sorry, did not set key on src. So I set sync-key onsrc, repeat sync, and IIRC, still no object sync'd.12:57
acolesand i seem to remember seeing in code that the sync point is only reset when sync-to is set, but not sync-key12:58
eranromacoles: a-ha. I will check that. and add a test12:58
eranromacoles: Question would be again: different patch?12:58
acoleseranrom: now, I'm not sure if that is a bug or a deliberate choice, because people may want to reset sync keys from time to time without resetting the sync points.12:58
*** links has quit IRC12:58
eranromacoles: hmm. thinking...12:59
acoleseranrom: different patch.12:59
acoleseranrom: but I would have expected the fact that the first sync attempt failed (not auth'd) to cause the sync points to not move ahead??13:00
acoleseranrom: caveat - i may have just got confused but I thought I would mention it since you are the sync expert ;)13:00
acoleseranrom: also (just quickly, i have to leave...) I will wait for next version of the symlinking patch before reviewing again, correct? assuming you will fix the .pending file issue?13:01
eranromacoles: That is correct. The sync point should not have moved. at least not both of them. At any rate I will test that. If it moves then this is clearly a bug.13:01
eranromacoles: So for the performance patch, I will upload a new one soon fixing the symlink issue you found.13:02
*** cdelatte has joined #openstack-swift13:02
acoleseranrom: ok, great. back later/13:02
*** lpabon has joined #openstack-swift13:05
*** cdelatte has quit IRC13:06
openstackgerritMerged openstack/swift: Document pretend_min_part_hours_passed
*** Zyric has quit IRC13:09
*** dslevin_ has joined #openstack-swift13:09
*** nakagawamsa has quit IRC13:10
eranromacloes: Thanks! back later as well13:13
*** tsg has quit IRC13:14
*** daddyjoseph97 has joined #openstack-swift13:15
*** briancurtin has quit IRC13:17
*** briancurtin has joined #openstack-swift13:17
*** rcernin has quit IRC13:21
*** briancurtin_ has joined #openstack-swift13:23
*** eranrom has quit IRC13:27
*** breitz has quit IRC13:32
*** breitz has joined #openstack-swift13:33
*** links has joined #openstack-swift13:33
*** Zyric has joined #openstack-swift13:52
*** venkat has quit IRC13:55
*** geaaru has quit IRC13:55
*** dslevin_ has quit IRC13:57
*** silor has joined #openstack-swift14:01
*** venkat has joined #openstack-swift14:03
*** dustins|out is now known as dustins14:10
*** haomaiwang has joined #openstack-swift14:11
*** venkat has quit IRC14:25
openstackgerritAlistair Coles proposed openstack/swift: Pass HTTP_REFERER down to subrequests
*** blmartin has joined #openstack-swift14:44
openstackgerritGanesh Maharaj Mahalingam proposed openstack/swift: Update SAIO setup for Ubuntu 14.04
gmmahaGood Morning14:52
*** petertr7_away is now known as petertr714:53
*** haomaiwang has quit IRC15:01
*** briancurtin_ has quit IRC15:01
*** haomaiwang has joined #openstack-swift15:01
*** rohit_ has joined #openstack-swift15:04
*** badari has joined #openstack-swift15:18
*** petertr7 is now known as petertr7_away15:20
*** haomaiwang has quit IRC15:25
*** daemontool has joined #openstack-swift15:27
*** janonymous_ has quit IRC15:47
*** links has quit IRC15:50
*** esker has joined #openstack-swift15:54
*** david-lyle_ has joined #openstack-swift15:55
*** esker has quit IRC15:56
*** esker has joined #openstack-swift15:56
*** david-lyle_ has quit IRC15:57
*** esker has quit IRC15:57
*** david-lyle has quit IRC15:59
*** diazjf has joined #openstack-swift16:04
*** david-lyle has joined #openstack-swift16:05
*** pgbridge has joined #openstack-swift16:08
*** lcurtis has joined #openstack-swift16:10
*** klrmn has joined #openstack-swift16:12
*** janonymous_ has joined #openstack-swift16:13
*** diogogmt has joined #openstack-swift16:17
notmynamegood morning16:19
gmmahagood morning notmyname16:21
pdardeaugood morning16:21
gmmahakota_: got a sec?16:23
kota_gmmah: sorry just in bed but...16:25
*** djc_ has joined #openstack-swift16:25
gmmahaohh then dont owrry16:26
gmmahanothing that important16:26
gmmahaand just realized its 1AM or so there.. :)16:26
kota_Yeah 1am exactly.16:27
gmmahaohh.. sorry for the ping :|16:27
djc_I am seeing a high number of client disconnect error messages in /var/log/messages. Any idea on what could be causing this? Here is a snippet:
kota_If you write down, i will see it tomorrow morning by scroll backing16:28
djc_I am also seeing a high number of TIME_WAIT (about 3000) when running netstat.16:30
*** minwoob has joined #openstack-swift16:32
*** badari has quit IRC16:36
gmmahakota_: sure. aded it to gerrit as well.. We sorta have to install pip that way as runnning pip install -U pip without removing the package will fail with 'pakcage maintained by the OS. Cannot upgrade' and after i remove it, I no longer have a pip package to do 'pip install -U pip'16:41
gmmahahence the down of the script and calling it to install pip16:41
*** marcusvrn_ has joined #openstack-swift16:42
-openstackstatus- NOTICE: Zuul is moving in very slow motion since roughly 13:30 UTC; the Infra team is investigating.16:43
*** petertr7_away is now known as petertr716:44
*** diazjf has quit IRC16:50
*** diazjf has joined #openstack-swift16:52
*** dmorita has joined #openstack-swift16:58
*** dmorita has quit IRC17:00
*** changbl has quit IRC17:00
*** dmorita has joined #openstack-swift17:02
*** djc_ has quit IRC17:03
*** diogogmt has quit IRC17:04
*** diogogmt has joined #openstack-swift17:08
*** lyrrad has joined #openstack-swift17:12
*** rledisez has quit IRC17:18
*** diazjf has quit IRC17:19
*** jistr has quit IRC17:19
*** changbl has joined #openstack-swift17:19
*** mac_ified has joined #openstack-swift17:21
*** diazjf has joined #openstack-swift17:23
*** arnox has quit IRC17:24
*** diazjf has quit IRC17:26
*** badari has joined #openstack-swift17:26
*** badari has quit IRC17:32
*** petertr7 is now known as petertr7_away17:47
*** dmorita has quit IRC17:49
*** gyee has joined #openstack-swift17:49
*** badari has joined #openstack-swift17:54
*** dmorita has joined #openstack-swift17:56
*** dmorita has quit IRC17:56
*** dmorita has joined #openstack-swift17:56
*** zhill has joined #openstack-swift17:56
*** esker has joined #openstack-swift18:05
*** esker has quit IRC18:12
*** eranrom has joined #openstack-swift18:15
*** badari has quit IRC18:16
*** zhill has quit IRC18:17
*** esker has joined #openstack-swift18:20
*** zhill has joined #openstack-swift18:21
*** zhill has quit IRC18:21
*** zhill has joined #openstack-swift18:21
*** petertr7_away is now known as petertr718:44
*** esker has quit IRC18:47
*** esker has joined #openstack-swift18:47
*** klrmn has quit IRC18:53
*** klrmn has joined #openstack-swift18:53
*** petertr7 is now known as petertr7_away18:54
*** gyee has quit IRC18:54
*** petertr7_away is now known as petertr718:54
*** esker has quit IRC18:57
*** acoles is now known as acoles_19:01
*** esker has joined #openstack-swift19:04
*** esker has quit IRC19:07
*** minwoob has quit IRC19:15
*** zhill has quit IRC19:17
*** zhill has joined #openstack-swift19:20
*** badari has joined #openstack-swift19:23
*** zul has quit IRC19:25
*** zul has joined #openstack-swift19:27
then3rdWith my juno upgrade having gone great, are there any gotchas that I might have missed for the jump to liberty? I haven't found any new or mondified configuration options that will be troubleshome. (so far) Just things to enable later.19:30
notmynamethen3rd: that sounds right19:31
*** badari has quit IRC19:34
*** ianbrown has quit IRC19:35
*** ianbrown has joined #openstack-swift19:35
then3rdCool. You guys have made this a cakewalk!19:37
openstackgerritBen Martin proposed openstack/swift: Print min_part_hours lockout time remaining
blmartinjust typed a long comment and gerrit made me re-sign in...... lost the comment....19:45
flaper87notmyname: so, there's no way to give permissions on a specific subdir. This means, I'll have to add swift-core to glance_store directly and grant access on the whole repo19:49
flaper87(et al ^)19:50
flaper87I'm ok with that19:50
flaper87I just want to know if you guys are ok with that as well. I'm obviously not expecting reviews from the swift team on glance-store other than the swift driver19:50
flaper87and there's obviously trust that the swift team will just approve patches there19:50
claygflaper87: we're a trustworthy bunch, except glange - he's sneaky19:51
flaper87I knew he was... I mean, his nick looks almost like glance and that cannot be good19:52
then3rdAre object-expirer and container-reconciler new, and should I set them up before configuring multiple regions? I can't figure out if they're recomended or required.19:53
*** blmartin has quit IRC19:53
claygthen3rd: -expirer is required to reap objects created with x-delete-at19:55
*** blmartin has joined #openstack-swift19:55
notmynameflaper87: sounds good to me19:56
claygthen3rd: the reconciler is only needed if you're running more than one object storage policy19:56
flaper87awesome, I'll proceed with the addition and communicate this to the m-l19:57
ZyricGood morning19:57
openstackgerritJohn Dickinson proposed openstack/swift: remove beta label from EC
notmynameclayg: acoles_: peluse: ^19:58
then3rdclayg: Okay, we're node using x-delete-at. Though, if we setup erasure coding or some other policy I'll need to worry about the reconciler. Thanks!20:00
claygthen3rd: werd20:00
claygnotmyname: did we loose the ec tagged bugs lp link on the priority reviews wiki?20:01
notmynameperhaps. link is
claygnotmyname: re beta label - i'm still sorta freaked out about this one ->
openstackLaunchpad bug 1496636 in OpenStack Object Storage (swift) "EC: Chunked transfer/commit protocol is *not* HTTP" [Undecided,New]20:04
claygI think i decided that we're going to subclass the eventlet HTTPProtocol class anyway and can have some sort of backwards compat shim in the putters to decide if it it should send 0\r\n\r\n before the commit or not depending on the version of eventlet running on the object node - but otherwise it make take a "shutdown your cluster upgrade eventlet and then turn it back on"20:05
*** diazjf has joined #openstack-swift20:06
claygnotmyname: other than that I think I agree things are business as usual with ec storage policies - and acoles has a lot of great patches we should get merged - and there's some bugs wrt reporting that hopefully we can prioritize along with all the other great stuff going on20:07
notmynameclayg: yeah. muy thinking is not that it's bug free or that there aren't places to improve. but (1) it's treated as a normal thing in swift that we're reasonably confident in and (2) with the current beta label on it, we get a chicken-and-egg problem of nobody wanting to run it20:10
notmynameI tried something for fun this morning20:12
claygnotmyname: well ok, I'll just feel bad I guess if we end up fixing that bug and having release notes that say you have to shutdown your entire cluster when you upgrade because you deployed an EC policy20:12
notmynameI checked out the very first commit of swift and ran those functional tests against current HEAD of master20:13
claygI mean - I *think* we can maybe do it rolling - but it may end up being some janky code20:13
*** silor has quit IRC20:13
clayghow'd we do!?20:13
notmynameresults are pretty good, I think20:13
*** diazjf has quit IRC20:14
notmynamewe've changed utf-8 to utf8 in a content type response20:14
notmynamewe changed the exact error body from "Invalid UTF8" to "Invalid UTF8 or contains NULL"20:15
clayg90% compat, not bad20:15
*** nadeem has joined #openstack-swift20:15
notmynamewe fail a test that was testing a bug in the first version (old swift used to return 412 if the account existed but no token give. now we always give 401 to any unauthed request so as not to leak info)20:16
clayglike a boss20:16
notmynamealso, the 403 vs 405 is because my SAIO has allow_account_management set to true20:16
notmynamethe tests expect a 405 method not allowed20:16
notmynameI don't know what the first errors are20:17
notmynamethe 413 and 416 errors20:17
openstackgerritTimur Alperovich proposed openstack/swift: Correctly handle keys starting with the delimiter.
notmynameso of the 14 errors/failures, I think 10 of them are known and expected and not "incompatibilities"20:17
claygtimur: slry!?  we're still not doing delimiter's correctly!?20:18
notmynameoh! the invalid length thing is because the old tests assume that every swift cluster in the world has a max object size of 5*2**30+220:20
notmynamenot sure about the range one20:20
notmynameso that's only 2 failures20:20
*** diogogmt has quit IRC20:26
*** esker has joined #openstack-swift20:32
*** esker has quit IRC20:32
*** blmartin has quit IRC20:32
*** esker has joined #openstack-swift20:33
then3rdWhen a storage node is down (and it's throwing rsync errors) the cluster doesn't automatically rebalance, correct? I would have to manaully do so (remove the node) with the ring-builder?20:39
*** silor has joined #openstack-swift20:40
claygthen3rd: right, if it's just like network connection errors that replica will just get skipped until the node comes back20:40
clayg... or the ring changes20:40
claygthe reason is of course because of the pain of the over-replication for a temporary network blip, or node reboot20:41
then3rdAwesome! So, it's not constantly rebuilding like ceph. Hooray for eventual consistency.20:47
*** dslevin has quit IRC20:48
then3rdWhat happens if so many storage nodes are down that an object can't be retrieved?20:49
glangeyou have to keep fixing hardware in a swift cluster20:50
glangeand not get too far behind20:50
glangethe software is only part of it20:51
then3rdThat part makessese. What if every storage node were simultaneously powered off.. what are the odds of dealing with data loss? Since it doesn't rebuild around the failuers, nothing really changes as long as all nodes come back?20:54
then3rdFrom what I understand the auditors would deal with any corruption.20:56
*** daemontool has quit IRC20:57
*** daemontool has joined #openstack-swift20:58
glangeyeah, when the nodes came back, everything would be fine20:58
glangethere is a small chance of data loss if 3 or more drives break at once20:59
glangeyou can do math to figure the chance out20:59
glangebut hardware isn't automatically taken out of the ring21:00
glangeso if something is down, you'll have less than 3 copies available21:00
glangeit'll only push to new/different hardware if the broken hardware is taken out of the ring21:01
glangeon puts, bad hardware is avoided by putting the object on handoff nodes21:01
then3rdAnother great answer. Exactly what I was hoping to hear.. it's all very logical.21:03
glangeit was designed by vulcans21:04
glangelike from star wars21:04
*** esker has quit IRC21:06
*** dslevin has joined #openstack-swift21:07
then3rdHow are you guys handling intergalactic latency? I could imagine a few light years between regions could present some difficulties.21:07
glangenetwork cables and worm holes, that's all I can say21:07
then3rdWonderful. :'-)21:09
glangeclayg can tell you more21:14
glangeI'm out!21:14
then3rdlater! thanks again!21:14
then3rdSo, if I had a large (or even small) culster, it would be safe to automatically (i.e ansible) upgrade multiple storage nodes simultaneously as long as the failure rate doesn't go up?21:17
*** zhill has quit IRC21:22
*** silor has quit IRC21:22
*** badari has joined #openstack-swift21:37
pdardeaumattoliverau: morning21:41
*** blmartin has joined #openstack-swift21:41
notmynameI just found a make pop-up store a block from the office.
*** dmorita has quit IRC21:43
*** dmorita has joined #openstack-swift21:43
*** zhill has joined #openstack-swift21:46
then3rdIt would seem that something isn't quite right with this proxy after upgrading. LookupError: Entry point 'swift' not found in egg 'ceilometer'21:51
then3rdI removed it from the pipeline and filter, starts fine.21:51
notmynamethen3rd: sounds like you have some middleware referenced that isn't found to be installed21:51
*** trifon has quit IRC21:51
openstackgerritMerged openstack/swift: Minor cleanup of repeated identical test assertions
then3rdnotmyname: Not sure why that would be, ceilometer hasn't been removed.21:58
*** zaitcev has quit IRC22:00
*** zaitcev has joined #openstack-swift22:03
*** ChanServ sets mode: +v zaitcev22:03
*** trifon has joined #openstack-swift22:07
*** ianbrown has quit IRC22:07
*** ianbrown has joined #openstack-swift22:08
*** zaitcev has quit IRC22:08
*** dustins is now known as dustins|out22:12
*** gyee has joined #openstack-swift22:13
then3rdGot it! Middleware used to meter Swift was previously packaged in Ceilometer and is now deprecated. It is now separated into it's own library: ceilometermiddleware.22:15
notmynameah, ok22:16
*** petertr7 is now known as petertr7_away22:19
*** zaitcev has joined #openstack-swift22:23
*** ChanServ sets mode: +v zaitcev22:23
mattoliverauthen3rd: you should blog about your upgrade experience, coming at this new, it would be interesting for others to learn and what we may be missing in any documentation :)22:32
*** alejandrito has joined #openstack-swift22:32
then3rdI hope I'm not asking too many noob questions or bothering too much. I don't have any co-workers to bounce questions off of, unfortunately. I'm a little young to have had much experince with IRC and etiquette.22:33
then3rdI've never blogged before either. I don't know how to find the time to. I can barely get my personal notes and repos organized, let along take time to blog it. Maybe I should...22:33
then3rdI have documented everything I've done for our wiki, though. Would be easy to rip-off and annotate for a blog.22:34
*** dslevin has quit IRC22:35
mattoliverauthen3rd: no I think it's awesome :)22:36
pdardeauthen3rd: i think you're questions here are great22:36
then3rdI've notice that there are some things missing in the openstack docs. Particularly some of the steps I've been trying to figure out this week. The answers are there but sometimes scattered about. I guess everyone has things customized differently too.22:37
mattoliverauthen3rd: ask more.. I just think it'll be interesting to see if we can improve documentation so more people can use swift ;)22:37
pdardeauyou can also revisit the logs on eavesdrop to help jog your memory for blogging22:37
mattoliverauthen3rd: or if you go openstack-ansible swift route, we can make sure we can/have captured all the upgrade points :)22:38
mattoliverauthen3rd: your doing a great job, keep it up :)22:38
then3rdI've had great things to say to my friends and coworkers. This community is awesome and helpful. It just keeps getting better.22:39
then3rdIf anyone has a linkedin and would like to connect, shoot me a PM.22:39
notmynamethen3rd: where are you geographically?22:40
openstackgerritEran Rom proposed openstack/swift: Container-Sync to iterate only over synced containers
*** rohit_ has quit IRC22:41
then3rdnotmyname: I've been in Utah Valley for the last couple years. I moved here from a lifetime in north-east Indiana.22:43
*** dslevin has joined #openstack-swift22:43
*** rohit_ has joined #openstack-swift22:43
notmynamecool. I've always wanted to go snow skiing in utah22:43
notmynameI've heard it's the best in the world :-)22:44
then3rdI've still never been skiing, that's what everyone says.22:44
then3rdnever been on a rollercoaster either.22:44
*** eranrom has quit IRC22:45
*** trifon has quit IRC22:45
*** eranrom has joined #openstack-swift22:45
then3rdWhere are you folks at?22:47
mattoliverauMelbourne Australia22:48
then3rdthat explains the au. :)22:48
mattoliverauYup ;)22:48
mattoliveraunotmyname: sounds like a midcycle to Utah over your winter is in order ;)22:49
mattoliverauNote it's going to be 102 here today, /me thinks I'll need a swimming break at some point :p22:51
notmynamethat's awesome22:51
notmynamethen3rd: I'm in San Francsico22:51
notmynamethen3rd: we've got people in here from NZ, Aus, Japan, India, Israel, Germany, France, UK, Ireland, Brazil, USA, Canada. and I probably left out some22:53
*** nadeem has quit IRC22:56
pdardeaunotmyname: then3rd: and Texas (a whole 'nuther country) ;-)22:58
then3rdCool! So notmyname, your're at SwiftStack.. your product would make my life easier (but not as much fun ;)23:02
*** vinsh has quit IRC23:03
notmynamethen3rd: yeah, I am at swiftstack23:04
notmynamethen3rd: happy to talk with you about that too ;-)23:04
notmyname(but I try to keep the sales stuff out of this channel) :-)23:05
*** km has joined #openstack-swift23:05
then3rdFor sure, for sure. We <3 opensource.23:05
*** badari has quit IRC23:09
*** vinsh has joined #openstack-swift23:09
*** ianbrown has quit IRC23:15
*** ianbrown has joined #openstack-swift23:16
*** dmorita has quit IRC23:22
*** badari has joined #openstack-swift23:27
*** dmorita has joined #openstack-swift23:29
*** dmorita has quit IRC23:30
*** dmorita has joined #openstack-swift23:30
*** pgbridge has quit IRC23:38
*** ho has joined #openstack-swift23:42
*** badari_ has joined #openstack-swift23:44
*** darrenc is now known as darrenc_afk23:44
*** badari has quit IRC23:46
*** lyrrad has quit IRC23:46
hogood morning!23:47
*** zaitcev has quit IRC23:48
*** zaitcev has joined #openstack-swift23:48
*** ChanServ sets mode: +v zaitcev23:48
*** esker has joined #openstack-swift23:48
mattoliverauho: morning23:52
homattoliverau: morning!23:53
pdardeauho: good morning!23:53
openstackgerritBen Martin proposed openstack/swift: Print min_part_hours lockout time remaining

Generated by 2.14.0 by Marius Gedminas - find it at!