| *** _JZ_ has joined #openstack-swift | 00:01 | |
| *** garthb__ has quit IRC | 00:04 | |
| *** haomaiwang has joined #openstack-swift | 00:21 | |
| *** haomaiwang has quit IRC | 00:26 | |
| *** lyrrad has quit IRC | 00:26 | |
| *** Suyi has joined #openstack-swift | 00:51 | |
| *** StraubTW has joined #openstack-swift | 01:00 | |
| *** Suyi has quit IRC | 01:26 | |
| *** StraubTW has quit IRC | 01:30 | |
| *** StraubTW has joined #openstack-swift | 01:31 | |
| *** rohit_ has quit IRC | 01:37 | |
| *** sgundur1 has left #openstack-swift | 01:52 | |
| *** mwheckmann has joined #openstack-swift | 02:01 | |
| notmyname | reminder to those online now, no meeting tomorrow | 02:01 |
|---|---|---|
| *** lakshmiS has quit IRC | 02:19 | |
| *** dmorita has quit IRC | 02:26 | |
| *** gyee has quit IRC | 02:27 | |
| *** dmorita has joined #openstack-swift | 02:27 | |
| openstackgerrit | Tim Burke proposed openstack/swift: Close all versioned_writes subrequests' app_iters https://review.openstack.org/307937 | 02:30 |
| *** dmorita has quit IRC | 02:32 | |
| *** delattec has quit IRC | 02:40 | |
| *** sekrit is now known as CIA | 02:44 | |
| *** sgundur1 has joined #openstack-swift | 02:44 | |
| *** sgundur1 has left #openstack-swift | 02:44 | |
| *** dmorita has joined #openstack-swift | 02:50 | |
| *** tsimp06 has joined #openstack-swift | 02:52 | |
| *** dmorita has quit IRC | 02:55 | |
| *** sgundur1 has joined #openstack-swift | 02:59 | |
| *** sgundur1 has left #openstack-swift | 02:59 | |
| *** mingdang1 has joined #openstack-swift | 03:06 | |
| janonymous | ntata: Hello, Please use this patch for testr and modify it according to requirements: https://review.openstack.org/#/c/257053/ | 03:16 |
| patchbot | janonymous: patch 257053 - swift - Using ostestr to run unittests | 03:16 |
| notmyname | janonymous: are you still working on that one? I think ntata is going to do a different one first | 03:17 |
| notmyname | or to rephrase, are you still interested in porting the unittests to use testr? | 03:18 |
| janonymous | notmyname: Hey | 03:18 |
| notmyname | janonymous: and do we get to see you next week in austin?! | 03:18 |
| janonymous | notmyname: that patch does the porting except the merge conflict part :) | 03:18 |
| notmyname | janonymous: from a quick glance at it (not a full review), I agree with tdasilva's -1. looks like a good start, but perhaps in the wrong place | 03:20 |
| janonymous | notmyname: ntata can continue on that after doing the priority first no problem, i am involved with my company's project's release so could not give much time on that | 03:20 |
| notmyname | ah, ok | 03:20 |
| *** links has joined #openstack-swift | 03:21 | |
| janonymous | notmyname: yes i am planning to see awesome swifties :) | 03:21 |
| notmyname | janonymous: eg the idea of having a single runner that takes the test type is interesting. I like that. but probably the wrapper scripts (.functests, etc) should be updated to call that, instead of doing it in something called "functests". that's confusing | 03:21 |
| janonymous | notmyname: Sure it could be used as a base patch, it needs improvement for sure | 03:21 |
| notmyname | janonymous: great! I'm glad you'll make it :-) | 03:21 |
| janonymous | notmyname: :) | 03:23 |
| janonymous | notmyname: are the session's topic decided ? | 03:25 |
| notmyname | janonymous: yep. they're scheduled, and you can see the ehterpads in the channel topic | 03:26 |
| janonymous | notmyname: great !! | 03:26 |
| *** asettle has quit IRC | 03:27 | |
| janonymous | notmyname: I will go through the etherpad :), Hope to see all of you in summit :) | 03:29 |
| *** sgundur1 has joined #openstack-swift | 03:29 | |
| *** sgundur1 has left #openstack-swift | 03:29 | |
| *** dmorita has joined #openstack-swift | 03:30 | |
| *** dmorita has quit IRC | 03:37 | |
| *** dmorita has joined #openstack-swift | 03:38 | |
| *** asettle has joined #openstack-swift | 03:43 | |
| *** nadeem has joined #openstack-swift | 03:50 | |
| *** ekarlso has quit IRC | 04:11 | |
| *** hosanai has quit IRC | 04:13 | |
| *** hosanai has joined #openstack-swift | 04:14 | |
| *** ChanServ sets mode: +v hosanai | 04:14 | |
| *** mingdang1 has quit IRC | 04:16 | |
| *** klrmn has quit IRC | 04:19 | |
| *** ekarlso has joined #openstack-swift | 04:24 | |
| *** dmorita has quit IRC | 04:25 | |
| *** SkyRocknRoll has joined #openstack-swift | 04:58 | |
| *** mwheckmann has quit IRC | 05:01 | |
| *** ppai has joined #openstack-swift | 05:31 | |
| *** asettle has quit IRC | 05:32 | |
| *** asettle has joined #openstack-swift | 05:33 | |
| *** nadeem has quit IRC | 05:37 | |
| *** nadeem has joined #openstack-swift | 05:47 | |
| *** chip_ has joined #openstack-swift | 05:49 | |
| *** chip_ has quit IRC | 05:52 | |
| *** chip_ has joined #openstack-swift | 05:53 | |
| *** asettle has quit IRC | 05:55 | |
| *** asettle has joined #openstack-swift | 05:55 | |
| *** ChubYann has quit IRC | 05:56 | |
| *** zaitcev has quit IRC | 06:05 | |
| *** chip_ has quit IRC | 06:16 | |
| *** rcernin has joined #openstack-swift | 06:21 | |
| *** dmorita has joined #openstack-swift | 06:25 | |
| *** dmorita has quit IRC | 06:29 | |
| *** sheel has joined #openstack-swift | 06:47 | |
| *** tesseract has joined #openstack-swift | 06:47 | |
| *** tesseract is now known as Guest67082 | 06:48 | |
| *** asettle has quit IRC | 06:55 | |
| *** asettle has joined #openstack-swift | 06:56 | |
| *** a1|away is now known as JelleB | 06:59 | |
| *** permalac has joined #openstack-swift | 07:00 | |
| *** janonymous has quit IRC | 07:01 | |
| *** janonymous has joined #openstack-swift | 07:03 | |
| *** asettle has quit IRC | 07:03 | |
| *** acoles_ is now known as acoles | 07:08 | |
| mattoliverau | hosanai, kota_: haven't seen you guys on today. This is a reminder that there is _no_ swift meeting tomorrow morning, so you can sleep in. | 07:10 |
| mattoliverau | acoles: hey man! | 07:10 |
| acoles | mattoliverau: hey! hows it going? have read your parenting comments in scrollback :D did I say nothing can prepare you? | 07:13 |
| kota_ | mattolivearu: hey man! just be online to look logs but unfortunately I'm going to be offline soon. | 07:13 |
| kota_ | mattoliveau: but thanks | 07:13 |
| kota_ | acoles: o/ | 07:13 |
| acoles | kota_: hi | 07:13 |
| janonymous | mattoliverau, kota_, acoles: o/ | 07:14 |
| kota_ | janonymous: hello | 07:14 |
| kota_ | janonymous: will you be in Austin/ | 07:14 |
| kota_ | ? | 07:14 |
| janonymous | kota_: yes :) | 07:15 |
| janonymous | kota_: wbu ? | 07:15 |
| *** chmouel has joined #openstack-swift | 07:15 | |
| kota_ | janonymous: sure, i'm looking forwared to meeting you there ;) | 07:15 |
| *** acoles is now known as acoles_ | 07:16 | |
| janonymous | kota_: Sure, me too :) | 07:16 |
| janonymous | kota_: when is your arrival and departure? | 07:18 |
| *** acoles_ is now known as acoles | 07:18 | |
| acoles | ah, no meeting today | 07:18 |
| acoles | cschwede: did you see ^^ | 07:19 |
| kota_ | janonymous: flight in monday, deart at friday morning, this time my travel is too shorter than previous ones for some private reasons. | 07:19 |
| cschwede | acoles: Good Morning Alistair! Not yet, so thanks for the heads up - a little more sleep tonight ;) | 07:20 |
| acoles | cschwede: yes! I for one need it! | 07:20 |
| janonymous | kota_: Ohhkay :) | 07:20 |
| kota_ | good morning cschwede. | 07:21 |
| janonymous | cschwede: o/ | 07:21 |
| acoles | mahatic_: hi just saw your comments on patches - are you able to work more on that today? (just wondering if i should pick anything up) | 07:21 |
| cschwede | kota_: janonymous: Hello & good afternoon :) | 07:21 |
| *** jlvillal has quit IRC | 07:22 | |
| *** jlvillal has joined #openstack-swift | 07:22 | |
| mahatic_ | acoles: hello, yup should be able to push a patchset today. Will do it as early I can | 07:24 |
| mattoliverau | acoles: all is well, looking forward to sleeping on the long flight to Austin ;) | 07:25 |
| mahatic_ | acoles: I couldn't get to them early in the day, had some quarterly updates to attend | 07:25 |
| *** rledisez has joined #openstack-swift | 07:32 | |
| acoles | mahatic_: great thank you | 07:34 |
| acoles | mattoliverau: yes, sleep is the new currency ;) | 07:35 |
| *** mmcardle has joined #openstack-swift | 07:35 | |
| *** Psy has joined #openstack-swift | 07:43 | |
| *** Psy has left #openstack-swift | 07:44 | |
| *** nadeem has quit IRC | 07:55 | |
| *** kirubak has joined #openstack-swift | 08:09 | |
| kirubak | kirubak | 08:09 |
| *** jordanP has joined #openstack-swift | 08:14 | |
| *** mingdang1 has joined #openstack-swift | 08:17 | |
| *** jmccarthy has quit IRC | 08:18 | |
| *** jistr has joined #openstack-swift | 08:18 | |
| *** jmccarthy has joined #openstack-swift | 08:18 | |
| *** joeljwright has joined #openstack-swift | 08:20 | |
| *** ChanServ sets mode: +v joeljwright | 08:20 | |
| *** dosaboy_ is now known as dosaboy | 08:24 | |
| *** asettle has joined #openstack-swift | 08:42 | |
| *** Guest66327 has quit IRC | 08:43 | |
| *** david-lyle has quit IRC | 08:59 | |
| *** david-lyle has joined #openstack-swift | 09:00 | |
| *** stantonnet has quit IRC | 09:02 | |
| *** mingdang1 has quit IRC | 09:03 | |
| *** stantonnet has joined #openstack-swift | 09:04 | |
| *** kei_yama has quit IRC | 09:09 | |
| *** asettle has quit IRC | 09:11 | |
| *** asettle has joined #openstack-swift | 09:17 | |
| *** asettle has quit IRC | 09:54 | |
| *** asettle has joined #openstack-swift | 09:54 | |
| *** asettle has quit IRC | 09:59 | |
| *** geaaru has joined #openstack-swift | 10:03 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: crypto - byok support in keymaster https://review.openstack.org/307811 | 10:08 |
| *** asettle has joined #openstack-swift | 10:10 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: crypto - byok support in keymaster https://review.openstack.org/307811 | 10:16 |
| *** hosanai has quit IRC | 10:18 | |
| openstackgerrit | Ankur Jain proposed openstack/swift: Invalid parameter does not return 400 https://review.openstack.org/299225 | 10:29 |
| ntt | Hi, I'm trying to use the directory path based configuration in openstack swift (kilo). It seems that if I enable for example openstack-swift-account.service at boot, it doesn't load /etc/swift/account-server/1.conf and /etc/swift/account-server/2.conf but try to load /etc/swift/account-server.conf (that doesn't exist). It works normally if I use swift-init account start | 10:38 |
| ntt | someone can help me pleasE? | 10:38 |
| *** oshritf has joined #openstack-swift | 10:45 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: crypto - Encrypt flag for probe tests https://review.openstack.org/216974 | 10:50 |
| *** jordanP has quit IRC | 10:55 | |
| mahatic_ | acoles: i think the rebases and changes are done - ready to review i guess | 11:08 |
| * mahatic_ is afk | 11:09 | |
| *** asettle has quit IRC | 11:20 | |
| *** asettle has joined #openstack-swift | 11:21 | |
| *** dmorita has joined #openstack-swift | 11:26 | |
| *** haomaiwang has joined #openstack-swift | 11:27 | |
| *** mingdang1 has joined #openstack-swift | 11:27 | |
| *** dmorita has quit IRC | 11:31 | |
| *** ppai has quit IRC | 11:35 | |
| *** mmcardle has quit IRC | 11:42 | |
| *** ppai has joined #openstack-swift | 11:49 | |
| *** asettle has quit IRC | 11:49 | |
| *** daemontool has joined #openstack-swift | 11:50 | |
| *** haomaiwang has quit IRC | 12:01 | |
| *** SkyRocknRoll has quit IRC | 12:05 | |
| *** ppai has quit IRC | 12:10 | |
| *** cdelatte has joined #openstack-swift | 12:17 | |
| *** ppai has joined #openstack-swift | 12:24 | |
| *** mmcardle has joined #openstack-swift | 12:35 | |
| *** asettle has joined #openstack-swift | 12:37 | |
| *** admin6 has joined #openstack-swift | 12:37 | |
| *** asettle has quit IRC | 12:42 | |
| *** links has quit IRC | 12:46 | |
| *** mwheckmann has joined #openstack-swift | 12:47 | |
| *** klamath has joined #openstack-swift | 12:51 | |
| *** klamath has quit IRC | 12:51 | |
| *** klamath has joined #openstack-swift | 12:52 | |
| *** oshritf has quit IRC | 12:52 | |
| *** NM has joined #openstack-swift | 12:58 | |
| *** aimeeU has joined #openstack-swift | 13:02 | |
| *** StraubTW has quit IRC | 13:08 | |
| *** kirubak has quit IRC | 13:12 | |
| tdasilva | good morning | 13:26 |
| *** jordanP has joined #openstack-swift | 13:31 | |
| *** haomaiwang has joined #openstack-swift | 13:31 | |
| *** StraubTW has joined #openstack-swift | 13:34 | |
| ntt | Hi, I'm trying to use the directory path based configuration in openstack swift (kilo). It seems that if I enable for example openstack-swift-account.service at boot, it doesn't load /etc/swift/account-server/1.conf and /etc/swift/account-server/2.conf but try to load /etc/swift/account-server.conf (that doesn't exist). It works normally if I use swift-init account start | 13:35 |
| ntt | someone can help me pleasE? | 13:35 |
| *** esker has quit IRC | 13:37 | |
| *** dmsimard has left #openstack-swift | 13:40 | |
| *** ppai is now known as ppai|afk | 13:43 | |
| *** haomaiwang has quit IRC | 13:44 | |
| *** cdelatte has quit IRC | 13:48 | |
| ppai|afk | ntt, hi | 13:50 |
| ppai|afk | ntt, it's because of the way the .service scripts have been written | 13:51 |
| ppai|afk | ntt, http://pkgs.fedoraproject.org/cgit/rpms/openstack-swift.git/tree/openstack-swift-account.service?h=f23 | 13:51 |
| admin6 | Hi all, I’ve have a starting cluster whit a 14 partition power for account and container rings (in replica 4) but only 6 disks currently setup into them. When i perform a rebalance, I got messages like : | 13:51 |
| admin6 | part, devs_for_part))) | 13:52 |
| *** links has joined #openstack-swift | 13:52 | |
| ppai|afk | ntt, take a look at this: http://pkgs.fedoraproject.org/cgit/rpms/openstack-swift.git/tree/openstack-swift-account@.service?h=f23 | 13:53 |
| admin6 | RingValidationWarning: The partition 2206 has been assigned to duplicate devices | 13:53 |
| admin6 | Do you think I should care about these errors, and wil they vanish if a | 13:54 |
| admin6 | if I add additional disks into my rings? | 13:54 |
| *** sgundur1 has joined #openstack-swift | 14:00 | |
| *** _JZ_ has quit IRC | 14:01 | |
| *** cdelatte has joined #openstack-swift | 14:02 | |
| glange | admin6: your partion power might be too small | 14:03 |
| glange | that's only 16k partitions | 14:03 |
| glange | it might be difficult to get an even distribution of data with that few partitions | 14:04 |
| glange | and that duplicate device error isn't good | 14:04 |
| admin6 | too small? I was thinking on the opposite that it were too big for such a few disks inside it | 14:07 |
| admin6 | and you’re right, I’ve only 16384 patition in these rings | 14:08 |
| *** ametts has joined #openstack-swift | 14:11 | |
| glange | I believe the more partitions, the better chance of even distrubuition of replicas | 14:11 |
| glange | like if you only had 5 partitions, it seems a lot more likely that one of them would have way more replicas than another | 14:11 |
| glange | and that chance reduces as you add more partitions | 14:11 |
| glange | we have some big clusters, but we have 1 million plus partitions | 14:12 |
| glange | plus, more partitions are better when they have to move | 14:13 |
| glange | but ask around, maybe what you are doing is ok -- those are just what I think/am guessing | 14:13 |
| admin6 | glange: So I’ll try with a much higher partition power to see if that helps, thanks. However, I had an object ring with erasure coding 9+3 (so declared with 12 "replicas"), that I started with very few disks and a partition power of 19 wich got same type of errors. Reducing the partition power to 18 grealy enhance the number of "duplicate devices" error messages | 14:18 |
| admin6 | (greatly enhance : drastically reduced) | 14:19 |
| glange | that's possible | 14:19 |
| *** mingdang1 has quit IRC | 14:23 | |
| *** esker has joined #openstack-swift | 14:29 | |
| *** esker has quit IRC | 14:29 | |
| *** esker has joined #openstack-swift | 14:30 | |
| *** arch-nemesis has joined #openstack-swift | 14:33 | |
| *** arch-nemesis has joined #openstack-swift | 14:34 | |
| *** ntt has quit IRC | 14:39 | |
| *** hosanai has joined #openstack-swift | 14:42 | |
| *** ChanServ sets mode: +v hosanai | 14:42 | |
| *** sgundur has joined #openstack-swift | 14:45 | |
| *** haomaiwang has joined #openstack-swift | 14:52 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: crypto - byok support in keymaster https://review.openstack.org/307811 | 14:53 |
| *** nadeem has joined #openstack-swift | 14:58 | |
| *** garthb has joined #openstack-swift | 14:58 | |
| acoles | mahatic_: your byok patch has been really useful to me for putting together a demo for next week | 14:59 |
| *** ppai|afk is now known as ppai | 15:01 | |
| *** ppai_ has joined #openstack-swift | 15:06 | |
| *** amit213 has quit IRC | 15:06 | |
| *** ppai has quit IRC | 15:08 | |
| *** amit213 has joined #openstack-swift | 15:08 | |
| *** haomaiwang has quit IRC | 15:09 | |
| mahatic_ | acoles: nice! good luck with the demo | 15:09 |
| clayg | acoles: hey! are you at home? I have a question for you boys... | 15:12 |
| acoles | clayg: i'm here | 15:12 |
| clayg | acoles: I went to see a band last night "The Savages" - they're from London and they're pretty good! | 15:12 |
| clayg | acoles: have your boys ever heard of these gals? | 15:12 |
| acoles | clayg: oic yeah i would need to ask the boys, i haven't heard of them | 15:13 |
| clayg | :) | 15:13 |
| acoles | where did they play? | 15:13 |
| clayg | k, they were pretty raucous - not sure if they're really that big in the London rock scene but we had a good time | 15:14 |
| clayg | The Fillmore | 15:14 |
| *** cdelatte has quit IRC | 15:15 | |
| *** cdelatte has joined #openstack-swift | 15:17 | |
| acoles | clayg: looks good. whenever I think I know what the boys are into it seems they have moved on :) | 15:17 |
| acoles | clayg: i was looking for gigs in Austin since its the music capital of somewhere | 15:19 |
| acoles | openstack falls between paul simon and F1 grand prix i think | 15:19 |
| *** jordanP has quit IRC | 15:21 | |
| clayg | acoles: of the WORLD - the whole world man - and they hid it in the middle of Texas to keep out the yellow-bellies | 15:21 |
| *** haomaiwang has joined #openstack-swift | 15:22 | |
| jrichli | lol | 15:22 |
| clayg | acoles: i'd go see paul simon although I probably don't know his solo stuff as well | 15:22 |
| clayg | is F1 grand prix... a... band? | 15:22 |
| acoles | lol | 15:23 |
| acoles | clayg: you're confusing it with The Cars...now that was A band | 15:24 |
| *** StraubTW has quit IRC | 15:26 | |
| clayg | yeah that'd be good - wikipedia says they're mostly still alive and doing shows as of 2015 - maybe there's still hope!? | 15:27 |
| *** thingee has left #openstack-swift | 15:27 | |
| *** esker has quit IRC | 15:28 | |
| *** esker has joined #openstack-swift | 15:28 | |
| notmyname | good morning | 15:32 |
| *** StraubTW has joined #openstack-swift | 15:32 | |
| *** links has quit IRC | 15:32 | |
| *** dmorita has joined #openstack-swift | 15:34 | |
| *** siva_krishnan_ has quit IRC | 15:34 | |
| *** dmorita has quit IRC | 15:35 | |
| *** esker has quit IRC | 15:36 | |
| *** haomaiwang has quit IRC | 15:36 | |
| *** siva_krishnan_ has joined #openstack-swift | 15:36 | |
| notmyname | looks like it's been repeated enough (thanks!) but reminder that there is not meeting this week | 15:42 |
| notmyname | and of course no meeting next week either, in lieu of the summit | 15:43 |
| *** cdelatte has quit IRC | 15:43 | |
| notmyname | acoles: how's "have a good feature branch for the summit" looking? | 15:45 |
| acoles | notmyname: good news is that with mahatic's help we should have some BYOK support | 15:46 |
| notmyname | cool | 15:46 |
| acoles | but i'm wary of merging copy middleware on the feature branch - feels like its making work, so we may not be quite there | 15:47 |
| acoles | i can propose it so its on gerrit | 15:47 |
| acoles | but unpicking it could be painful | 15:47 |
| openstackgerrit | Merged openstack/swift: crypto - Update container with correct etag on POST https://review.openstack.org/305995 | 15:48 |
| tdasilva | acoles: we have one +2 :) | 15:48 |
| *** lakshmiS has joined #openstack-swift | 15:49 | |
| jrichli | re: unpicking could be painful: +1 | 15:49 |
| acoles | tdasilva: yeah i saw, on decoupling versioned writes correct? :) but IDK, i'm guessing the actual middleware patch has had less review time? | 15:49 |
| *** siva_krishnan_ has quit IRC | 15:49 | |
| tdasilva | acoles: yeah, i'm just looking for anything to be optimistic about :) | 15:50 |
| acoles | tdasilva: heh | 15:50 |
| *** cdelatte has joined #openstack-swift | 15:50 | |
| acoles | notmyname: i'm scripting a demo cos i know my hands will be too shakey to type! | 15:51 |
| jrichli | yeah, I think we all did a great job getting most of the branch "todos" worked. but COPY not being there is the big one. still some little todos, but there was a llot | 15:52 |
| jrichli | of progress | 15:52 |
| *** gyee has joined #openstack-swift | 15:54 | |
| notmyname | acoles: that is absolutely the thing to do :-) | 15:54 |
| notmyname | (scripting a demo) | 15:54 |
| *** mmotiani_ has joined #openstack-swift | 15:57 | |
| *** mmotiani_ has quit IRC | 15:57 | |
| *** nadeem has quit IRC | 15:58 | |
| *** mmotiani_ has joined #openstack-swift | 15:59 | |
| *** ntata- has joined #openstack-swift | 16:00 | |
| *** siva_krishnan_ has joined #openstack-swift | 16:02 | |
| timburke | good morning | 16:08 |
| *** cdelatte has quit IRC | 16:10 | |
| *** _JZ_ has joined #openstack-swift | 16:11 | |
| *** cdelatte has joined #openstack-swift | 16:12 | |
| *** ppai_ has quit IRC | 16:13 | |
| *** cdelatte has quit IRC | 16:14 | |
| *** lakshmiS has quit IRC | 16:15 | |
| *** lakshmiS has joined #openstack-swift | 16:15 | |
| *** cdelatte has joined #openstack-swift | 16:16 | |
| *** ChubYann has joined #openstack-swift | 16:17 | |
| *** hosanai has quit IRC | 16:17 | |
| *** SkyRocknRoll has joined #openstack-swift | 16:18 | |
| admin6 | Hi chat, Could one explain me what means error message type : "RingValidationWarning: The partition 1968 has been assigned to duplicate devices" when rebalancing a ring. I’m still fighting with this. | 16:19 |
| clayg | admin6: how old is this ring? | 16:20 |
| admin6 | it’s a fresh one declared today | 16:20 |
| *** mingyu has joined #openstack-swift | 16:21 | |
| clayg | admin6: it *means* that multiple replicas of the partition 1968 (good year) have been assigned to a device - it shouldn't happen - so we validate against. | 16:21 |
| clayg | it's like a constraint | 16:21 |
| clayg | admin6: that's *fascinating* - do you think you could reproduce the error? If you created a ring with the same params again? | 16:22 |
| *** jistr has quit IRC | 16:22 | |
| clayg | oh... maybe it's less facinating that I thought - that's a Warning - not an error | 16:22 |
| admin6 | clayg: yes I can, I’m just playing with new rings to understand this error | 16:23 |
| clayg | if you're running idk, swift 2.5? 2.4? somewhere in there it was just broken - it's when we started emiting the warning | 16:23 |
| admin6 | clayg: warning ;-) | 16:23 |
| clayg | it's fixed now! get you some swift 2.7! you should almost definately not get that invalid state on a fresh ring with swift 2.7 | 16:23 |
| *** lyrrad has joined #openstack-swift | 16:24 | |
| admin6 | I think It’s swift 2.6, (at least client python swift-client is 2.6) | 16:25 |
| clayg | i have no idea if the client and storage system have related version numbers | 16:26 |
| clayg | how did you install swift? | 16:26 |
| admin6 | sorry it’s version 2.5.0 : version embebed in ubuntu 15.10 | 16:27 |
| clayg | bummer :'( | 16:28 |
| clayg | luckily ring balancing is out of band to the servers and the format hasn't changed - so you can use swift-ring-builder from 2.7 with services installed on swift 2.5 (although I'd recommend just upgrading everyting because swift 2.7 is *awsome*) | 16:29 |
| *** arch-nemesis has quit IRC | 16:29 | |
| clayg | but... i suppose the part placement algorithm *did* work back then - it just took some be carefuls - e.g. you have to have as many devices as you do replicas? Can't put 3 replicas of each part on two devices without duplicate assignments | 16:31 |
| *** jmccarthy1 has quit IRC | 16:31 | |
| admin6 | ok. I think I’ll choose the upgrade solution :-) Are you interrested in getting the commands I used to reproduce the context? | 16:31 |
| *** cdelatte has quit IRC | 16:31 | |
| clayg | there's also some futzy stuff if you have zones that don't have roughly equal weights in them - or fewer failure-domains than replicas at any given tie | 16:32 |
| admin6 | clayg: no, I always have more disks than the number of replicas in all the rings (even the erasure coding one with 12 "replicas") | 16:32 |
| clayg | not... terribly so? as i recall it was sort of trivial to break the old placement stuff - like there was litterly dozens of things that could go wrong and many of them weren't even that uncommon or esoteric | 16:33 |
| clayg | ok, gl with the upgrade solution - if you wanna do a swift-ring-builder-analyzer template that might be useful - i could at least sanity check you -> http://docs.openstack.org/developer/swift/overview_ring.html#module-swift.cli.ring_builder_analyzer | 16:35 |
| clayg | admin6: but you could just upgrade too - i bet new swift will be great - eyes forward! | 16:36 |
| *** cschwede has quit IRC | 16:36 | |
| admin6 | clayg; Ok, many thanks for your expertise | 16:36 |
| *** tmoreira has quit IRC | 16:37 | |
| *** tmoreira has joined #openstack-swift | 16:41 | |
| *** cschwede has joined #openstack-swift | 16:43 | |
| *** arch-nemesis has joined #openstack-swift | 16:43 | |
| *** david-nesher has joined #openstack-swift | 16:45 | |
| *** Guest67082 has quit IRC | 16:46 | |
| *** mingyu has quit IRC | 16:47 | |
| *** zul has quit IRC | 16:48 | |
| *** cdelatte has joined #openstack-swift | 16:50 | |
| *** joeljwright has quit IRC | 16:52 | |
| *** rledisez has quit IRC | 16:52 | |
| *** chuck__ has joined #openstack-swift | 16:54 | |
| *** mingyu has joined #openstack-swift | 16:54 | |
| *** sgundur has quit IRC | 16:57 | |
| *** sbezverk has joined #openstack-swift | 16:58 | |
| *** thumpba has joined #openstack-swift | 16:58 | |
| sbezverk | Hello, we have to change default rsync port to different value, do you see any problem in using a non-default port and is there any dependecies hardcoded in configuration files for a default port? | 16:59 |
| *** NM has quit IRC | 16:59 | |
| *** sgundur- has joined #openstack-swift | 16:59 | |
| *** sgundur- has left #openstack-swift | 17:00 | |
| clayg | sbezverk: yes I believe that doesn't work by default? | 17:01 |
| *** sgundur has joined #openstack-swift | 17:01 | |
| clayg | well... i could be wrong... is possible device metadata get's interpolated into the rsync_module template - so you might be able to annotate devices - i'm not sure it's described in the docs? | 17:02 |
| sbezverk | +clayg thank you for your reply in kolla project we run swift in docker containers and due to some security issues have to change default port | 17:02 |
| clayg | sbezverk: ok, it's a good use case - let me check something | 17:03 |
| sbezverk | +clayg I cannot find any references or guidelines for this non-default port scenario | 17:03 |
| clayg | yeah - it's a little janky -> http://docs.openstack.org/developer/swift/misc.html#swift.common.utils.rsync_module_interpolation | 17:04 |
| clayg | looking at the source the device "meta" key *is* interpolated into the rsync_module - but there's no defined k/v syntax - so it'd be a weird coupling - but it might work - i've never tried it | 17:05 |
| *** SkyRocknRoll has quit IRC | 17:07 | |
| sbezverk | +clayg are you saying that it might work without any additional configuration? | 17:07 |
| clayg | no | 17:07 |
| clayg | it might work with additional sort of baroque undocumented configuration - but it also might now | 17:08 |
| clayg | *not | 17:08 |
| clayg | i was gunna bring up my dev instance and see if I could get you a little script to pay with | 17:08 |
| sbezverk | +clayg I would really appreciate if you do that | 17:09 |
| sbezverk | +clayg at this point for us it is a show stopper, as we have to run as root | 17:09 |
| notmyname | FYI, the proposed design summit split that was talked about a while back seems to be a done deal. the conversation seems to only be about the details of how each thing will work instead of if it will happen http://lists.openstack.org/pipermail/foundation-board/2016-April/000226.html | 17:10 |
| openstackgerrit | Andy McCrae proposed openstack/swift: Allow fallocate_reserve to be a percentage https://review.openstack.org/288011 | 17:12 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/174343 | 17:13 |
| *** klrmn has joined #openstack-swift | 17:16 | |
| notmyname | glange: are you going to be at the summit? | 17:18 |
| *** _JZ_ has quit IRC | 17:19 | |
| clayg | sbezverk: so you can deifnately stick a devices meta field into your rsync module - i think the syntax would line up such that you can make the rsync command come out pointint to a non-default port if you just fill in whatever port your rsync daemons are configured to listen on as "meta" on the device when building the ring and then update your rsync_module config opiton on your *-replicator sections -> https://gist.github.c | 17:21 |
| *** haomaiwang has joined #openstack-swift | 17:21 | |
| notmyname | clayg: that cut off in the middle of the gist url | 17:22 |
| clayg | boooo! | 17:22 |
| clayg | https://gist.github.com/clayg/93bcc37ec4f5fc7ab028f07ac4305148 | 17:22 |
| sbezverk | +clayg thank you | 17:22 |
| *** NM has joined #openstack-swift | 17:22 | |
| *** zaitcev has joined #openstack-swift | 17:22 | |
| *** ChanServ sets mode: +v zaitcev | 17:22 | |
| *** diazjf has joined #openstack-swift | 17:23 | |
| *** mmcardle has quit IRC | 17:24 | |
| *** mingdang1 has joined #openstack-swift | 17:24 | |
| sbezverk | +clayg when we deploy kolla we use master source and would like to continue doing so. Do you think swift team could expose rsync port parameter in the configuration file, so we do not have to hack the code. | 17:25 |
| clayg | sbezverk: the rsync_module template may need to be more like... idk "rsync://{replication_ip}:{meta}/object" ??? | 17:25 |
| *** ober37 has joined #openstack-swift | 17:25 | |
| clayg | sbezverk: the default seems to be "{replication_ip}::object" which gets joined into in front of '<device>/objects/<part>' | 17:26 |
| clayg | sbezverk: testing from the command line rsync doesn't seem to like "{replication_ip}:{meta}::object" - but was ok with "rsync://{replication_ip}:{meta}/object" - so you will probably have to play around with the exact syntax | 17:27 |
| clayg | sbezverk: yeah i mean I think it would just be documenting this configuration? I'm trying to tell you how you can configure swift for this - if it works you could document it for the next guy - probably somewhere in http://docs.openstack.org/developer/swift/replication_network.html | 17:28 |
| *** pc6411 is now known as pchng | 17:28 | |
| *** mingdang1 has quit IRC | 17:28 | |
| sbezverk | +clayg sure thing at first I thought it would require code change | 17:29 |
| sbezverk | +clayg if it is just to figure out right syntax for configuration item then it is exactly what we need. | 17:30 |
| *** yarkot has joined #openstack-swift | 17:32 | |
| tdasilva | notmyname: so basically they are replacing mid-cycle meetups with an developer oriented event planned by the foundation | 17:33 |
| clayg | sbezverk: ok, I think i have enough info in the gist - you need to update the config option - but also annotate your devices in the ring with metadata | 17:33 |
| *** diazjf has quit IRC | 17:33 | |
| notmyname | tdasilva: yes. (maybe?) | 17:33 |
| clayg | notmyname: they can take our design summits - but they can never take our FREEEDOOOOOM^wMid-Cycles | 17:35 |
| clayg | i love how the reaction to wow - the project organized mid-cycles are more productive than the foundation organized event - let's have the foundation organize the mid-cycle! | 17:35 |
| tdasilva | lol | 17:36 |
| tdasilva | nice way of looking at it | 17:36 |
| notmyname | that's the way it looks, but it's also something that we should be a part of next week and as any discussion continues | 17:37 |
| clayg | tdasilva: it's probably terribly unproductive acctually - i'm sort of a cynic | 17:37 |
| notmyname | (but to be fair, the way I saw it happen was "here's an idea..." followed by "well now that this idea is happening...") | 17:37 |
| tdasilva | but also reducing the developers meeting time from 4 to 2 times a year | 17:37 |
| tdasilva | clayg: no worries, i'm the same | 17:37 |
| notmyname | tdasilva: but it doesn't. there's still the summits | 17:37 |
| notmyname | and meeting together less often is the wrong way to go! | 17:37 |
| tdasilva | notmyname: yeah, but there wouldn't be a "design" time at the summit | 17:38 |
| notmyname | and if there is still some need to have other in-person meetings (mid-mid-cyles?), then that's 6 international trips a year for some people! that's terrible | 17:38 |
| tdasilva | just marketing summit with user <> developer time | 17:38 |
| clayg | tdasilva: you try putting more than 5 swift developers in the same city and see if there's no design going on | 17:38 |
| tdasilva | lol | 17:38 |
| tdasilva | clayg: can't argue with that | 17:38 |
| notmyname | tdasilva: but there is. the summits still have a place for the devs to talk to users/ops about what needs to be done | 17:39 |
| notmyname | and the new event is for figuring out how to do those things | 17:39 |
| notmyname | (I don't think you can actually separate those conversations, though. that's like uber-waterfall design) | 17:39 |
| notmyname | "We still expect many key developers will attend the the community-wide Summit events" <--- tdasilva, that includes you. congratulations | 17:40 |
| tdasilva | notmyname: talk about the "what" and in 3 months talk about the "how" | 17:40 |
| clayg | notmyname: I basically just hear you saying we'll have 4 foundation events a year to schedule our project organized events inbetween - if that means we would have done both jrichli's and acoles's events w/i the same OS relase I'm totally down | 17:41 |
| clayg | ... acctually it sounds exahusting | 17:41 |
| tdasilva | clayg: that would be 6 events | 17:41 |
| notmyname | someone pointed out to me that we're getting close to 10% of a year (or more) in travel related to swift. that's a lot more than some people signed up for | 17:42 |
| clayg | yeah seems like a lot - i'm good on... 4 events a year - crap... that's what we already have tho? | 17:42 |
| tdasilva | right, yeah, i've thought of that 10% number also | 17:42 |
| notmyname | yeah. right now we're at 4 times a year. once a quarter, basically | 17:42 |
| clayg | hrmmm... | 17:43 |
| clayg | maybe the *can* take our mid-cycles :'( | 17:43 |
| *** Suyi has joined #openstack-swift | 17:45 | |
| tdasilva | i'm also wondering about the plan to put these new PTGs in "cheaper" cities. I would this it's still a 1K people event, so there will be a need for a medium size city | 17:46 |
| notmyname | tdasilva: "don't worry. it's just the devs. you can put them in the cheap city" ;-) | 17:46 |
| *** dmorita has joined #openstack-swift | 17:48 | |
| *** dmorita has quit IRC | 17:48 | |
| notmyname | here's the cross-project session for discussing it | 17:48 |
| *** dmorita has joined #openstack-swift | 17:48 | |
| notmyname | https://www.openstack.org/summit/austin-2016/summit-schedule/events/9478 | 17:48 |
| clayg | tdasilva: was there a post on the ML recently about this? I've never heard of this "PTG"? | 17:48 |
| tdasilva | i bet you can't find a venue to host 1k people in westford or longmont | 17:49 |
| notmyname | clayg: yeah, I just linked it above (probably when you were writing that rsync gist) | 17:49 |
| notmyname | clayg: http://lists.openstack.org/pipermail/foundation-board/2016-April/000226.html | 17:49 |
| zaitcev | Albuquerque easily hosts 1000-people events, and it's only 600,000 people in the metro area. | 17:50 |
| notmyname | and the two most recent posts on http://ttx.re talk about it | 17:50 |
| *** mwheckmann has quit IRC | 17:50 | |
| notmyname | https://www.youtube.com/watch?v=e8TUwHTfOOU | 17:50 |
| tdasilva | clayg: so yeah, i'm putting together this email that notmyname just linked with an earlier email when they were proposing the idea. One of the reasons was that a smaller devs event could be hosted in "cheaper" cities, didn't have to be tokyo or paris or barcelona | 17:50 |
| timburke | zaitcev: but what if we miss that left turn on the way home and never get back? | 17:51 |
| jrichli | well, at least http://ttx.re has some better arguments than the mail I had read in the past | 17:52 |
| jrichli | not that i like it ... | 17:52 |
| ahale_ | "dev kings listening to grievances in their throne hall" - I've not been to a summit since boston, is there actually a throne hall now? | 17:53 |
| clayg | notmyname: oh sorry - you didn't need to re-link i had it in scrollback - dunno why it's not in my inbox yet tho? does that information only matter to "foundation-board"? good think notmyname is like hacker news for openstack ML's | 17:53 |
| *** sgundur- has joined #openstack-swift | 17:53 | |
| *** daemontool has quit IRC | 17:53 | |
| *** sgundur has quit IRC | 17:54 | |
| clarkb | ahale_: there have been ops feedback sessions since atlanta iirc | 17:54 |
| clarkb | clayg: it is an open list but you have to subscribe or troll the archive | 17:54 |
| clayg | the rooms in the hotel in Toyko did sorta have a throne room feel to it yeah... | 17:54 |
| tdasilva | gotta write that down somewhere...could make to a slide one day: "notmyname is like hacker news for openstack ML's" ;) | 17:55 |
| clarkb | the board has a separate private list for private things iirc | 17:55 |
| redbo | That doesn't seem very four opens. | 17:56 |
| clarkb | I think it has to do with legal issues | 17:56 |
| glange | notmyname: I'm not going to the summit -- we are sending quite a few people though | 17:56 |
| clarkb | which necessarily cannot be public? | 17:56 |
| notmyname | glange: ok | 17:56 |
| clarkb | there was a whole thing about it a couple years back | 17:56 |
| clayg | 'and become true "OpenStack" contributors' <- totally a goal of mine | 17:56 |
| notmyname | yeah, the openness of the foundation ML was a big deal...right what clarkb said | 17:57 |
| notmyname | so the legal stuff about pay and stuff like that is still private, but everything else is now open (wasn't originally) | 17:57 |
| *** esker has joined #openstack-swift | 17:59 | |
| *** ChubYann has quit IRC | 18:07 | |
| *** StraubTW has quit IRC | 18:13 | |
| *** garthb has quit IRC | 18:19 | |
| *** ChubYann has joined #openstack-swift | 18:20 | |
| *** StraubTW has joined #openstack-swift | 18:20 | |
| *** nadeem has joined #openstack-swift | 18:29 | |
| *** nadeem has quit IRC | 18:31 | |
| *** haomaiwang has quit IRC | 18:35 | |
| *** daemontool has joined #openstack-swift | 18:35 | |
| *** mwheckmann has joined #openstack-swift | 18:37 | |
| *** haomaiwang has joined #openstack-swift | 18:37 | |
| *** nadeem has joined #openstack-swift | 18:37 | |
| *** garthb has joined #openstack-swift | 18:51 | |
| *** mwheckmann has quit IRC | 18:54 | |
| *** joeljwright has joined #openstack-swift | 19:00 | |
| *** ChanServ sets mode: +v joeljwright | 19:00 | |
| *** tqtran has joined #openstack-swift | 19:00 | |
| *** StraubTW has quit IRC | 19:01 | |
| acoles | joeljwright: did you see there is no meeting tonight? | 19:02 |
| joeljwright | yes, Isaw, thanks :) | 19:02 |
| joeljwright | gives me more time for other work... | 19:03 |
| joeljwright | acoles: how's your Austin prep coming? | 19:03 |
| *** rcernin has quit IRC | 19:05 | |
| *** aimeeU has quit IRC | 19:06 | |
| acoles | joeljwright: its coming along ok thanks | 19:07 |
| acoles | good night | 19:07 |
| jrichli | night! | 19:07 |
| joeljwright | night! | 19:08 |
| *** acoles is now known as acoles_ | 19:08 | |
| *** mwheckmann has joined #openstack-swift | 19:08 | |
| *** StraubTW has joined #openstack-swift | 19:08 | |
| *** joeljwright has quit IRC | 19:08 | |
| *** StraubTW has quit IRC | 19:08 | |
| *** StraubTW has joined #openstack-swift | 19:09 | |
| *** Suyi has quit IRC | 19:12 | |
| *** Suyi has joined #openstack-swift | 19:12 | |
| *** haomaiwang has quit IRC | 19:13 | |
| *** sheel has quit IRC | 19:15 | |
| *** nadeem has quit IRC | 19:27 | |
| openstackgerrit | Daisuke Morita proposed openstack/swift: Update Metadata after 'sync' op in account/container replicator https://review.openstack.org/302494 | 19:27 |
| *** haomaiwang has joined #openstack-swift | 19:27 | |
| *** nadeem has joined #openstack-swift | 19:28 | |
| *** siva_krishnan_ is now known as siva_krishnan | 19:28 | |
| openstackgerrit | Merged openstack/swift: Require account/container metadata be UTF-8 https://review.openstack.org/285754 | 19:30 |
| *** rcernin has joined #openstack-swift | 19:31 | |
| *** ametts has quit IRC | 19:33 | |
| *** haomaiwang has quit IRC | 19:33 | |
| *** haomaiwang has joined #openstack-swift | 19:36 | |
| *** haomaiwang has quit IRC | 20:02 | |
| *** mariusv has quit IRC | 20:08 | |
| *** nadeem has quit IRC | 20:11 | |
| *** mingyu has quit IRC | 20:13 | |
| *** haomaiwang has joined #openstack-swift | 20:17 | |
| *** thumpba has quit IRC | 20:23 | |
| *** NM has quit IRC | 20:32 | |
| *** geaaru has quit IRC | 20:34 | |
| *** haomaiwang has quit IRC | 20:36 | |
| *** haomaiwang has joined #openstack-swift | 20:38 | |
| *** haomaiwang has quit IRC | 20:39 | |
| *** nadeem has joined #openstack-swift | 20:42 | |
| *** asettle has joined #openstack-swift | 20:42 | |
| *** mingyu has joined #openstack-swift | 20:43 | |
| *** nadeem has quit IRC | 20:45 | |
| *** asettle has quit IRC | 20:45 | |
| *** nadeem has joined #openstack-swift | 20:46 | |
| *** mingyu has quit IRC | 20:48 | |
| *** _JZ_ has joined #openstack-swift | 20:52 | |
| *** daemontool has quit IRC | 20:56 | |
| *** esker has quit IRC | 20:56 | |
| *** m_kazuhiro has joined #openstack-swift | 20:59 | |
| notmyname | zaitcev: have you played with ceph's bluestore? based on my initial read about it, it looks pretty interesting | 21:13 |
| notmyname | granted, I've only read a blog about it... | 21:16 |
| pdardeau | i'm trying to follow-up on keystone/swift integration pain points at summit | 21:17 |
| notmyname | pdardeau: +1 | 21:18 |
| pdardeau | anyone remember the specifics? | 21:18 |
| pdardeau | i'm trying to rope in some keystone guys so that we can have some cross-project discussion on it | 21:18 |
| notmyname | pdardeau: a lot of it is around CLI/SKD integration issues. session objects, when to use which auth version and how to figure that out without 2 dozen different CLI options | 21:18 |
| pdardeau | i remember: scalability limits with keystone, keystone api versioning pain, too simplistic policy.json, desire for something like saio but for keystone | 21:19 |
| notmyname | keeping token info across requests in the CLI | 21:19 |
| *** m_kazuhiro has quit IRC | 21:19 | |
| notmyname | ah, the policy.json was a comment donagh had made about the samples, not what's possible | 21:19 |
| notmyname | but yeah, we've got the policy.json work too that hisashi is working on | 21:20 |
| pdardeau | if anyone is interested in meeting up with keystone guys at summit, please let me know and i'll try to set up informal discussion | 21:20 |
| notmyname | from a dev side, setting up a simplistic, stable keystone for testing would help with further dev/integration work from the swift side | 21:20 |
| notmyname | pdardeau: thanks | 21:21 |
| pdardeau | notmyname: thx for clarifications | 21:21 |
| notmyname | there's the "received wisdom" (which AFAIK hasn't been retested in a while) about keystone not able to handle the load from busy swift clusters | 21:22 |
| *** sbezverk has quit IRC | 21:25 | |
| *** sbezverk has joined #openstack-swift | 21:27 | |
| *** _JZ_ has quit IRC | 21:27 | |
| stevemar | notmyname: pdardeau you guys are more than welcome to crash a keystone design session or two | 21:31 |
| stevemar | keystoners don't bite :) | 21:31 |
| notmyname | stevemar: so are fernet tokens a thing everyone should be using today? or not yet? or not ever? | 21:34 |
| pdardeau | stevemar: cool. i might do that | 21:34 |
| stevemar | notmyname: they should totally be used today | 21:34 |
| stevemar | notmyname: IIRC, mfisch is using them today in production at TWC | 21:34 |
| stevemar | pretty sure rax and ibm clouds are using them too | 21:35 |
| notmyname | stevemar: cool. but he does nova. I want to hear what Vinsh says for using them with swift ;-) | 21:35 |
| stevemar | notmyname: i would hope it didn't matter what service is using it | 21:36 |
| notmyname | stevemar: with fernet, does each auth request get a new token (within some short amount of time)? IOW are they cachable or would swift see a new one every time for every request? | 21:36 |
| stevemar | the policy stuff is sadly still an issue :( | 21:36 |
| notmyname | stevemar: right. just that nova loads are generally lower than swift loads :-) | 21:36 |
| stevemar | notmyname: they would see a new one, its the way they are validated that is different | 21:37 |
| notmyname | which means that the load to keystone (req/sec) would go up when using fernet. specifically, it goes up by the cache-hit rate with uuid tokens | 21:38 |
| stevemar | with fernet you don't have to go back to keystone to validate, they are validated offline | 21:38 |
| stevemar | let me double check if they are cachable... | 21:38 |
| * stevemar looks at code | 21:38 | |
| *** CaioBrentano has quit IRC | 21:41 | |
| notmyname | stevemar: thanks for the info | 21:42 |
| Vinsh | Fernet all the way +1 | 21:43 |
| stevemar | Vinsh: \o/ | 21:43 |
| notmyname | Vinsh: cool, that's good to hear | 21:43 |
| Vinsh | been swifting and ferneting for a solid year. Transition wasn't even a thing users detected. | 21:43 |
| redbo | I don't know how applicable it is to our auth (which is only keystone-ish), but we had some problems with people whose scripts authenticate frequently, and they get new tokens every time now. | 21:44 |
| *** cdelatte has quit IRC | 21:44 | |
| * Vinsh same team as mfisch btw. | 21:45 | |
| stevemar | notmyname: so, in keystonemiddleware (keystone_authtoken), it's smart enough to cache tokens so you don't have to do that for every request | 21:45 |
| stevemar | Vinsh: nice to meet you :) | 21:45 |
| Vinsh | Fernet Tokens are cachable, some digging I did on that last week confirms. Hi stevemar :) | 21:45 |
| stevemar | Vinsh: i know mfisch, unfortunately :O | 21:45 |
| notmyname | stevemar: thanks | 21:46 |
| Vinsh | stevemar: Indeed :) | 21:46 |
| Vinsh | I thought I read in the summit talk he is doing that fernet will be the standard soon and all other tokens deprecated (in the O release). | 21:48 |
| timburke | notmyname: there's a decent start on allowing swiftclient to use sessions in patch 298968 - i even did a follow-up to add a v1 plugin in patch 300697 so swift devs could get a feel for it against tempauth/swauth | 21:49 |
| patchbot | timburke: https://review.openstack.org/#/c/298968/ - python-swiftclient - Adding keystoneauth sessions support | 21:49 |
| patchbot | timburke: https://review.openstack.org/#/c/300697/ - python-swiftclient - Add v1password keystoneauth plugin | 21:49 |
| timburke | that first patch still doesn't have us instantiating any sessions ourselves, but it'll allow other projects to pass in sessions when creating Connections | 21:51 |
| *** lakshmiS has quit IRC | 21:53 | |
| notmyname | timburke: that sounds like progress | 21:53 |
| timburke | definitely. and we can work backwards from there a bit: start instantiating our own sessions given some auth args to Connection; allow SwiftService to accept a session and safely share it across threads; have SwiftService instantiate a session from auth args... | 21:56 |
| notmyname | pdardeau: ^ ;-) | 21:56 |
| pdardeau | notmyname, timburke got it, thx! | 21:59 |
| timburke | now that we've got the optparse -> argparse transition out of the way, the ultimate goal is to get rid of most/all of our --os-* options and let keystoneauth handle them through loading.register_auth_argparse_arguments and loading.load_auth_from_argparse_arguments | 21:59 |
| *** asettle has joined #openstack-swift | 22:01 | |
| jamielennox | timburke: whilst i love people using the keystone loading for that, you might want to jump straight to os-client-config | 22:06 |
| sbezverk | +clayg ping | 22:07 |
| *** marcusvrn_ has quit IRC | 22:13 | |
| timburke | jamielennox: it's taken me this long just to wrap my head around the keystone bits. one step at a time :-) | 22:20 |
| jamielennox | timburke: :) it's great swift is looking at it because it | 22:23 |
| jamielennox | 's the last holdout, just saying you'll have mordred around pretty quick replacing it with os-c-c for the whole OS_CLOUD= thing | 22:23 |
| clayg | sbezverk: pong | 22:23 |
| *** gyee has quit IRC | 22:25 | |
| mattoliverau | morning | 22:26 |
| *** asettle has quit IRC | 22:26 | |
| *** asettle has joined #openstack-swift | 22:27 | |
| clayg | lol @ timburke | 22:28 |
| *** mingyu has joined #openstack-swift | 22:33 | |
| *** StraubTW has quit IRC | 22:33 | |
| sbezverk | +clayg I tried configuring follwoing your guidelines | 22:33 |
| sbezverk | +clayg I generated rings using new R syntax | 22:33 |
| sbezverk | I see now this message in the log, I wanted to check with you if it is expected | 22:34 |
| sbezverk | swift-object-replicator: Can't find itself 192.168.80.25 with port 6000 in ring file, not replicating | 22:34 |
| *** mwheckmann has quit IRC | 22:35 | |
| *** arch-nemesis has quit IRC | 22:36 | |
| clayg | sbezverk: no that's probably not right - what "R syntax" | 22:36 |
| *** mingyu has quit IRC | 22:37 | |
| sbezverk | +clayg object.builder add z1-192.168.80.25:6000R192.168.80.25:10873/d1 1 | 22:42 |
| sbezverk | in the doc you gave me the link, they suggested to use this format when building rings | 22:42 |
| clayg | sbezverk: ok yeah that format is fine - but you need d1_<rsync_port> for the rsync_module with the {meta} thing to work | 22:48 |
| clayg | the replication ip and port is for replication object servers - if you're not running a object-server process on 10873 that's probably not what you ment to put in the ring | 22:48 |
| clayg | r1z1-192.168.80.25:6000/d1_10873 should let you use a rsync_module template like 'rsync://{ip}:{meta}/object' or something close to that | 22:49 |
| clayg | sbezverk: ^ | 22:50 |
| *** nadeem has quit IRC | 22:52 | |
| sbezverk | +clayg so I need to rebuild ring with 6000/d1_10873 right? | 22:56 |
| clayg | sbezverk: worth a shot | 22:56 |
| clayg | i have to bounce pretty quick - if you're still fighting with it you could ask someone else here - or drop me a message and I'll probably see it later | 22:57 |
| sbezverk | +clayg sure thing, will try it, who would you suggest to ping with this? | 22:58 |
| clayg | oh idk - just ask - people like to help - except glange | 22:59 |
| clayg | glange: ;) | 22:59 |
| sbezverk | +clayg thanks | 23:05 |
| *** mingdang1 has joined #openstack-swift | 23:06 | |
| *** Suyi has quit IRC | 23:07 | |
| *** Suyi has joined #openstack-swift | 23:07 | |
| *** km__ has joined #openstack-swift | 23:08 | |
| *** km__ is now known as Guest81621 | 23:08 | |
| sbezverk | folks could you please check new rings I rebuild for a non-default rsyncd port: http://paste.openstack.org/show/494902/ | 23:16 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 23:18 |
| *** ozialien has joined #openstack-swift | 23:20 | |
| *** Lickitysplitted_ has quit IRC | 23:36 | |
| *** Lickitysplitted has joined #openstack-swift | 23:36 | |
| *** kei_yama has joined #openstack-swift | 23:36 | |
| *** thumpba has joined #openstack-swift | 23:38 | |
| zaitcev | notmyname: Sorry, I have not touched bluestore. I don't even know what that is. I know that Sage wanted to create a user-mode filesystem or store... But I thought that was called "newstore". | 23:39 |
| notmyname | zaitcev: I think that was the code name | 23:40 |
| *** lakshmiS has joined #openstack-swift | 23:41 | |
| *** _JZ_ has joined #openstack-swift | 23:44 | |
| *** asettle has quit IRC | 23:47 | |
| *** asettle has joined #openstack-swift | 23:47 | |
| *** mingyu has joined #openstack-swift | 23:48 | |
| *** mwheckmann has joined #openstack-swift | 23:50 | |
| *** jamielennox is now known as jamielennox|away | 23:51 | |
| *** sbezverk has quit IRC | 23:52 | |
| *** mingyu has quit IRC | 23:53 | |
| *** sbezverk has joined #openstack-swift | 23:56 | |
| *** rickyrem has quit IRC | 23:58 | |
| *** rickyrem has joined #openstack-swift | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!