Monday, 2016-02-08

*** amotoki has joined #openstack-release00:16
openstackgerritDavanum Srinivas (dims) proposed openstack/releases: [WIP] Oslo Releases for Week of Feb 8th
openstackgerritDavanum Srinivas (dims) proposed openstack/releases: [WIP] Oslo Releases for Week of Feb 8th
*** dims has quit IRC01:01
*** dims has joined #openstack-release01:24
openstackgerritDavanum Srinivas (dims) proposed openstack/releases: Oslo Releases for Week of Feb 8th
openstackgerritDavanum Srinivas (dims) proposed openstack/releases: [WIP] oslo.db release
*** dims has quit IRC01:46
*** dims has joined #openstack-release03:13
*** dims has quit IRC03:35
*** dims has joined #openstack-release03:42
*** dims has quit IRC04:08
*** amotoki has quit IRC04:13
*** amotoki has joined #openstack-release04:27
*** david-lyle has quit IRC04:30
*** dstanek has quit IRC04:31
*** dstanek has joined #openstack-release04:33
*** amotoki has quit IRC04:39
*** amotoki has joined #openstack-release04:40
*** david-lyle has joined #openstack-release04:45
*** amotoki has quit IRC04:56
*** amotoki has joined #openstack-release05:11
*** amotoki_ has joined #openstack-release06:08
*** amotoki has quit IRC06:11
openstackgerritAngus Lees proposed openstack/releases: Bump oslo.privsep to 1.0.0
openstackgerritMerged openstack/reno: add earliest_version option to scanner
*** amotoki_ has quit IRC08:55
*** openstackgerrit has quit IRC09:02
*** openstackgerrit has joined #openstack-release09:02
ttxdhellmann: weird, I don't even get an error09:11
*** amotoki has joined #openstack-release09:14
*** amotoki has quit IRC09:32
*** dtantsur|afk is now known as dtantsur09:58
*** dims has joined #openstack-release12:07
*** amotoki has joined #openstack-release12:13
*** dims has quit IRC12:20
*** dims has joined #openstack-release12:22
*** gordc has joined #openstack-release12:30
dimsttx : dhellmann : i had to approve dhellmann 's blocking octavia from projects.txt to get things merged over the weekend -,n,z12:33
dimsttx : dhellmann : had an emergency osprofiler release as well for a gate break for cinder -
*** amotoki has quit IRC12:35
ttxdims: ack12:38
*** krotscheck_dcm is now known as krotscheck12:39
*** amotoki has joined #openstack-release12:58
*** krotscheck has quit IRC13:06
*** amotoki has quit IRC13:15
*** krotscheck has joined #openstack-release13:18
*** dtantsur is now known as dtantsur|brb13:35
*** dtantsur|brb is now known as dtantsur14:12
*** superdan is now known as dansmith14:45
*** sigmavirus24_awa is now known as sigmavirus2415:01
dhellmannttx, fungi: regarding the release announce script, how about if I just change the default to use my email address for now to let messages go through, and then maybe we can set up an alias in the foundation email system?15:21 might be a good name for the alias15:21
fungidhellmann: i missed the error. what's happening with it now?15:24
*** amotoki has joined #openstack-release15:24
fungiare we sending from an address which fails sender callback verification?15:24
dhellmannfungi : the email goes through to the server, and is discarded somewhere along the way. I *think* because is not subscribed to the list, but when I added it to the list of senders who aren't subscribers that didn't fix it15:25
fungioh, i can try to hunt down the cause if you have a recent example release15:25
dhellmannI don't know if mailman is doing any other verification, or if the message even made it through to mailman15:25
dhellmannyeah, osprofiler, let me get the link15:25
fungii have access to mta logs on the listserv15:25
*** doug-fish has joined #openstack-release15:42
*** amotoki has quit IRC15:46
dhellmann#success is live with release history and upcoming release schedules15:51
openstackstatusdhellmann: Added success to Success page15:51
ttxnow we just need an s15:51
*** jgriffith_away is now known as jgriffith16:45
ttxdhellmann: would you have 5 minutes to discuss what's left of the RC process ?16:52
dhellmannttx: the oslo meeting is about to end, and then yes16:52
dhellmannttx: o/16:56
ttxok so here is how we used to proceed, for reference16:56
ttxshortly before RC1 we would cut a stable branch using rccut.sh16:56
ttxthat script would create the stable/$series branch and move FixComitted bugs to FixReleased16:57
ttxthen when we get the RC1 signoff we would tag the RC1 version on that release branch16:57
ttxthen when we get the RC2 signoff we would tag the RC2 version on that release branch, etc16:58
ttxThat was done using rcdelivery.sh16:58
ttxwhich would do the tagging and uploading16:58
ttxfinally when the final date came we run rcdelivery to retag the last RC as final version16:58
ttxso new process would be...16:59
ttxwe would tag RC1 from a request to releases repo16:59
dhellmannon the master branch?16:59
ttxcut a stable branch from that manually afterwards16:59
ttxon both16:59
ttxthe reason for the previous process was that you needed to push a setup.cfg bump to master17:00
ttxso you would wait for that to happen and cut the stable branch from the previous commit17:00
ttxto be sure you close the hole17:00
dhellmannyeah, you clarified that the branch is created after the tag, at the tag point17:00
dhellmannwhich was what I was really trying to understand17:00
ttxright, but now we don't do the setup.cfg bump17:00
*** jgriffith is now known as jgriffith_away17:00
ttxso we can tag and branch alright, without creating a window for pain17:01
dhellmannwe could rename to just make_stable_branch.sh17:01
ttxyeah, I was just looking into it to see if there was anything special in it17:01
dhellmannit does the work to update .gitreview too, so it's useful17:01
dhellmannit also used to verify that there was a milestone in launchpad, but I just took that out recently17:01
ttxthen RCx are requested on SHAs on the stable/$series branch17:01
dhellmannwhere x>117:02
dhellmannthe script requires a tag at the commit where the branch is going to be created17:02
ttxi think the release script would work alright, as long as we don't trigger announces (and since those are still prereleases, they shouldn't)17:02
ttxthe final retag should be the same17:03
dhellmannwe really need a test project to try all of this with17:03
ttxyeah, that could be useful17:03
dhellmannyeah, retagging is no problem with the release script17:03 did extra tricks like checking that the final is actually the same as the last RC17:04
*** jgriffith_away is now known as jgriffith17:04
ttxbut that was probably overzealous17:04
dhellmannwe could add that validation to the releases repo, if you think we need it17:04
ttxdhellmann: it was actually meant to validate the tarballs17:04
ttxi.e. check that you did not produce a corrupted one17:05
dhellmannso it didn't look at the tags, but the contents?17:05
ttxyes, it compared the tar contents17:05
ttxso it's a bit harder17:05
dhellmannok, so that's happening later in the process and -- right17:05
dhellmannnot impossible, but not something we're going to automate quickly right now17:06
ttxso it looks like we don't really need the rc* scripts that much anymore17:06
dhellmannyeah, all of this consistency is removing special case scripts :-)17:06
ttxAfter all the tagging we used to run to build a unique release page in LP, but we won't do that either17:08
ttxwe'll just check that reno produced a magic unified release notes doc17:08
ttxok, so it looks like we have everything we need17:09
ttxtwo things potentially left:17:09
ttx- rename the branching script to something more generic17:09
ttx- run a live test with something other than openstack/glance post-FF17:09
ttxi.e. some test repo we would run pre-FF17:10
ttxas part of the "Pre-final-run sprint" maybe17:10
ttxat R-617:11
ttxIt's a bit tricky though, since we want jobs defined and all17:11
openstackgerritDoug Hellmann proposed openstack-infra/release-tools: rename tag for making stable branches
dhellmann^^ step 117:12
dhellmannttx: yeah, we should make a little skeleton project like openstackreleasetest and set it up with all the same jobs as the other projects17:12
ttxmaybe we could leverage openstack-dev/sandbox17:13
dhellmannwe can use it to test the machinery, then remove the deliverables file from the releases repo17:13
dhellmanndo we ever commit changes there? we'll want to be able to do things like add reno notes, etc., too, I suspect17:13
*** dims has quit IRC17:14
ttxhmm, yeah, probably simpler to setup something else17:14
ttxwe just don't delete Gerrit repos so it could look funny17:14
dhellmannwell, we can keep using it as we mess with the tools17:14
dhellmannwhen we get to the point where we have more of this automated, it will be useful to be able to test, so I expect it to be useful for a while17:15
ttxok, good idea then17:15
dhellmanndo you want to set that up, or should I?17:16
ttxdhellmann: please do -- I'll be traveling this week and then in vacation next week, so not a lot of time to push it through17:16
dhellmannok, I'll add it to my list17:17
ttxworst case scenario we'll push it at start of the sprint week17:17
dhellmannfor the name, do you like release-test?17:17
ttxsounds good17:17
ttxdhellmann: on we noted "After Mitaka 1 or 2, re-consider the release candidate process to see if we need to be very strict with a deadline or if we can allow project teams to have some flexibility"17:19
ttxI don't remember what we had in mind there17:19
dhellmannI think we were worried about whether folks would submit requests on time17:19
ttxwas it about the rc1 deadline ?17:19
ttxwe always had some flexibility there17:20
dhellmannI think most of the teams have the process down, and the tools are working, so I think it's safe to keep doing what we're doing17:20
dhellmannyeah, we wrote that around the same time we said we would stop having meetings and office hours, and not chase down release liaisons17:20
ttxthe only trick is to make sure they do at least a rc1 so we can fallback to that. But then it's not very different from making sure we at least have an intermediary release before end of cycle17:21
ttxso there will be some chasing to do on managed projects17:21
ttxbut less than usual17:21
dhellmannwe should probably send a process email separate from the countdown emails so folks know what we expect. do you have time to start drafting that?17:21
ttxthat will serve as a reminder for intermediary released projects that haven't done one already17:23
dhellmanngood, thanks17:23
dhellmannfungi : did you find anything in the logs to explain the delivery issues?17:28
fungidhellmann: haven't looked yet, just about worked my way down to it now17:28
dhellmannfungi : ack, just circling back myself17:29
fungifrom the job log: data: (250, 'OK id=1aS5Kq-0007AT-MO')17:30
fungiso i should be able to find 1aS5Kq-0007AT-MO in the exim mainlog17:30
openstackgerritDoug Hellmann proposed openstack/releases: add a flag for whether to report the PyPI link for releases
*** dims has joined #openstack-release17:32
openstackgerritDoug Hellmann proposed openstack/releases: skeleton version of newton schedule
openstackgerritDoug Hellmann proposed openstack/releases: fill in the rest of the newton schedule
fungi2016-02-06 16:03:44 1aS5Kq-0007AT-MO <= H=([]) [2001:4802:7801:103:be76:4eff:fe20:897b] P=esmtp S=100217:57
fungi2016-02-06 16:03:44 1aS5Kq-0007AT-MO => openstack-dev <> R=mailman_router T=mailman_transport17:57
fungi2016-02-06 16:03:44 1aS5Kq-0007AT-MO Completed17:57
fungiso exim inbound on the server handed it off to mailman fine17:57
dimsfungi : looking at why release emails didn't get out? (osprofiler one over the weekend did not make it as well)17:58
fungidims: that's the one i'm looking at17:58
dimsfungi : ah thanks :)17:59
*** sigmavirus24 is now known as sigmavirus24_awa17:59
fungiso hunting within mailman's logs now to see what it did with the message once it got it17:59
*** doug-fish has quit IRC18:08
*** doug-fish has joined #openstack-release18:09
*** doug-fish has quit IRC18:13
*** dtantsur is now known as dtantsur|afk18:28
fungidhellmann: dims: so with jeblair's help i was finally able to track it down to the osprofiler release announcement getting rejected because was not subscribed to openstack-dev18:46
dhellmannfungi : oh, right, I updated openstack-announce but not openstack-dev18:48
* dhellmann slaps forehead18:48
dhellmannfungi : do you have admin rights on openstack-dev to add to the list of allowed senders?18:49
fungidhellmann: i have admin rights transitively by being able to reset the admin password as a sysadmin on the server, but i am not a list admin no18:49
fungiyou want one of the people listed on the bottom of the listinfo page for the -dev ml18:50
dhellmannfungi : ok, perhaps ttx does18:50
fungidhellmann: yep, looks like just ttx at the moment18:50
dhellmannfungi : so that's ttx18:50
fungilet's just keep spamming his irc highlight trigger18:51
fungihe loves that, i'm sure18:51
dhellmannI'll email him, too18:51
dimsdhellmann : ah thanks!18:52
dhellmannyeah, I suspected that was the problem but didn't look closely enough at which list that announcement was going to so "fixed" it in the wrong place18:53
dhellmannthanks to fungi's fresh eyes he got to the bottom of it18:53
*** sigmavirus24_awa is now known as sigmavirus2419:02
*** doug-fish has joined #openstack-release19:16
*** jgriffith is now known as jgriffith_away19:16
*** doug-fish has quit IRC19:21
*** doug-fish has joined #openstack-release19:22
dimsawesome thanks fungi19:43
fungiif nothing else, i'm good at staring at stuff and repeating what i see19:45
dimsdhellmann : i can take care of oslo.config -
dimslooks like mriedem is not around today, will need eyes on the big review as well dhellmann
dhellmanndims : I'll take a look in a few minutes20:06
dimsdhellmann : ack no rush, just lining it up :)20:07
*** jgriffith_away is now known as jgriffith20:13
*** rjaiswal has joined #openstack-release20:27
openstackgerritDoug Hellmann proposed openstack-infra/release-tools: rename script for making stable branches
*** dims_ has joined #openstack-release20:51
*** dims has quit IRC20:52
openstackgerritMerged openstack/releases: release oslo.config 3.5.0
dims_fungi : this looks like it succeeded too
*** onovy has joined #openstack-release21:20
onovydhellmann: i'm here21:20
dhellmannonovy : I'm reading your comment on
dhellmannso you're saying that the old default backend has some copyright issues, but the new default backend will not?21:21
onovymaybe have copyright issues21:21
dhellmannis jerasure part of some other library?21:21
onovyit's not part of other library, just dynamically linked21:21
dhellmannok, so you're using jerasure through PyECLib now?21:21
dhellmannfwiw, I think this is all just fine, I merely want to make sure I understand21:22
onovydefault is "don't use EC lib"21:22
onovybut example configuration is "use jerasure"21:22
onovyin liberty21:22
onovyand unit tests have hardcoded check for jerasure in liberty21:22
onovy(already fixed in master/mitaka)21:22
dhellmannand for mitaka you want to make that liberasure instead?21:23
dhellmannthe example configuration, I mean21:23
dhellmannonovy : +2, thanks for clarifying that21:24
*** dims has joined #openstack-release21:26
*** dims_ has quit IRC21:29
openstackgerritDoug Hellmann proposed openstack/reno: add release note for earliest-version feature
dhellmannttx, dims: I'm working on an email to announce releases.o.o. I could use some input, it feels a bit light on details.
dimsdhellmann : we could use this email to highlight how the tie in with the governance/ repo, tools in release-tools/ etc to help spread some knowledge22:09
dimsdhellmann : i added some sentences at the bottom, please feel free to add/modify if you feel appropriate22:09
*** jgriffith is now known as jgriffith_away22:10
dhellmanndims : I wonder if we should remove the stuff about keeping it up to date, to focus on the audience of consumers of the releases? I think I got a little off track there at line 7.22:11
dimsdhellmann : i think it's ok for now as we have to build up more information by the time we hit Mitaka, then it will be more focused towards consumers of releases22:14
dimsback in a little bit.22:15
*** jgriffith_away is now known as jgriffith22:17
dhellmanndims, ttx: I missed adding a release note to the last reno change:
*** dims has quit IRC22:31
*** doug-fish has quit IRC22:41
*** doug-fish has joined #openstack-release22:44
*** doug-fis_ has joined #openstack-release22:46
*** doug-fish has quit IRC22:48
*** doug-fis_ has quit IRC22:50
*** dims_ has joined #openstack-release22:51
dims_dhellmann : +122:52
*** doug-fish has joined #openstack-release22:59
*** doug-fish has quit IRC23:03
*** doug-fish has joined #openstack-release23:04
*** sigmavirus24 is now known as sigmavirus24_awa23:04
*** doug-fish has quit IRC23:14
*** doug-fish has joined #openstack-release23:14
*** dims_ has quit IRC23:14
*** doug-fish has quit IRC23:15
*** doug-fish has joined #openstack-release23:15
*** dims_ has joined #openstack-release23:19
*** doug-fish has quit IRC23:23
*** gordc has quit IRC23:24
*** doug-fish has joined #openstack-release23:26
*** doug-fish has quit IRC23:27
*** doug-fish has joined #openstack-release23:28
*** doug-fish has quit IRC23:38
*** doug-fish has joined #openstack-release23:40
*** doug-fish has quit IRC23:46
*** doug-fish has joined #openstack-release23:50
*** doug-fish has quit IRC23:50
*** doug-fish has joined #openstack-release23:50
*** doug-fish has quit IRC23:54

Generated by 2.14.0 by Marius Gedminas - find it at!