Friday, 2015-04-24

thingeettx: slowest build...3 hours later :)00:00
*** nikhil_k is now known as nikhil_k-afk00:48
*** flaper87 has quit IRC01:27
*** flaper87 has joined #openstack-relmgr-office01:28
*** flaper87 has quit IRC01:33
*** flaper87 has joined #openstack-relmgr-office01:39
*** asalkeld is now known as asalkeld_lunch03:27
*** david-ly_ has joined #openstack-relmgr-office06:01
*** david-lyle has quit IRC06:03
ttxfungi: thx for the heads-up!06:57
ttxKiall: for Designate, we are missing... backport for , backport for , and decision whether should be fixed for RC2 or not07:15
openstackLaunchpad bug 1447683 in Designate "/v2/zones is missing the nameservers sub resource" [Critical,Fix committed] - Assigned to Graham Hayes (grahamhayes)07:15
Kiallttx - humm 2 back ports I'll look at in a couple. The extra bug, yea, I mentioned it here later last night as one last one...07:33
ttxKiall: so I should add it to RC2 ?07:36
KiallYes, please :)07:36
ttxKiall: ok doing that now07:36
KiallI'll be at a laptop in 30 mins and can sort the other 2 backports out.07:36
*** asalkeld_lunch has left #openstack-relmgr-office07:42
*** zz_johnthetubagu is now known as johnthetubaguy07:51
Kiallttx: backport of one here: and the other here
Kiallttx: also, proposal bot gave us this.. not sure anymore if we should land it?
ttxKiall: you should, otherwise first project to require designateclient might break the world07:57
KiallOkay, perfect :)07:57
ttxIt's basically to make sure all stable/kilo testing uses consistent deps07:57
KiallOn it's way in.. Thanks :)07:57
* Kiall loves this gate at time of the morning .. everything get's a test slave straight away ;)07:59
Kiallttx: Okay, last two backports passed the check queue and just need their +A :)08:19
ttxon it08:28
ttxKiall: about
openstackLaunchpad bug 1445123 in Designate kilo "Under high load, DB Connection Pool get's exhausted" [High,In progress] - Assigned to Kiall Mac Innes (kiall)08:28
ttxfixed is marked "partial"08:29
ttxcan we consider this bug fixed by it ? Or open a separate bug for what's left ?08:29
Kiallttx: ah yes, I had intended on a better fix, but it involves too much re-work so we settled for a change that fixes under all but the worst load :)08:30
ttxok, will mark bug fixed then08:30
ttxnow if only LP wouldn't timeout08:32
KiallOne day, we'll get JIRA to moan about ;)08:34
KiallAnd - All changes have landed :) Should we pull the trigger, or give it the rest of the day just in case?09:09
ttxflaper87: ping me when around10:04
ttxKiall: hojw about we pull trigger in a couple of hours if things stil look good ?10:05
*** flaper87 has quit IRC10:05
*** flaper87 has joined #openstack-relmgr-office10:05
ttxflaper87: around?10:08
flaper87ttx: I am10:09
ttxflaper87: looking intop a Zaqar RC2 to include,n,z10:09
ttxanything you'd like to add ?10:10
flaper87nope, I'll approve those 2 and we're good to go10:10
Kiallttx: sounds good to me10:10
ttxflaper87: sounds good, I'll make it happen10:11
sdaguemorning ttx, let me know if there is anything I can help with on the release or un thaw today11:34
sdaguettx: you want to unblock your requirements -2s at this point?11:38
sdagueI flushed out mine yesterday11:38
ttxsdague: ack11:40
ttxsdague: We need to decide if we announce requirement unfreeze, and also if we clear the few stable/kilo libs that wanted to release11:40
ttxAfter that I think we are good11:40
ttxflushing my -2s now11:41
ttxsdague: you still have one @
sdagueso, I think there are 2 bandit proposes11:42
sdague and
sdagueI guess I should -2 the newer one and unblock the older one11:43
sdagueit would have been nice if the security team had talked to each other about that :)11:43
ttxat least renew the -2 justification message11:43
ttxsdague, dhellmann: so I'd say that we wait for the last RC2s (should be done in a couple of hours) and announce requirements and library release unfreeze11:45
ttxso that we can get oslo.messaging and the keystone libs a kilo point release11:46
sdaguehonestly, I let a couple of things go through yesterday that should be fine for master11:46
ttxsdague: on the release front, mostly concerned with
openstackLaunchpad bug 1446583 in oslo-incubator "services no longer reliably stop in stable/kilo" [Critical,Fix committed] - Assigned to Julien Danjou (jdanjou)11:46
ttxsdague: is the final call there to work around it in grenade ?11:47
ttxor still being discussed ?11:47
ttxKiall: I'll proceed with Designate RC2 now unless you scream11:51
sdaguettx: we put a hack into stable/kilo devstack11:51
sdaguewhere we always shutdown servers with "kill; sleep 1; kill"11:52
ttxshould we kill the stable/kilo proposed backport ?11:52
sdaguewe still want that in stable/kilo11:52
sdagueit can wait until the release11:52
sdaguewe're also going to need to have something in the release notes acknowledging that kill SERVICE may not stop that service11:53
ttxok good11:53
sdaguebut testing is unblocked11:54
sdagueI want to remove the work around in devstack as soon as we can11:54
sdaguefyi - this requirements file reorg might be useful -
ttxflaper87: zaqar RC2 up @
ttxKiall: designate RC2 up @12:00
ttxthingee, david-ly_ nikhil_k: ping me when you wake up13:08
Kiallttx: excellent :)13:15
mesteryttx: A new security issue (DoS against Neutron DHCP agent) was filed 10 hours ago. I think we may need an RC3 for this:
openstackLaunchpad bug 1447883 in neutron "Restrict netmask of CIDR to avoid DHCP resync is not enough" [Undecided,In progress] - Assigned to Kevin Benton (kevinbenton)13:36
ttxmestery: itmay13:37
mesteryttx: We're goign to land the master patch and get a backport up just in case13:38
ttxfungi, sdague: updated to match new stable/* usage13:47
nikhil_k-afkttx: hi13:50
*** nikhil_k-afk is now known as nikhil_k13:51
fungittx: awesome, i meant to ask if we should. likely infra-manual needs some tweaks to remove references to "proposed"13:52
fungii'll put together a patch in a bit13:52
*** johnthetubaguy is now known as zz_johnthetubagu13:52
*** dansmith is now known as superdan13:59
ttxnikhil_k: I think we are good for RC214:03
ttxnikhil_k: all targeted bugs fixes merged14:04
nikhil_kttx: ok, how open are we for RC3 possibility if something comes up by Mon?14:04
nikhil_kttx: I understand that you have been very busy and rc3 might be a big pain14:05
nikhil_kttx: however, I wanted to set the right expectations with folks so trying to get early feedback14:05
ttxnikhil_k: we can still do a RC3 next week if something we can't fix post-release appears14:05
ttxIf it's a "regular bug" it will wait until after release14:06
nikhil_ksounds good14:06
nikhil_kttx: not a regular one, but there was some discussion that did not fan out for rc3 yet14:06
nikhil_kpan out*14:06
ttxnikhil_k: At this point we hope RC2 will be good enough. If something comes up we'll reassess14:07
ttxin all cases the thing would need to be fixed in master first14:07
nikhil_kttx: about the py-client14:07
nikhil_kcan we bump the req to 0.17.1 (I need to check if not done already)14:07
ttxnikhil_k: we'd rather not -- we'll document the recommended min version though in the release notes14:08
nikhil_kttx: this one
nikhil_kttx: can you please comment there? I seem to be the bottleneck in this communication14:09
nikhil_kthis == lib release schedule14:09
ttxI fully agree on Doug's -2 there14:09
nikhil_kStuart is pressing on it14:09
nikhil_kok, just a agree would suffice :)14:09
nikhil_kplease :)14:09
ttxcutting Glance rc2 now14:10
nikhil_kHope nothing was affected by the py-client release. Should I be aware of anything if it did?14:12
ttxnikhil_k: nope it went well14:13
ttxwe'll likely unblock the others once I'm done with RC2s14:13
ttxthingee, david-ly_ ping me when around14:13
nikhil_kttx: unblock whom/what ?14:14
nikhil_kare other RC2s waiting on Glance?14:14
ttxnikhil_k: library releases14:14
nikhil_kfor a sec, I went wat14:15
ttxglance done14:15
nikhil_kttx: .awesome.14:15
nikhil_kHave a great day!14:15
ttxyou too!14:17
*** zz_johnthetubagu is now known as johnthetubaguy14:20
ttxdhellmann: around?14:30
ttxredrobot: ping14:51
redrobotttx pong14:51
ttxredrobot: so we seem to have a hard time merging stuff14:51
redrobotttx yeah, one of our contributors is looking at the intermittent failures14:52
redrobotttx I suspect is a problem with a specific version of openssl14:52
ttxBe that or
ttxredrobot: we'll probably have to delay Barbican RC2 to Monday14:52
* ttx will try to stay away from OpenSTack this weekend14:52
redrobotttx fine by me.  I'll be poking at those failures as well.  Hopefully we can get them sorted out today.14:53
ttxredrobot: that would be swell. If a fix is necessary for stable/kilo for it to pass more reliably, you can propose it thezre. I'll pick it up and merge it on Monday morning14:53
dhellmannttx: here15:02
ttxdhellmann: we're about to unfreeze libs and requirements15:04
ttxdhellmann: and therefore closing the-big-thaw15:04
ttxdhellmann: any last thoughts ?15:04
ttxdhellmann: I think we'll do external dep capping after release15:04
ttxnot sure I'm ready to take risks there15:05
dhellmannI've had enough fun this week15:05
ttxsdague was supportive of the "unfreeze now"15:05
ttxI'd like to cut the last RC2s before15:05
ttxso waiting on david-ly_ and thingee15:05
dhellmannI think we're ready to open liberty up, and I can spend some time reviewing req changes today, although it might be better to wait for monday15:06
ttxUnfreezing doesn't mean "approving all"15:06
sdaguedhellmann: yeh, so I undid all my -2s15:06
sdaguethen did real reviews15:06
dhellmannsdague: I noticed a flurry of email from gerrit15:06
ttxIt's just thawed, not hot yet15:06
sdagueI did approve a couple that had +2 on them already that seemed safe15:06
ttxsame for lib releases. We can unblock them, but I'd encourage people to not release on Friday15:06
dhellmannyeah, let's wait to mention that they are unblocked until monday just to be safe15:07
sdagueyeh, mtreinish is asking about tempest-lib atm15:07
ttxSo I will wait for the last two RC2s, then send an email about DepFreeze end and unleash libs15:07
sdaguettx: or not send the email until monday15:07
sdaguewe can soft unfreeze today15:07
ttxsdague: well, I feel guilty of having overrun the DepFreeze mandate already15:08
ttxwas supposed to be unfrozen weeks ago. But I can see the argument15:08
ttxI can send it Monday morning15:08
thingeettx: I'll be at the office in 15 minutes to look.15:10
ttxthingee: noted15:10
dhellmannttx: I'm going afk for an errand, bbiab15:11
ttxdhellmann: wasn't nitpicking on this commit message on, was nitpicking on *that* one15:12
ttxdhellmann: sorry for the confusion in nitpicking15:12
* ttx doesn't nitpick on comit messages, only on commit messages.15:12
sdagueor comet messages15:12
dhellmannttx: ah, I wondered :-)15:13
ttxdavid-ly_: I think I'll tag Horizon RC2 in the coming 30 minutes. Scream if you don't agree15:20
david-ly_ttx: agree15:22
*** david-ly_ is now known as david-lyle15:22
ttxI'll tag now then15:23
ttxhorizon rc2 on its way15:23
ttxit's done, will announce together with Cinder15:30
ttxonce thingee arrives in office and gives me the signoff15:30
thingeettx: hi15:31
thingeettx: looking15:31
ttxFrom where I stand, all targeted things are in15:31
ttxthingee: there is
openstackLaunchpad bug 1446954 in Cinder "upload_volume_to_image --force ignores fact that a volume is already attached" [Undecided,Fix committed] - Assigned to John Griffith (john-griffith)15:32
ttxbut that could be post-released15:32
thingeettx: introduces a config opt15:33
thingeelooking closer15:34
ttxarh, my weekend is now in jeopardy15:34
thingeettx: yeah this is a valid bug.15:41
ttxsure, how critical is it though ? This also impacts doc15:41
ttxi.e. is this a regression ?15:42
thingeethis has existed for some releases back according to the commit message15:42
ttxor did we live with it happily for the last year15:42
thingeeit's all drivers except lvm that have this issue. but yeah we've been fine for a while with this15:43
ttxthis fix breaks so many freezes I would accept it only if it is critical15:44
thingeeok, it's not.15:44
thingeeit can be avoided by not passing a force flag15:44
ttxok, let's document it15:45
ttxin release notes15:45
ttxas a known issue15:45
* thingee updates release notes15:45
ttxand tag RC215:45
ttxthingee: anything else on the balance ?15:45
thingeenothing else has been raised to me. I've checked the cinder channel log as well15:46
* thingee checks new bugs one last time15:46
ttxkey word at this point is "regression" :)15:47
ttxthough I might also consider "security", "data loss" and "terrorism"15:48
thingeettx: ok LGTM15:48
thingeehurry before I change my mind! :P15:49
* thingee is just another one of the cats in the herd15:49
ttxit's tagged already15:50
ttxjust waiting on tarball now15:50
ttxI'm faster than cats15:50
thingeettx: enjoy your weekend and thanks for all your work on coordinating everything.15:52
ttxnp! Happy to leave for the weekend with a clean slate15:53
ttxlast weekend was not exactly that15:53
ttxOK, I'll call it a week and stop early for a change :)15:59
ttxTalk to you all on Monday15:59
morganfainbergttx: enjoy the weekend!16:52
*** johnthetubaguy is now known as zz_johnthetubagu18:15
*** gordc has joined #openstack-relmgr-office19:55
*** bknudson has joined #openstack-relmgr-office19:55
dhellmanngordc: so tell me a little about what you're trying for, so I understand what we're doing19:55
bknudsonsee if I can find the m/l thread...19:55
gordcrough summary is pycadf <0.7.0 is bringing in oslo.messaging as a requirement... but it should really be a test-requirement since it's not mandatory.19:57
* dhellmann waits for his browser to stop beach-balling19:57
gordci think because it's mandatory it's bringing in a bunch of other dependencies.19:57
dhellmannok, I guess dropping a requirement is less of an issue than adding one19:57
bknudsonalso, here's the keystonemiddleware change that's failing:
dhellmannhmm, how did this work before?19:59
dhellmannor has it been broken for a while?19:59
bknudsonwe just created the keystonemiddleware stable/juno branch20:00
bknudsonand then I was working through other issues with it.20:00
bknudsondhellmann: unless you were wondering about something else that worked?20:01
dhellmannbknudson: so creating the juno branch for keystonemiddleware uncovered this?20:01
bknudsondhellmann: well, the stable/juno branch never worked since it was created.20:02
dhellmanndoes this show up in other juno branches?20:02
bknudsonwho else uses pycadf?20:02
dhellmannin kilo,
bknudsonalso, keystonemiddleware has oslo.messaging in test-requirements.txt and pycadf in requirements.txt20:03
bknudsonand pip resolver seems to really have an issue when they're in separate files.20:04
dhellmannand keystonemiddleware doesn't depend on oslo.messaging?20:04
gordcdhellmann: we moved the audit middleware to keystonemiddleware.. it's optional20:05
gordcbknudson: unless there's somethign else using it.20:05
bknudsonkeystonemiddleware/tests/'oslo.messaging.get_transport', mock.MagicMock())20:05
bknudsonThat's where keystonemiddleware requirement on oslo.messaging comes from so it's in test-requirements.txt20:06
bknudsonin some ways you could say pycadf is optional since audit middleware is optional...20:06
bknudsonbut we never handled it that way20:06
bknudsonkeystonemiddleware/            self._notifier = oslo.messaging.Notifier(20:07
bknudsonwait a minute... oslo.messaging is also not just in the tests.20:07
bknudsonoh, but it's on optional import.20:07
bknudsonso it should be in test-requirements.txt20:08
bknudsonI had proposed other changes to move pycadf into test-requirements.txt or to move oslo.messaging into requirements.txt.20:09
bknudsoneither of those would also resolve the issue.20:09
bknudsonand then we wouldn't need a stable/juno branch of pycadf.20:10
gordcbknudson: i'd probably choose the former.20:10
bknudsonat least for this issue.20:10
dhellmannit's ok to create the branch, I'm just trying to make sure I understand the whole goal20:10
gordcbut i think we should probably have stable/juno for pycadf regardless of everything.20:10
dhellmannso keystonemiddleware uses pycadf and pycdf uses oslo.messaging?20:11
dhellmannbut keystonemiddleware doesn't use oslo.messaging directly?20:11
dhellmanngordc, bknudson: ^^20:12
bknudsondhellmann: keystonemiddleware doesn't use oslo.messaging directly, and actually it's explicitly an optional import.20:13
bknudsonthe only use of oslo.messaging is in the tests.20:13
bknudsonpycadf uses oslo.messaging, but apparently that's optional too.20:14
dhellmannand how about pycadf?20:14
bknudsonbut even though it's optional pycadf has it in requirements.txt20:14
bknudsonat least in that release.20:14
gordcdhellmann: it's in pycadf because we needed a placeholder for middleware when we moved it out of oslo-incubator20:14
dhellmannah, right, and where is that middleware now?20:15
gordcit's in pycadf's  test-requirements from >=0.7.020:15
gordcthe middleware is in keystonemiddleware... (and being dropped in pycadf 1.x)
dhellmannok, I think I'm caught up20:16
dhellmannso creating a stable/juno for pycadf at 0.6.0 will let you release 0.6.1 with oslo.messaging moved to test-requirements.txt20:16
dhellmannand that more accurately reflects the real requirements for the lib at that point20:16
gordcdhellmann: yep.20:17
gordcbknudson: sound good?20:17
bknudsongordc: yes.20:18
bknudsonI think that'll fix keystonemiddleware20:18
bknudsonbecause then oslo.messaging won't be installed by requirements.txt processing.20:18
bknudsononly by test-requirements.txt20:19
dhellmannbranch is done, that change will fix up the .gitreview file for you20:19
gordcdhellmann: cool cool. thanks!20:19
gordcbknudson: ok if i cut 0.6.1 on monday?20:20
dhellmannyes, please don't release anything today :-)20:20
gordcdhellmann: :) you don't like when i release things and disappear for the weekend?20:20
bknudsongordc: yep, all I should need to do is recheck when it's released.20:20
dhellmanngordc: as long as my name isn't attached, and I don't have to fix it...20:20
gordchaha cool cool20:21
gordcwill do it monday. thanks again dhellmann20:21
bknudsonif a release of a stable branch causes things to break we are really messed up.20:21
dhellmannnp, thanks for walking through it all with me to make sure I had it clear20:21
dhellmannbknudson: true, but you'd be surprised20:21
bknudsonjust don't call it 6.0.1 or something!20:21
gordcbknudson: too much pressure now :\20:22
gordcbknudson: i'll ping you monday when i've release.20:22
bknudsongordc: you going to post the review?20:22
bknudsonat least do that today then we can see if jenkins works at all.20:23
gordcbknudson: done.20:23
* gordc switches back to #os-keystone20:24
*** gordc has left #openstack-relmgr-office20:24
bknudson,n,z ?20:24
mesteryttx: This one is ready for Neutron RC3:
openstackLaunchpad bug 1447883 in OpenStack Security Advisory "Restrict netmask of CIDR to avoid DHCP resync is not enough" [Undecided,New]21:28
mesteryttx: Given it's a security issue, I think we need to role this for an RC3.21:28
*** bknudson has quit IRC21:49

Generated by 2.14.0 by Marius Gedminas - find it at!