Wednesday, 2019-10-09

« Tuesday, 2019-10-08 Index Thursday, 2019-10-10 »
openstackgerritlidong proposed openstack/heat master: Fix Firewall policy resource  https://review.opendev.org/67397300:40
*** maddtux has joined #heat01:02
*** ayoung has quit IRC01:19
*** ayoung has joined #heat01:20
*** tkajinam_ has joined #heat01:26
*** tkajinam has quit IRC01:28
*** ayoung has quit IRC02:19
*** ayoung has joined #heat02:19
*** ricolin has joined #heat02:21
openstackgerritRico Lin proposed openstack/heat master: [DMT][TEST] Check gate status againist fixes  https://review.opendev.org/68725202:28
*** ramishra has joined #heat02:37
*** k_mouza has joined #heat03:05
*** maddtux has quit IRC03:09
*** maddtux has joined #heat03:10
*** k_mouza has quit IRC03:10
openstackgerritRico Lin proposed openstack/heat master: Don't mark snapshot complete while task in progress  https://review.opendev.org/68741703:22
openstackgerritRico Lin proposed openstack/heat-tempest-plugin master: Allow retry when wc_notify not successed  https://review.opendev.org/68696203:31
*** ramishra has quit IRC03:43
*** skramaja has joined #heat04:31
*** ayoung has quit IRC04:55
*** ayoung has joined #heat04:55
ricolinNot sure what should we do with No valid host found in test jobs like https://4f8930fe3493299a9a80-c38099e6f683500395476cc7835a36a5.ssl.cf5.rackcdn.com/569582/2/gate/grenade-heat/37c8d54/logs/grenade.sh.txt.gz04:56
ricolinmaybe to give a general retry(with a enable config) when it happen?04:57
*** ramishra has joined #heat04:59
*** ayoung has quit IRC05:12
*** ayoung has joined #heat05:13
openstackgerritlidong proposed openstack/heat master: Fix Firewall policy resource  https://review.opendev.org/67397305:43
*** jtomasek has joined #heat06:50
*** jtomasek has quit IRC06:52
*** jtomasek has joined #heat07:03
*** ayoung has quit IRC07:24
*** ayoung has joined #heat07:25
*** ivve has joined #heat07:41
*** ayoung has quit IRC07:52
*** ayoung has joined #heat07:54
*** k_mouza has joined #heat08:04
*** k_mouza has quit IRC08:09
*** k_mouza has joined #heat08:10
*** tkajinam_ has quit IRC08:10
openstackgerritpengyuesheng proposed openstack/heat-cfntools master: Sync Sphinx requirement  https://review.opendev.org/68747808:21
*** ramishra_ has joined #heat08:27
*** ramishra has quit IRC08:29
*** e0ne has joined #heat08:39
*** gfidente|afk has quit IRC09:01
*** ayoung has quit IRC09:04
*** ayoung has joined #heat09:07
*** ttsiouts has joined #heat09:13
*** gfidente has joined #heat09:14
*** k_mouza has quit IRC09:26
*** k_mouza has joined #heat09:50
*** k_mouza has quit IRC09:51
*** k_mouza_ has joined #heat09:52
*** maddtux has quit IRC10:11
*** ttsiouts has quit IRC10:11
*** sharmistha has joined #heat10:16
sharmisthaHi can anyone help me in solving this issue. I have created a project and in that project created a network. Default security group was created by default.. now when i delete the stack, all the resources are getting deleted but the default security-group is not being deleted10:38
sharmisthahow can i delete it usong heat10:38
*** ayoung has quit IRC10:38
sharmistha?10:38
*** ayoung has joined #heat10:39
*** rcernin has quit IRC10:48
ricolinsharmistha, sounds like a bug of heat if Default security group was created by heat but not deleted, can you provide more information by create a story under https://storyboard.openstack.org/#!/project/openstack/heat10:54
*** sharmistha has quit IRC10:58
*** sharmistha has joined #heat11:22
sharmistharicolin: sorry got disconnected. default security group was not explicitly created by heat. Only project and networks were created using heat which resulted in creation of default security group11:22
*** k_mouza_ has quit IRC11:24
*** k_mouza has joined #heat11:25
*** gshippey has joined #heat11:29
*** ramishra_ has quit IRC11:30
*** ttsiouts has joined #heat11:32
ricolinsharmistha, implicit counts too:)11:36
ricolinsharmistha, in case you didn't see my previous message11:37
ricolin can you provide more information by create a story under https://storyboard.openstack.org/#!/project/openstack/heat11:37
*** ramishra has joined #heat11:49
sharmistharicolin: oh okay.. sure will share the scenario in a story12:08
*** ayoung has quit IRC12:12
*** ayoung has joined #heat12:15
sharmistharicolin: I have created a story for the issue: https://storyboard.openstack.org/#!/story/200669212:20
*** ttsiouts has quit IRC12:24
*** ttsiouts has joined #heat12:25
*** ttsiouts_ has joined #heat12:28
*** ttsiouts has quit IRC12:28
*** sharmistha has quit IRC12:37
*** ayoung has quit IRC12:38
*** ayoung has joined #heat12:41
ricolino/12:59
ricolinfeel free to ask anything:)13:04
zanebo/13:07
ricolinzaneb, I updated https://review.opendev.org/#/c/687417/13:10
zanebricolin: thanks for adding tests to https://review.opendev.org/687417 !13:10
ricolinlol13:10
*** ayoung has quit IRC13:12
ricolinso take a look on the new changes13:12
zanebapproved13:12
ricolinnice:)13:12
zanebwe'll want to backport that one to Train13:12
zanebsince whatever changed in Nova seems to have changed in Train13:12
ricolinzaneb, let's backport it soon master merged13:13
zaneb+113:13
zanebricolin: how do you know that wc_notify is failing?13:14
*** ayoung has joined #heat13:14
ricolinwc is failing frequently in this year, I assume there's something unstable13:16
ricolinand it looks like make the improvement https://review.opendev.org/#/c/687252/13:16
zanebI know the tests are failing, but how do you know it's because wc_notify is returning an error code?13:16
ricolinI found that part from sahara13:17
ricolinthere's similar check in sahara as well13:18
ricolinto make sure the curl return success13:18
zanebaha. anything in the heat-api logs to indicate why it was failing? got a link to one of those failures?13:19
ricolinhttps://github.com/openstack/sahara/blob/master/sahara/service/heat/templates.py#L4813:19
*** jcoufal has joined #heat13:19
ricolinzaneb, I didn't see any useful log, just think it might be such case, than search around for curl return code13:20
zanebah, ok13:20
zaneblooks like they added that in Pike13:21
ricolinwhich I think we should introduce some logs13:21
zaneb"because sometimes heat can be unavailable"13:21
ricolinKind of worry about that statement too, need to find out if that's the gate env issue or actually heat lightweight api issue13:22
ricolinjust create a story today for it https://storyboard.openstack.org/#!/story/200668513:24
dpeacocko/ sorry I'm so late :-)13:26
*** ayoung has quit IRC13:26
ricolindpeacock, o/13:26
ricolindpeacock, how's your day13:27
dpeacockricolin: So far so good - lots of bugs - the usual :-D  How about yours?13:27
*** ayoung has joined #heat13:27
*** jawad_axd has joined #heat13:28
ricolinzaneb, ramishra I think we need to approve patches in https://review.opendev.org/#/q/topic:create-train+(status:open+OR+status:merged)13:28
ricolinto make train13:28
*** bnemec has quit IRC13:29
ricolindpeacock, feel great today, it's 21:30 but still feel I can make some works:)13:30
* ricolin think maybe because he take a coffee around 18:00?13:30
dpeacockhaha13:30
*** bnemec has joined #heat13:31
zanebricolin: done, I think. let me know if there's any I missed13:32
zanebricolin: btw not sure if you've been introduced, but dpeacock is working with ramishra now13:33
zanebdpeacock: ricolin is our glorious leader13:33
dpeacockthanks for the introduction zaneb!  ricolin nice to meet you formally :-)13:33
ricolinzaneb, heat-translator need it too, but I think there's gate issue there (I can +2 on that too)13:35
ricolindpeacock, Hi!13:35
ricolinso happy to have you around! we need more hands!:)13:36
dpeacockricolin: I'm learning as fast as I can!  My plan is to be useful! :-)13:36
ricolindpeacock, that's the best plan I hard today!13:38
ricolinzaneb, I need some advice on https://storyboard.openstack.org/#!/story/200668013:39
ricolinI'm thinking maybe we can regenerate db session when that failure happens, but not sure if that's the best plan13:40
zanebwhich transaction was rolled back though? that's what I don't get13:46
*** k_mouza_ has joined #heat13:50
*** jawad_axd has quit IRC13:51
ricolinI don't know13:51
*** jawad_axd has joined #heat13:52
*** mikecmpbll has joined #heat13:53
*** k_mouza has quit IRC13:56
*** ivve has quit IRC13:56
*** ivve has joined #heat13:56
*** jawad_axd has quit IRC13:57
*** openstackstatus has quit IRC13:58
zanebricolin: best guess is that we are raising UpdateInProgress at https://opendev.org/openstack/heat/src/branch/master/heat/db/sqlalchemy/api.py#L468 then tenacity is retrying and on the retry sqlalchemy complains we can't reuse the session14:10
zanebpossssibly because we are using subtransactions=True???14:12
ricolinzaneb, might be14:17
ricolinzaneb, you mean this retry? https://github.com/openstack/heat/blob/master/heat/engine/resource.py#L41914:25
zanebyes that one14:26
ricolinzaneb, how about we move exception.UpdateInProgress out side of transaction process14:27
ricolinthat might works14:28
zanebwell, the whole point of raising it there is that we do want to rollback that transaction14:28
zanebI would have thought the with session.begin() context manager would see the exception, rollback the transaction, and leave things in a state where we can start another transaction14:29
zanebbut it appears that it may not be doing that last bit14:30
zanebwe may have to summon zzzeek14:34
*** jdwidari has joined #heat14:34
zzzeekzaneb: heat hasn't moved to the modern oslo.db enginefacade ?14:35
zanebI believe not14:35
zanebI think we tried at one point but the patch ended up being abandoned14:35
zzzeekzaneb: too bad14:36
*** ayoung has quit IRC14:36
zzzeekzaneb: so yes, if you throw an exeption it has to propagate out to the topmost begin/commit pair14:37
zanebthat makes sense, but afaict that *is* the topmost one14:37
zzzeekI assume you are getting this?  https://docs.sqlalchemy.org/en/13/faq/sessions.html#this-session-s-transaction-has-been-rolled-back-due-to-a-previous-exception-during-flush-or-similar14:37
zaneb"This session is in 'inactive' state, due to the SQL transaction being rolled back; no further SQL can be emitted within this transaction."14:38
zzzeekyep14:38
zanebat the start of what should be a new top-level transaction14:38
zzzeekso yes when that subtransactions=True flag is used, that makes it more likely to get this kind of confusion14:38
*** ayoung has joined #heat14:38
zzzeekbut also there's a "subtransaction" in the flush process14:38
zanebcan you confirm that you need to pass subtransactions=True for both the top-level and lower-level transactions? or is it only needed for lower-level ones?14:39
zzzeekzaneb: subtranscations=True is only needed when you call begin() on a Session that already had begin() called.14:39
zzzeekzaneb: it is nothing more than an "opt in" to, "i want to use nested begin/commit and I know what im getting into"14:39
zaneboh, awesome. so we could get rid of subtransactions=True there and maybe this would go away14:40
zzzeekzaneb: if you aren't nesting begin/commit pairs then absolutely14:40
zzzeekzaneb: but..you probably are14:40
zanebwe are but we know which one is at the top14:41
zaneband that's the one that seems to be causing problems14:41
zzzeekzaneb:  if you are at the top and you just called rollback() then you shouldnt be getting this error14:41
zanebI agree :)14:41
zzzeekzaneb: so...taking out the use of the flag may at least give you a heads up where the nesting is happening14:41
zzzeekzaneb: is this something you can reproduce ?14:42
zanebnot easily I don't think. it's a flaky gate test14:42
zzzeekzaneb: but it invovles a retry decorator ?14:42
ricolinwe might be able? but not easy14:42
zanebyes14:42
zzzeekbecause they make this all very crazy14:42
zzzeekthis is why i made the enginefacade14:43
zanebwe could possibly hack something to raise that exception unconditionally. that would probably reproduce it reliably14:43
zzzeekwhich people also dont' like that much but it is better14:43
zanebthe way heat-engine uses the db to store its own state is a bit different from the way e.g. nova-api uses it to just plug in some data from the user. iirc that's why it turned out not to be an easy fit14:44
zzzeekzaneb: im sure heat uses function calls which call other functions in a single greenlet or thread, and that's all enginefacade does14:45
ricolinzzzeek, jsut to make sure, session.rollback() is what you referring `rollback` above right?14:46
zzzeekzaneb: it's literally just, intesad of "session.begin(); try: foo; session.commit(); finally: session.close(), it's a decorator or context manager that does that14:46
zzzeekricolin: yes14:47
zzzeekzaneb: then enginefacade does the "nesting" thing for you.  if two functions that have @enginefacade has one calling the other, enginefacade calls begin()/commit() at the top14:47
zanebthat sounds like it should work then. I'm not sure why we gave up on the patch in that case14:49
*** ayoung has quit IRC14:49
ricolinzaneb, do you still have the link to the patch or the discussion14:49
zanebyes I keep it here on my desktop always :P14:49
* zaneb looks14:50
openstackgerritMerged openstack/heat master: Don't mark snapshot complete while task in progress  https://review.opendev.org/68741714:51
zzzeekzaneb: enginefacade turned out to be difficult to integrate in existing test suites which had to change their creational patterns14:51
*** ayoung has joined #heat14:51
zzzeekzaneb: it's my job to help you integrate it into heat, though :)14:51
zaneboh wait, apparently we merged it https://review.opendev.org/#/c/330800/14:51
zanebbut only for a handful of functions...14:52
zzzeekzaneb: intersting, in that patch I still see context.session.begin(subtransactions=True)14:52
zzzeekOK so what this patch is doing is using the newer enginefacade to start up, then it immeidately calls get_legacy_facade() and the rest of the program does everything the old way :)14:53
*** jawad_axd has joined #heat14:53
zzzeekzaneb: but you're off to a great start14:53
zaneblol14:53
zzzeek"https://docs.sqlalchemy.org/en/13/faq/sessions.html#this-session-s-transaction-has-been-rolled-back-due-to-a-previous-exception-during-flush-or-similar"14:54
zzzeekoops14:54
zzzeek"The next change will decorate the db api functions with reader and writer so queries in a function are performed in a single transaction. "14:54
* zaneb searches for the next change to see how that worked out14:55
zanebspolier: that was the last change in https://bugs.launchpad.net/heat/+bug/147972314:56
openstackLaunchpad bug 1479723 in OpenStack Heat "Replace deprecated EngineFacade from oslo_db" [High,In progress] - Assigned to Steve Baker (steve-stevebaker)14:56
ricolinsounds like we got one more task to finish14:57
zzzeekzaneb ricolin  I dont know if switching to enginefacade will solve your bug here but at least you're on the same patterns everyone else is using14:57
ricolinzzzeek, IIUC I assume to call rollback at top level might solve that bug?14:59
zzzeekricolin: that's the theory anyway14:59
zzzeekricolin: the session maintains a stack14:59
zzzeekricolin: if you close out the stack, it's happy14:59
ricolinzzzeek, thanks15:02
*** ivve has quit IRC15:03
ricolinzaneb, what else can you summon!!!15:03
zanebyou don't want to know15:04
zanebzzzeek: many thanks. we will try some stuff and keep you posted with the results15:04
zzzeekzaneb: ok, I hope we can fix this15:04
ricolinzzzeek, will keep you up to date with this15:07
*** ayoung has quit IRC15:12
*** ayoung has joined #heat15:13
*** jdwidari has quit IRC15:13
*** ramishra has quit IRC15:21
gregworkis there any notion as a cloud operator of doing a stack create where that create is done inside a target tenant?15:24
gregworkopenstack stack create -e env -t tmpl --target-tenant that stackname15:24
gregworkand it deploys said stack in their tenant, consuming their quotas15:25
gregworkas it is, im having to add myself implicitly to their project, then download a keystone openrc and source -> those commands15:25
zanebgregwork: get a token for that tenant15:25
*** jdwidari has joined #heat15:26
gregworkzaneb: i havent tried that, the docs say openstack --os-options-for-project token issue .. can a cloud operator just specify the project? or do i need a user in that project and in that case is it not the same thing i described earlier ?15:31
*** ttsiouts_ has quit IRC15:33
*** ttsiouts has joined #heat15:33
zanebgregwork: tbh I'm not sure what the current status is. if you are a global admin you may be able to get a token for any project without having a role on that project specifically. but in any case that is a question for #openstack-keystone15:33
*** skramaja has quit IRC15:35
gregworkalright, but assuming i sort that out the method is to OS_WHATEVERTOKEN=id and stack create away ?15:35
gregworkalso assuming OS_PROJECT= is correct15:35
gregworkmy goal is to build a project as a substack with OS::Heat::Stack .. then return the project id as output and create resources in that project15:37
gregworki dont know if that can be done in one stroke15:37
*** ttsiouts has quit IRC15:37
*** jdwidari has quit IRC15:42
zanebit cannot15:44
*** jdwidari has joined #heat15:48
*** jdwidari has quit IRC15:50
*** openstackgerrit has quit IRC15:52
*** ayoung has quit IRC16:01
*** ayoung has joined #heat16:02
*** ayoung has quit IRC16:09
*** ayoung has joined #heat16:10
*** ayoung has quit IRC16:14
*** ayoung has joined #heat16:17
*** openstackgerrit has joined #heat16:30
openstackgerritMerged openstack/heat master: Ignore not found when delete raw template  https://review.opendev.org/56958216:30
openstackgerritZane Bitter proposed openstack/heat master: DB: Remove unnecessary subtransactions=True  https://review.opendev.org/68761716:34
*** k_mouza has joined #heat16:36
*** ramishra has joined #heat16:37
*** k_mouza has quit IRC16:39
*** k_mouza_ has quit IRC16:40
*** bnemec has quit IRC16:45
*** bnemec has joined #heat16:47
*** jawad_ax_ has joined #heat17:02
*** mikecmpbll has quit IRC17:02
*** e0ne has quit IRC17:02
*** jawad_axd has quit IRC17:02
*** ramishra has quit IRC17:27
*** ayoung has quit IRC17:37
*** ayoung has joined #heat17:38
*** ricolin has quit IRC17:41
*** k_mouza has joined #heat18:00
*** k_mouza has quit IRC18:04
*** ayoung has quit IRC18:12
*** ayoung has joined #heat18:13
*** e0ne has joined #heat18:25
*** k_mouza has joined #heat18:40
*** k_mouza has quit IRC18:44
*** gshippey has quit IRC18:48
*** e0ne has quit IRC19:00
*** e0ne has joined #heat19:08
*** zaneb has quit IRC19:33
*** zbitter has joined #heat19:33
*** e0ne has quit IRC19:56
*** jcoufal has quit IRC20:00
*** ayoung has quit IRC20:06
*** ayoung has joined #heat20:08
openstackgerritZane Bitter proposed openstack/heat master: DB: Remove unnecessary subtransactions=True  https://review.opendev.org/68761720:31
openstackgerritZane Bitter proposed openstack/heat master: DNM try to reproduce DB session test failures  https://review.opendev.org/68767020:31
openstackgerritZane Bitter proposed openstack/heat master: Make tags handling more robust  https://review.opendev.org/68767220:41
openstackgerritMerged openstack/heat-tempest-plugin master: Allow to pass --insecure options into VMs  https://review.opendev.org/59207420:43
openstackgerritZane Bitter proposed openstack/heat master: Make tags handling more robust  https://review.opendev.org/68767220:44
openstackgerritZane Bitter proposed openstack/heat master: Make tags handling more robust  https://review.opendev.org/68767220:51
openstackgerritZane Bitter proposed openstack/heat master: Make tags handling more robust  https://review.opendev.org/68767220:52
gregworkwhat is the correct way to pass %index% from a resource group to a yaql expression?  eg. $.data.this[x] where x is %index%21:02
gregworki get an unknown #indexer error when i try passing $.data.this["%index%"]21:02
gregworkhttps://pastebin.com/mfzK9c2p21:06
gregworkmaybe some funny escaping? or maybe you cant pass %index% to yaql21:07
*** ayoung has quit IRC21:07
*** ayoung has joined #heat21:08
*** ayoung has quit IRC21:21
*** ayoung has joined #heat21:24
*** gfidente has quit IRC21:25
gregworkmaybe i am running into this: https://bugs.launchpad.net/heat/+bug/147061421:26
openstackLaunchpad bug 1470614 in OpenStack Heat "OS::Heat::ResourceGroup does not substitute index_var within intrinsics." [Medium,Triaged] - Assigned to Joe D'Andrea (joedandrea)21:26
gregwork:/21:26
*** radeks_ has quit IRC21:31
gregworki tried using str_replace on the yaql expression and %index% appears to be visible within a sub scope of str_replace21:36
gregworksigh, if there is a way to pull this off id love to hear about it or if it is a hard gap in heat, i can stop trying :)21:37
gregworks/appears/appears to not/21:37
*** ayoung has quit IRC21:43
*** ayoung has joined #heat21:44
*** rcernin has joined #heat22:14
*** flwang has joined #heat22:24
*** flwang has quit IRC22:57
*** tkajinam has joined #heat23:02
*** openstackstatus has joined #heat23:18
*** ChanServ sets mode: +v openstackstatus23:18
« Tuesday, 2019-10-08 Index Thursday, 2019-10-10 »

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!