| *** sgordon has quit IRC | 00:00 | |
| *** SnowDust has quit IRC | 00:02 | |
| *** AndrewMichalik has quit IRC | 00:06 | |
| *** brint has quit IRC | 00:09 | |
| *** sabeen1 has quit IRC | 00:10 | |
| *** IlyaE has joined #heat | 00:12 | |
| *** pablosan has quit IRC | 00:24 | |
| *** agileclipse has quit IRC | 00:25 | |
| *** zaneb has quit IRC | 00:30 | |
| *** gokrokve has joined #heat | 00:36 | |
| *** dims_ is now known as dims | 00:37 | |
| *** gokrokve has quit IRC | 00:40 | |
| *** gokrokve has joined #heat | 00:41 | |
| *** pvaneck has quit IRC | 00:41 | |
| *** brint has joined #heat | 00:41 | |
| *** gokrokve_ has joined #heat | 00:42 | |
| *** gokrokve has quit IRC | 00:46 | |
| *** brint_ has joined #heat | 00:57 | |
| sdake2 | SpamapS - yes a piece of data that I'm averse to unified agents ;-) | 00:58 |
|---|---|---|
| russellb | stevebaker: added some info to https://bugs.launchpad.net/heat/+bug/1261433 | 00:58 |
| uvirtbot | Launchpad bug 1261433 in heat "heat-tempest-xx already has an action (DELETE) in progress" [High,Triaged] | 00:58 |
| russellb | stevebaker: happened 39 times in the last 3 days | 00:58 |
| stevebaker | russellb: OK, thanks. jasond is looking into the followup fix | 00:59 |
| russellb | and just put in an elastic-recheck query for it | 01:00 |
| russellb | thanks for being on top of it :) | 01:00 |
| stevebaker | no problem | 01:01 |
| *** brint has quit IRC | 01:03 | |
| *** brint_ is now known as brint | 01:03 | |
| *** swygue has joined #heat | 01:07 | |
| *** rpothier_ has joined #heat | 01:08 | |
| *** agileclipse has joined #heat | 01:13 | |
| *** IlyaE has quit IRC | 01:15 | |
| *** agileclipse has quit IRC | 01:15 | |
| *** fandikurnia01 has quit IRC | 01:17 | |
| *** rwsu has quit IRC | 01:18 | |
| *** rwsu has joined #heat | 01:20 | |
| openstackgerrit | Richard Lee proposed a change to openstack/heat: Add filter and pagination to stack_get_all https://review.openstack.org/63039 | 01:20 |
| openstackgerrit | Anderson Mesquita proposed a change to openstack/heat: Unscoped List Stacks https://review.openstack.org/63041 | 01:25 |
| *** brint has quit IRC | 01:26 | |
| *** brint has joined #heat | 01:26 | |
| *** arbylee has quit IRC | 01:35 | |
| *** arbylee has joined #heat | 01:36 | |
| *** nosnos has joined #heat | 01:38 | |
| *** andersonvom has quit IRC | 01:38 | |
| *** andersonvom has joined #heat | 01:38 | |
| *** arbylee has quit IRC | 01:40 | |
| *** faramir1 has joined #heat | 01:41 | |
| *** andersonvom has quit IRC | 01:42 | |
| *** harlowja has quit IRC | 01:51 | |
| openstackgerrit | A change was merged to openstack/heat: Convert [Cinder]VolumeAttachment to new Schema format https://review.openstack.org/61783 | 01:58 |
| *** andersonvom has joined #heat | 02:03 | |
| *** harlowja has joined #heat | 02:03 | |
| *** slong has quit IRC | 02:07 | |
| *** slong has joined #heat | 02:10 | |
| *** rongze has joined #heat | 02:12 | |
| *** arbylee has joined #heat | 02:28 | |
| *** nosnos_ has joined #heat | 02:30 | |
| openstackgerrit | huangtianhua proposed a change to openstack/heat: Fix copy/paste errors in API docstrings https://review.openstack.org/62536 | 02:32 |
| *** nosnos has quit IRC | 02:33 | |
| *** erkules_ has joined #heat | 02:40 | |
| SpamapS | sdake2: I think if you have that much experience with them failing then you have quite a bit of context that many of us do not. :) | 02:40 |
| *** erkules has quit IRC | 02:40 | |
| *** yogesh has quit IRC | 02:43 | |
| sdake2 | security is a big concern | 02:46 |
| sdake2 | imagine a monolithic agent deployed on hundreds of thousands of openstack vms where an exploit is found in some way | 02:47 |
| sdake2 | that would be very very verrry bad for openstack | 02:47 |
| SpamapS | sdake2: thats why the agent has to be dumb and preferrably _NOT_ root. | 02:47 |
| sdake2 | i get that cves are a fact of life, but restarting something like nova is different then restarting hundreds of thousands of vms | 02:47 |
| sdake2 | well ideally the agent wouldn't run as root, but some services need enhanced capabilities | 02:48 |
| *** nati_ueno has quit IRC | 02:48 | |
| sdake2 | agree agent has to be dumb | 02:49 |
| sdake2 | dead simple | 02:49 |
| sdake2 | do one thing do it well | 02:49 |
| SpamapS | This agent is specifically meant to not run as root. | 02:49 |
| sdake2 | don't import alot of third party dependencies as that could increase the attack surface | 02:49 |
| SpamapS | And to not do anything except the narrow things that each project needs it to. | 02:49 |
| sdake2 | yes, although i think it is a bit preferrable that they run in separate process boundaries for security concerns | 02:50 |
| sdake2 | so that has been one reason a unified agent has failed | 02:50 |
| SpamapS | Like Trove needs to run SQL commands, maybe upload to swift. Savanna just wants to execute hadoop commands. | 02:50 |
| sdake2 | (security wasn't job #1) | 02:50 |
| sdake2 | dependency management is important | 02:50 |
| sdake2 | the more dependencies, the less likely someone is to import the agent into the vm | 02:50 |
| sdake2 | that has derailed a different unified agent | 02:51 |
| SpamapS | Thats where I see the salt idea as better than write our own. Salt is just a framework that has a lot of system management plugins. You can run it without any of those. | 02:51 |
| sdake2 | salt introduces a new attack surface | 02:51 |
| sdake2 | especially the "glue" partt of things | 02:51 |
| SpamapS | so does our own comms library | 02:51 |
| sdake2 | so just listing out what has failed, not debating alternatives at this point :) | 02:52 |
| sdake2 | people try to turn unified agents into sea salt makers | 02:52 |
| sdake2 | i've seen it happen on several occaisons | 02:52 |
| sdake2 | its natural for developers to want to grow the agent to do more things | 02:53 |
| sdake2 | once that happens, the ocean boiling begins | 02:53 |
| sdake2 | as you can see from our long drawn out requirements thread | 02:53 |
| sdake2 | server dependencies have been a particular problem in the past, because they have to be securified, ha-ified, documented, etc | 02:53 |
| SpamapS | Yeah I've been trying to steer it into "do not invent anything" and "do not pile anything else on this" | 02:53 |
| sdake2 | you can steer, but it seems like the thread has gotten out of hand | 02:54 |
| sdake2 | reference sea salt making | 02:54 |
| lifeless | SpamapS: what did you think of my troll^Widea? | 02:56 |
| sdake2 | agents turn into a huge political gang bang | 02:56 |
| sdake2 | i've seen that happen on multiple occasions | 02:56 |
| sdake2 | everyone thinks they run the show | 02:56 |
| sdake2 | when in fact nobody is actually running anything ;-) | 02:57 |
| sdake2 | i could go on | 02:57 |
| sdake2 | if yo uwant me to bore you more ;-) | 02:57 |
| sdake2 | memory footprint, always a favorite topic of everyone | 02:58 |
| SpamapS | lifeless: well Dmitry seems to be the one most interested in putting actual time in, and he is not interested in the salt idea, so I steered him toward your idea and he's doing that now. | 02:58 |
| sdake2 | python is really not a suitable choice for an agent, because of its memory footprint and the fact that it shaves memory off the top of the system *always8 | 02:58 |
| lifeless | SpamapS: ok cool | 02:58 |
| sdake2 | c agent = language war | 02:59 |
| SpamapS | sdake2: that is one reason we've focused on fork/execcing the actual thing. os-collect-config is pretty small. | 02:59 |
| sdake2 | so assuming you have a 512mb vm, and you have a 20mb agent, how much memory does it cost, if you have 1 million vms | 02:59 |
| SpamapS | sdake2: that is why they wrote a C++ agent. | 03:00 |
| SpamapS | which is an easier language to sell than C | 03:00 |
| SpamapS | middle ground | 03:00 |
| sdake2 | c++ doesn't fix that problem, see my most recent post comparing c++ to c re memory consumption | 03:00 |
| sdake2 | c++ has same footprint as python memory wise | 03:00 |
| SpamapS | That was a really unfair comparison. | 03:00 |
| sdake2 | seemed fair enough to me ;-) | 03:01 |
| SpamapS | runtime != actual running footprint. | 03:01 |
| lifeless | clearly we should use forth | 03:01 |
| SpamapS | lifeless: go .. just say it. We're doing it in go. | 03:01 |
| SpamapS | I bet go's memory footprint ends up identical to python | 03:01 |
| sdake2 | we saw what happened with go + juju ;-) | 03:01 |
| sdake2 | c is a nonstarter for an agent in openstack imnsho | 03:02 |
| sdake2 | c++ is also a nonstarter for similar reasons | 03:02 |
| SpamapS | you know.. reading salt minion.. | 03:02 |
| SpamapS | it is pretty self contained | 03:02 |
| lifeless | what you need is zeromq to have a process starting interface | 03:03 |
| lifeless | :P | 03:03 |
| sdake2 | dbus ;-) | 03:03 |
| lifeless | more seriously | 03:03 |
| lifeless | memory footprint isn't an issue | 03:03 |
| lifeless | /active/ memory footprint is | 03:03 |
| *** liang has joined #heat | 03:03 | |
| sdake2 | agree active | 03:03 |
| sdake2 | eg, res in top | 03:03 |
| lifeless | the problem with Python, go, and many other languages | 03:03 |
| lifeless | is that they have a mark and sweep gc | 03:04 |
| lifeless | which makes memory == active memory on the order of seconds to minutes | 03:04 |
| lifeless | you can either drive foot print down to drive active down | 03:04 |
| lifeless | or change languages to ones with only refcounting / manual memory management | 03:04 |
| lifeless | if you're in such constrained circumstances | 03:04 |
| sdake2 | there are so many competing requirements for a unified agent | 03:05 |
| sdake2 | and tbh, I don't really see the projects that need such agent have enough use cases to deal with to make something reasonable | 03:05 |
| sdake2 | focusing on just the specific problems they want solved makes alot of sense | 03:06 |
| sdake2 | yet, I have seen that fail - google matahari for an example | 03:06 |
| *** vkozhukalov has joined #heat | 03:06 | |
| sdake2 | pretty much the only cats that ever managed to figure it out where the snmp guys in the 70s | 03:07 |
| sdake2 | since then, snakeeyes | 03:07 |
| sdake2 | cim is a complete and utter failure | 03:07 |
| sdake2 | microsoft has unified agents, not sure how they managed to make it happen | 03:08 |
| sdake2 | but they were one company with alot of top/down hierarchial control | 03:08 |
| sdake2 | I hate to say it can't be done, because anything that can be thought up can be done | 03:08 |
| sdake2 | but I wont be the one doing it ;-) | 03:08 |
| sdake2 | i do like the idea of injecting agents via oslo.bootstrap or some other process | 03:09 |
| sdake2 | this solves the distribution problem | 03:09 |
| *** swygue has quit IRC | 03:11 | |
| sdake2 | i would say I have seen every strategy, tactic, plan, project management, organizatoinal management approach to a unified agent | 03:12 |
| sdake2 | and every one of them has failed for one reason or another | 03:12 |
| sdake2 | my general thought is if *those* army of people couldn't figure it out.. | 03:12 |
| sdake2 | well you know the rest | 03:12 |
| sdake2 | ok done ranting ;-) | 03:12 |
| sdake2 | but then again, many people didn't think heat could be done | 03:14 |
| sdake2 | and here we are ;-) | 03:14 |
| sdake2 | so its entirely possible I am the wrong person to ask for input since I am corrupted by experience | 03:14 |
| sdake2 | maybe a naive mind is just what a unified agent needs to get rolling | 03:21 |
| *** jergerber has joined #heat | 03:27 | |
| *** arbylee has quit IRC | 03:28 | |
| *** jergerber has quit IRC | 03:30 | |
| *** nati_ueno has joined #heat | 03:35 | |
| *** jergerber has joined #heat | 03:37 | |
| *** swygue has joined #heat | 03:44 | |
| *** yogesh has joined #heat | 03:54 | |
| *** Alienyyg has quit IRC | 03:55 | |
| *** yogesh has quit IRC | 03:58 | |
| *** nati_ueno has quit IRC | 04:01 | |
| SpamapS | lifeless: people want something "better than HTTP polling" .. which is why the zeromq/oslo.messaging conversation happens. | 04:01 |
| SpamapS | sdake2: btw | 04:03 |
| SpamapS | root 1199 0.1 0.6 20520 13316 ? Ss Dec18 0:29 /usr/bin/python /usr/local/bin/os-collect-config | 04:03 |
| asalkeld | Web Socket transport? | 04:04 |
| grncdr | seriously what's wrong with HTTP polling? | 04:04 |
| SpamapS | grncdr: useless chatter | 04:04 |
| asalkeld | https://github.com/FlaPer87/marconi-websocket | 04:05 |
| asalkeld | marconi are thinking about web socket | 04:05 |
| grncdr | hm, I should probably read this thread that sdake has been ranting about all day... | 04:05 |
| SpamapS | but, don't fucking use it in production. Not yet! ;) | 04:05 |
| SpamapS | ossum | 04:05 |
| asalkeld | :) he is just prototyping it | 04:06 |
| SpamapS | grncdr: to tell you the truth, I think it is the right way to go | 04:06 |
| grncdr | what is? | 04:06 |
| SpamapS | But there is that issue that both sides have to be expecting that it might get chopped off in the middle because web requests aren't supposed to take a long time. | 04:06 |
| SpamapS | grncdr: websocket | 04:07 |
| SpamapS | Somebody suggested that it is only popular because it works in browsers without modification but I think it is also just a ubiquitous and thus likely to work perversion of HTTP. | 04:07 |
| grncdr | I agree with somebody ;) | 04:08 |
| grncdr | but I have no context for or expertise with wtf you guys are talking about :) | 04:08 |
| *** blomquisg has quit IRC | 04:08 | |
| *** guohliu has joined #heat | 04:11 | |
| SpamapS | grncdr: If you are going to put an agent on every DBaaS node you spin up and every Hadoop aaS node you spin up.. useless chatter becomes a measurable factor in cost calculation. | 04:13 |
| grncdr | hm, so I'll try not to put my foot in my mouth too much, while still responding... ;) | 04:15 |
| SpamapS | grncdr: it relates to Heat in that we'd also like to have a way for things to contact Heat from private networks. I'm less convinced that one code base will do more than the private network traversal though. | 04:16 |
| grncdr | so, regarding HTTP polling vs. websockets vs. some other protocol over TCP sockets | 04:17 |
| grncdr | HTTP polling works everywhere and has easy semantics, and is not _necessarily_ much chattier than websockets if you use long-polling | 04:18 |
| grncdr | if long-polling is *still* too chatty, then drop HTTP altogether | 04:19 |
| SpamapS | right that is how I feel too | 04:20 |
| SpamapS | and if we just did that to the ec2 metadata service... | 04:20 |
| grncdr | I don't know, I'm not super passionate about it, but it doesn't seem reasonable to use websockets outside of browsers when there's a plethora of well-implemented and much lighter-weight text-based protocols out there | 04:21 |
| SpamapS | I'm starting to think that is the "it is already done" answer and we would just need to enable POST and help the long polling work better. | 04:21 |
| SpamapS | grncdr: Consider a deployment where users want to deploy a hadoop cluster in a private network, one where only one machine even has a default gateway. | 04:22 |
| grncdr | SpamapS: you may have to dumb it down/explain why that matters… | 04:23 |
| SpamapS | grncdr: Right now, in order to make the EC2 metadata service work for those networks, we let neutron sit on that network and listen to 169.254.169.254 port 80. It then looks up the source IP, finds the instance id, and fetches the metadata from nova. | 04:23 |
| SpamapS | grncdr: HTTP makes that exceedingly straight forward. | 04:24 |
| *** AndrewMichalik has joined #heat | 04:25 | |
| grncdr | hm, so I feel like there should be a generic way to (selectively) expose OpenStack HTTP services inside a private network | 04:26 |
| grncdr | like, there should be a neutron resource that does just that (minus the IP lookup part I guess) | 04:27 |
| grncdr | I guess you could do it by configuring the router… but honestly I'm talking out my ass here ;) | 04:27 |
| *** ccs has quit IRC | 04:28 | |
| *** gokrokve_ has quit IRC | 04:28 | |
| openstackgerrit | A change was merged to openstack/heat: Convert AWS network resources to new Schema format https://review.openstack.org/61818 | 04:29 |
| SpamapS | grncdr: thats the thing, even just IP is difficult at this level. | 04:29 |
| *** Tross has joined #heat | 04:30 | |
| SpamapS | grncdr: because the idea is that these nodes have precious data that you don't want to leak out. They are then contacted and filtered by a head node that pokes the data out in limited fashion. | 04:30 |
| *** ccs has joined #heat | 04:32 | |
| *** gokrokve has joined #heat | 04:36 | |
| *** AndrewMichalik has quit IRC | 04:37 | |
| *** rcleere has joined #heat | 04:41 | |
| *** AndrewMichalik has joined #heat | 04:44 | |
| *** rongze has quit IRC | 04:44 | |
| *** _skraynev has joined #heat | 04:44 | |
| *** nati_ueno has joined #heat | 04:46 | |
| *** Tross has quit IRC | 04:52 | |
| *** ccs has quit IRC | 04:52 | |
| *** akuznetsov has joined #heat | 04:54 | |
| _skraynev | Morning | 04:59 |
| lifeless | SpamapS: so don't poll :) | 05:01 |
| stevebaker | SpamapS: how about allowing either the neutron proxy or nova metadata server to have other endpoints that it proxies to? | 05:01 |
| lifeless | so the way the nova metadata server works | 05:01 |
| lifeless | with neutron | 05:02 |
| stevebaker | then we can implement longpoll | 05:02 |
| lifeless | is that there is an agent that lives into the user network namespace and proxies out to the API endpoint | 05:02 |
| _skraynev | stevebaker: hello, will you be here 30 min yet? I want discuss with you tempest, but from mobile it is dificult? | 05:02 |
| sdake2 | grncdr i wasn't ranting all day, it was more like a 1 hour get it off my chest session ;-) | 05:02 |
| *** chandankumar has joined #heat | 05:02 | |
| stevebaker | _skraynev: I will be here ... but ... wine | 05:03 |
| _skraynev | stevebaker: it is ok) | 05:04 |
| stevebaker | I'll probably make sense, mostly | 05:04 |
| stevebaker | lifeless: so the agent could proxy to other endpoints too? | 05:04 |
| *** rcleere has quit IRC | 05:06 | |
| *** nosnos_ has quit IRC | 05:06 | |
| lifeless | stevebaker: in principle yes, but really I'd be inclined to push it all through the one metadata API | 05:06 |
| lifeless | stevebaker: if that API needs to be richer, surely we just fix it there. | 05:07 |
| *** nosnos has joined #heat | 05:07 | |
| *** urulama has joined #heat | 05:07 | |
| stevebaker | lifeless: it would be nice if heat didn't have to push changes to nova metadata when something changed, that is where I'm coming from | 05:08 |
| *** rushiagr has joined #heat | 05:08 | |
| *** _skraynev has quit IRC | 05:08 | |
| *** _skraynev has joined #heat | 05:08 | |
| *** Seal has joined #heat | 05:10 | |
| *** rushiagr2 has joined #heat | 05:12 | |
| *** rushiagr has quit IRC | 05:16 | |
| *** Tross has joined #heat | 05:16 | |
| *** AndrewMichalik has quit IRC | 05:17 | |
| *** rongze has joined #heat | 05:18 | |
| *** _skraynev has quit IRC | 05:19 | |
| *** faramir1 has quit IRC | 05:21 | |
| *** rongze has quit IRC | 05:23 | |
| *** sdake2 has quit IRC | 05:28 | |
| *** radix_ has quit IRC | 05:30 | |
| asalkeld | woot, lifeless your data reproduces the bug | 05:31 |
| asalkeld | ERROR 1005 (HY000) at line 47: Can't create table 'test_migrations_utils.stack' (errno: 150) | 05:31 |
| *** rushiagr2 is now known as rushiagr | 05:40 | |
| *** harlowja is now known as harlowja_away | 05:40 | |
| *** erkules_ is now known as erkules | 05:45 | |
| skraynev | stevebaker: So now I am ready) As far as I remember you talk that tempest test will became available after release F20. Currently I want know : what plans do you have now? | 05:45 |
| *** Seal has quit IRC | 05:46 | |
| *** harlowja_away is now known as harlowja | 05:59 | |
| *** jergerber has quit IRC | 06:02 | |
| *** nati_ueno has quit IRC | 06:05 | |
| openstackgerrit | Jenkins proposed a change to openstack/heat: Imported Translations from Transifex https://review.openstack.org/62551 | 06:06 |
| *** yogesh has joined #heat | 06:09 | |
| *** Seal has joined #heat | 06:10 | |
| *** yogesh_ has joined #heat | 06:16 | |
| *** yogesh has quit IRC | 06:16 | |
| *** rongze has joined #heat | 06:20 | |
| *** yogesh_ has quit IRC | 06:23 | |
| *** nkhare has joined #heat | 06:24 | |
| *** Seal has quit IRC | 06:25 | |
| *** rongze has quit IRC | 06:25 | |
| *** yogesh has joined #heat | 06:27 | |
| *** rongze has joined #heat | 06:27 | |
| *** rongze has quit IRC | 06:28 | |
| *** Seal has joined #heat | 06:30 | |
| lifeless | asalkeld: great | 06:33 |
| asalkeld | lifeless, checkout my comments in the bug | 06:33 |
| lifeless | stevebaker: so - neutron doesn't push data to nova, but nova exposes neutron data in metadata | 06:33 |
| asalkeld | https://bugs.launchpad.net/heat/+bug/1261605 | 06:34 |
| uvirtbot | Launchpad bug 1261605 in heat "migration 30 -> 31 (031_stack_lock) fails" [Critical,In progress] | 06:34 |
| lifeless | stevebaker: so I don't quite follow your concern | 06:34 |
| *** rushiagr has quit IRC | 06:36 | |
| *** Seal has quit IRC | 06:37 | |
| *** nati_ueno has joined #heat | 06:37 | |
| *** rongze has joined #heat | 06:37 | |
| *** yogesh has quit IRC | 06:44 | |
| *** yogesh_ has joined #heat | 06:44 | |
| *** rushiagr has joined #heat | 06:44 | |
| *** yogesh_ has quit IRC | 06:46 | |
| *** Seal has joined #heat | 06:47 | |
| lifeless | asalkeld: we didn't set latin1 | 06:47 |
| asalkeld | ok | 06:47 |
| *** SushilKM has joined #heat | 06:47 | |
| lifeless | SpamapS: ^ | 06:47 |
| asalkeld | lifeless, I have deleted all copies of your sqldump, I can just use a trivial dump and change the tables to latin1 | 06:49 |
| asalkeld | (so no worries of the actual contents) | 06:49 |
| asalkeld | also that tenant_idx and stack_status_idx makes no difference | 06:49 |
| asalkeld | (you extra index) | 06:50 |
| lifeless | cool | 06:50 |
| lifeless | SpamapS: do you know about those extra indices? | 06:50 |
| *** amritanshu_RnD has joined #heat | 06:52 | |
| *** vkozhukalov has quit IRC | 06:52 | |
| stevebaker | asalkeld: i think there are about 4 different places utf8 can/should be set for mysql, and they all need to be to get utf8 end-to-end | 06:52 |
| asalkeld | stevebaker, well we need to decide if this is a valid bug | 06:53 |
| *** akuznetsov has quit IRC | 06:59 | |
| *** yogesh has joined #heat | 07:01 | |
| *** SushilKM has quit IRC | 07:02 | |
| lifeless | asalkeld: we didn't do anything to deliberately get that | 07:02 |
| lifeless | asalkeld: so permitting it to happen is a bug | 07:02 |
| stevebaker | lifeless: maybe os-db-create needs to do a: | 07:03 |
| stevebaker | create database heat default character set utf8; | 07:03 |
| asalkeld | I am making a test, to check for non-utf8 | 07:03 |
| lifeless | maybe | 07:03 |
| *** Seal has quit IRC | 07:03 | |
| lifeless | CONSTRAINT `stack_ibfk_2` FOREIGN KEY (`user_creds_id`) REFERENCES `user_creds` (`id`) | 07:03 |
| lifeless | ) ENGINE=InnoDB DEFAULT CHARSET=utf8 | | 07:03 |
| lifeless | for show create table stack in a recent overcloud | 07:04 |
| lifeless | I want to say that something in the heat space has changed | 07:04 |
| lifeless | in the last month or so | 07:04 |
| stevebaker | lifeless: sqlalchemy 015_grizzly specifies utf8 on the tables, so I don't _think_ anything has changed recently | 07:05 |
| SpamapS | lifeless: extra indices? no? | 07:05 |
| asalkeld | some sql people: | 07:06 |
| asalkeld | select TABLE_COLLATION from information_schema.TABLES where TABLE_COLLATION!='utf8_general_ci'; | 07:06 |
| lifeless | SpamapS: heat db on the underclou | 07:06 |
| lifeless | SpamapS: has two indices that new heat db's don't have | 07:06 |
| SpamapS | regarding the utf8 problem on indexes: it is ridiculous to have utf8 UUID's. Anyway, I'm still not ready to champion that bug just yet. | 07:06 |
| asalkeld | I want to say "if there is no "utf8" in the table_collation field | 07:07 |
| SpamapS | lifeless: let me look at that | 07:07 |
| lifeless | SpamapS: and has CHARSET latin1 | 07:07 |
| lifeless | SpamapS: on the stack table | 07:07 |
| lifeless | SpamapS: this is what make migration 31 barf | 07:07 |
| SpamapS | lifeless: the charset latin1 may have been some default leaking from our de-tuned mysql | 07:07 |
| lifeless | SpamapS: overcloud heat doesn't have the latin1 | 07:07 |
| lifeless | SpamapS: it also doesn't have the extra indices | 07:08 |
| SpamapS | lifeless: right. Hard to say why the table would be latin1. | 07:08 |
| lifeless | SpamapS: I know, right? | 07:08 |
| stevebaker | SpamapS, lifeless, is it because you're calling create database explicitly instead of letting sqlalchemy migrate fu do it? | 07:08 |
| SpamapS | lifeless: undercloud / is full again :( | 07:09 |
| SpamapS | lifeless: I'm guessing those extra indexes are my fault. :) | 07:15 |
| SpamapS | lifeless: ~/.mysql_history has them being added. | 07:15 |
| SpamapS | lifeless: I vaguely remember tinkering with them. | 07:16 |
| SpamapS | and indeed, server and db are defaulted to latin1 instead of utf8 as they should be | 07:16 |
| *** hub_cap has joined #heat | 07:17 | |
| SpamapS | lifeless: ok so I think we probably need to assert our mysql configuration to set character_set_server to utf8. | 07:21 |
| stevebaker | but I wonder why the utf8 on the migrate table create wasn't applied? | 07:21 |
| *** yogesh has quit IRC | 07:21 | |
| *** rushiagr has quit IRC | 07:22 | |
| *** rushiagr has joined #heat | 07:22 | |
| *** faramir1 has joined #heat | 07:22 | |
| SpamapS | stevebaker: do we actually assert utf8 anywhere? | 07:28 |
| *** ccs has joined #heat | 07:29 | |
| asalkeld | SpamapS I don't think so | 07:30 |
| SpamapS | ok, so what is happening then is that our clients _do_ set utf8 but our servers do not... | 07:31 |
| SpamapS | this leads to a client which sets "nothing" getting "latin1" | 07:31 |
| SpamapS | and, most likely, the client that was in use when the database was being dickered with (probably by me) was one such client | 07:32 |
| SpamapS | this gets really ugly if you already have utf-8 characters in the database, but AFAICT, we shouldn't | 07:32 |
| SpamapS | there's no real bug then in Heat. | 07:33 |
| *** tomek_adamczewsk has joined #heat | 07:36 | |
| *** ccs has quit IRC | 07:36 | |
| stevebaker | I shall mark that Essentiall/Invalid then :) | 07:39 |
| *** ccs has joined #heat | 07:39 | |
| *** Seal has joined #heat | 07:39 | |
| *** SushilKM has joined #heat | 07:39 | |
| stevebaker | I mean Critical/Invalid | 07:40 |
| *** liang has quit IRC | 07:48 | |
| *** rongze has quit IRC | 07:49 | |
| *** rongze has joined #heat | 07:49 | |
| SpamapS | stevebaker: this is why Drizzle removed charsets and just made everything utf8. | 07:50 |
| *** gokrokve has quit IRC | 07:50 | |
| *** nati_ueno has quit IRC | 07:54 | |
| *** tomek_adamczewsk has quit IRC | 07:55 | |
| *** rushiagr is now known as rushiagr_away | 07:55 | |
| *** tomek_adamczewsk has joined #heat | 07:56 | |
| asalkeld | o well at least I setup a new pgp key :O | 07:56 |
| *** che-arne has quit IRC | 07:56 | |
| *** yogesh has joined #heat | 07:57 | |
| openstackgerrit | Mitsuru Kanabuchi proposed a change to openstack/heat: Implements resource type NetworkGateway https://review.openstack.org/62287 | 07:59 |
| sgran | /win 31 | 08:01 |
| sgran | arg | 08:01 |
| *** SushilKM__ has joined #heat | 08:03 | |
| *** SushilKM has quit IRC | 08:04 | |
| *** harlowja is now known as harlowja_away | 08:04 | |
| *** TonyBurn has joined #heat | 08:11 | |
| *** rushiagr_away is now known as rushiagr | 08:12 | |
| *** tomek_adamczewsk has quit IRC | 08:14 | |
| *** tomek_adamczewsk has joined #heat | 08:16 | |
| *** akuznetsov has joined #heat | 08:22 | |
| *** vkozhukalov has joined #heat | 08:29 | |
| *** tomek_adamczewsk has quit IRC | 08:35 | |
| *** pshchelo has joined #heat | 08:35 | |
| *** tomek_adamczewsk has joined #heat | 08:36 | |
| shardy | morning | 08:40 |
| therve | Hi shardy | 08:41 |
| therve | So I missed the meeting (again), but the new times are just, like 4 to 6h later? | 08:41 |
| therve | Why not a bigger difference? | 08:42 |
| therve | I guess to keep NZ happy? :) | 08:42 |
| openstackgerrit | Simon Pasquier proposed a change to openstack/heat: Simplify update_with_template() method https://review.openstack.org/62832 | 08:52 |
| *** Tross has quit IRC | 08:52 | |
| *** ifarkas has joined #heat | 09:01 | |
| *** tomek_adamczewsk has quit IRC | 09:02 | |
| *** tomek_adamczewsk has joined #heat | 09:02 | |
| shardy | therve: Hmm, I didn't think we decided the new time, we were going to post a ML thread with a doodle poll IIRC | 09:03 |
| shardy | therve: Then alternate like ceilometer does, keeping the current time every other week | 09:03 |
| lifeless | stevebaker: so I think invalid is missing the point :) | 09:03 |
| lifeless | stevebaker: if we can get it wrong others will | 09:04 |
| lifeless | stevebaker: there should be a cross-check in there somewhere to prevent that being unnoticed | 09:04 |
| therve | shardy, Yeah there is a doodle, the times are 0-2 UTC | 09:05 |
| *** tomek_adamczewsk has quit IRC | 09:05 | |
| *** tomek_adamczewsk has joined #heat | 09:05 | |
| therve | At least for me that would make it a no-go | 09:05 |
| shardy | therve: ouch | 09:05 |
| shardy | therve: Yeah, pretty sure I'll be sleeping then too | 09:06 |
| therve | I'd be okay if the current one was perfect, but it's also slightly inconvenient :/ | 09:06 |
| shardy | Yeah, nearly all my meetings end up eating up my free evenings during the week | 09:07 |
| shardy | No way to make it convenient for everyone tho.. | 09:07 |
| SpamapS | lifeless: this is one of those meta-bugs in MySQL.. _everyone_ screws this up. | 09:07 |
| shardy | therve: Did you see the minutes, where we discussed the instance-users proposals? | 09:08 |
| shardy | therve: consensus seems to be to proceed with option (2) where we still create users associated with the credentials, but put them in a separate heat-specific domain | 09:10 |
| *** tomek_adamczewsk has quit IRC | 09:11 | |
| therve | shardy, Yeah I read it. Looking at stevebaker's section right now too | 09:16 |
| lifeless | SpamapS: sure, but can we be one of the projects that *notices* this? | 09:16 |
| SpamapS | lifeless: I tend to agree with you that the bug should be kept open. The bug is actually that the error is misleading and unhelpful. I want to make it clear though, it is not easy to predict when this will fail. | 09:19 |
| SpamapS | lifeless: have had it bite me in production more times than I can count. | 09:19 |
| SpamapS | and by it I mean "charset WTF" | 09:19 |
| openstackgerrit | Stephen Sugden proposed a change to openstack/heat: Add PortCheck resource https://review.openstack.org/62078 | 09:19 |
| *** ccs has quit IRC | 09:20 | |
| shardy | therve: I'm not sure about stevebaker's proposal, I see the advantages, but I don't like the fact that we can't do any policy enforcement, because the keypair isn't backed by a real user | 09:20 |
| shardy | therve: i.e whose token will the engine return when it validates the signature? | 09:21 |
| therve | Yeah I'm just curious :) | 09:21 |
| shardy | I also don't like the fact that it makes Heat the authentication service | 09:21 |
| therve | But I agree, 2) looks like the most promising, I like it | 09:21 |
| shardy | but I would support it being added as an optional mode, in addition to domain users | 09:21 |
| *** ccs has joined #heat | 09:21 | |
| lifeless | SpamapS: can we check the schema once on attaching or something ? | 09:22 |
| lifeless | SpamapS: care to reopen it then? | 09:22 |
| shardy | I don't see why we couldn't do both, in a way which is easily selectable via the heat.conf | 09:22 |
| shardy | then heat standalone folks could use (5) and everyone else could use the (arguably more robust) (2) | 09:22 |
| SpamapS | lifeless: what we can do is check it in db_sync | 09:23 |
| shardy | We need an answer to the whose-token-is-it question first tho | 09:23 |
| therve | shardy, Do we still need trusts with the second option? | 09:25 |
| *** akuznetsov has quit IRC | 09:26 | |
| SpamapS | lifeless: re-opened | 09:28 |
| shardy | therve: No, not unless you want the user inside the instance to impersonate the stack owning user | 09:29 |
| shardy | therve: which in many/most cases probably isn't required | 09:29 |
| therve | We have a trick anyway for that already, no? | 09:29 |
| shardy | therve: although it could make things simpler if we did, as it would avoid having to map user ids to resources, or do a lookup | 09:30 |
| therve | Right | 09:30 |
| therve | _load_user_creds | 09:30 |
| shardy | therve: Yeah, it's either do a policy check of userid-> resource, or do impersonation and allow the user to access any resource in the stack | 09:30 |
| shardy | comes down to whether we want a user per stack or a user per resource | 09:30 |
| shardy | The other issue is we'd need a way to defeat the tenant scoping if we don't use impersonation, which is maybe not desirable | 09:31 |
| shardy | we could have some resource DB lookup which filters on the unique stack_user_id or something I guess | 09:32 |
| *** Tross has joined #heat | 09:32 | |
| shardy | either way with or without trusts, I don't think it will be super complex | 09:33 |
| shardy | first step is to port heat_keystoneclient to v3 only, which I've already started | 09:33 |
| shardy | another keystone bug I need to fix first.. | 09:33 |
| *** ccs has quit IRC | 09:36 | |
| *** giulivo has joined #heat | 09:41 | |
| *** urulama has quit IRC | 09:56 | |
| openstackgerrit | Chmouel Boudjnah proposed a change to openstack/python-heatclient: Show better error when json fail to parse template https://review.openstack.org/62311 | 09:59 |
| *** SergeyLukjanov has joined #heat | 10:03 | |
| openstackgerrit | Liang Chen proposed a change to openstack/heat: Enable parallel while deleting neutron resources https://review.openstack.org/62580 | 10:05 |
| *** nkhare is now known as nkhare_afk | 10:06 | |
| *** faramir1 has quit IRC | 10:10 | |
| *** rongze has quit IRC | 10:11 | |
| *** liang has joined #heat | 10:12 | |
| *** derekh has joined #heat | 10:15 | |
| *** rongze has joined #heat | 10:23 | |
| *** yogesh has quit IRC | 10:24 | |
| *** pshchelo has quit IRC | 10:25 | |
| *** pshchelo has joined #heat | 10:27 | |
| *** rushiagr is now known as rushiagr_away | 10:30 | |
| *** akuznetsov has joined #heat | 10:32 | |
| openstackgerrit | huangtianhua proposed a change to openstack/heat: "Size" should be required para on Volume resource https://review.openstack.org/63083 | 10:41 |
| *** rushiagr_away is now known as rushiagr | 10:47 | |
| *** urulama has joined #heat | 10:50 | |
| *** aignatov has joined #heat | 10:51 | |
| *** tomek_adamczewsk has joined #heat | 10:53 | |
| *** SushilKM__ has quit IRC | 10:59 | |
| *** nosnos has quit IRC | 11:00 | |
| *** Seal has quit IRC | 11:01 | |
| *** tomek_adamczewsk has quit IRC | 11:07 | |
| *** blues-man has joined #heat | 11:10 | |
| pshchelo | hi all, I have a question - in my recent patch that was approved and merged, the last patch-set actually did not fix the error that commit message announced, and I hugely apologize for that :( how should I proceed? | 11:12 |
| *** rongze has quit IRC | 11:13 | |
| shardy | pshchelo: Post another patch which fixes the problem ;) | 11:13 |
| blues-man | hello folks | 11:14 |
| shardy | pshchelo: If you wish, you can post a patch reverting the one we committed first | 11:14 |
| shardy | pshchelo: which patch? | 11:14 |
| pshchelo | https://review.openstack.org/#/c/61816/3 | 11:14 |
| pshchelo | while I was fighting the json, the original intent got lost.. | 11:15 |
| pshchelo | just noticed it when rebasing now on my other patches that are in progress | 11:16 |
| shardy | pshchelo: Please just post a followup patch explaining the issue and fixing the problem | 11:17 |
| shardy | pshchelo: reference that commit or review URL in the commit message so we can connect the two patches | 11:17 |
| pshchelo | shardy: ref it in commit message? | 11:18 |
| pshchelo | ok, got you | 11:18 |
| *** guohliu has quit IRC | 11:18 | |
| shardy | In the detailed part of the commit message, say patch merged under review <url> introduces <problem explanation>, this fixes <problem> | 11:19 |
| openstackgerrit | Pavlo Shchelokovskyy proposed a change to openstack/heat: Fix misused assertTrue in unit tests https://review.openstack.org/63089 | 11:26 |
| pshchelo | shardy: done | 11:27 |
| *** alexheneveld has joined #heat | 11:27 | |
| *** aignatov has quit IRC | 11:28 | |
| *** rpothier has joined #heat | 11:32 | |
| *** rpothier_ has quit IRC | 11:34 | |
| *** SushilKM__ has joined #heat | 11:34 | |
| *** yogesh has joined #heat | 11:35 | |
| *** nkhare_afk is now known as nkhare | 11:35 | |
| *** yogesh has quit IRC | 11:39 | |
| *** rushiagr has quit IRC | 11:40 | |
| *** bgorski has joined #heat | 11:45 | |
| *** aignatov has joined #heat | 11:47 | |
| *** bgorski has quit IRC | 11:49 | |
| *** rushiagr has joined #heat | 11:50 | |
| *** rushiagr has quit IRC | 11:55 | |
| *** rushiagr has joined #heat | 11:55 | |
| *** nkhare is now known as nkhare_afk | 11:56 | |
| *** swygue has quit IRC | 11:59 | |
| *** bgorski has joined #heat | 12:02 | |
| *** rongze has joined #heat | 12:04 | |
| *** rpothier_ has joined #heat | 12:15 | |
| *** aignatov has quit IRC | 12:16 | |
| *** rpothier has quit IRC | 12:18 | |
| *** aignatov has joined #heat | 12:20 | |
| *** rongze_ has joined #heat | 12:24 | |
| *** giulivo has quit IRC | 12:24 | |
| *** rongze has quit IRC | 12:28 | |
| *** rpothier_ has quit IRC | 12:29 | |
| *** agileclipse has joined #heat | 12:37 | |
| *** nkhare_afk has quit IRC | 12:40 | |
| *** pshchelo has quit IRC | 12:40 | |
| *** urulama has quit IRC | 12:46 | |
| *** giulivo has joined #heat | 12:47 | |
| openstackgerrit | A change was merged to openstack/heat: Add nested resource references to resource group https://review.openstack.org/61614 | 12:50 |
| *** SushilKM__ has quit IRC | 12:51 | |
| *** blues-man has quit IRC | 12:51 | |
| *** akuznetsov has quit IRC | 12:53 | |
| *** rushiagr has quit IRC | 13:09 | |
| *** pshchelo has joined #heat | 13:09 | |
| openstackgerrit | Pavlo Shchelokovskyy proposed a change to openstack/heat: Fix asserttion of types https://review.openstack.org/61436 | 13:14 |
| *** pshchelo has quit IRC | 13:19 | |
| *** pafuent has joined #heat | 13:22 | |
| *** pafuent has left #heat | 13:23 | |
| sdake | morning | 13:28 |
| shardy | Hi sdake | 13:28 |
| sdake | hey shardy | 13:28 |
| sdake | cloud borat has become less offensive, and less funny as a result :( | 13:31 |
| *** rpothier has joined #heat | 13:36 | |
| *** blomquisg has joined #heat | 13:36 | |
| *** SushilKM__ has joined #heat | 13:36 | |
| *** Seal has joined #heat | 13:36 | |
| *** andersonvom has quit IRC | 13:39 | |
| *** Seal has quit IRC | 13:41 | |
| *** chandankumar_ has joined #heat | 13:44 | |
| *** chandankumar_ has quit IRC | 13:44 | |
| *** aweiteka has joined #heat | 13:45 | |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat: Avoid using python build-in class property and exit https://review.openstack.org/62814 | 13:45 |
| *** akuznetsov has joined #heat | 13:47 | |
| *** tomek_adamczewsk has joined #heat | 13:51 | |
| *** urulama has joined #heat | 13:55 | |
| *** urulama_ has joined #heat | 13:56 | |
| *** akuznetsov has quit IRC | 13:58 | |
| *** derekh has quit IRC | 14:00 | |
| *** urulama has quit IRC | 14:00 | |
| *** chandankumar has quit IRC | 14:06 | |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat: Removed unused variables https://review.openstack.org/62827 | 14:09 |
| openstackgerrit | Liang Chen proposed a change to openstack/heat: Enable parallel while deleting neutron resources https://review.openstack.org/62580 | 14:10 |
| *** blues-man has joined #heat | 14:12 | |
| openstackgerrit | A change was merged to openstack/heat: Use property name constants in Ceilometer alarm https://review.openstack.org/62413 | 14:13 |
| *** rushiagr has joined #heat | 14:14 | |
| *** sergmelikyan has quit IRC | 14:27 | |
| sdake | shardy did we cancel the heat meeting on the 1st? | 14:28 |
| *** andersonvom has joined #heat | 14:28 | |
| sdake | I was half-attending the irc meeting yesterday and don't recall the outcome | 14:28 |
| shardy | sdake: Yes, cancelled for the next two meetings | 14:28 |
| *** rushiagr has quit IRC | 14:29 | |
| *** rushiagr has joined #heat | 14:29 | |
| * sdake shakes fists at zimbra | 14:29 | |
| skraynev | shardy: oh. Is there no meeting 25 Dec and 1 Jun ? | 14:30 |
| sdake | some young enterprising individual should put that on the heatmeetings page ;-) | 14:30 |
| shardy | skraynev: No, due to Christmas and New Years holidays | 14:30 |
| skraynev | shardy: Well, what about 5 Jun? | 14:31 |
| sdake | what happens jun 5th? | 14:31 |
| skraynev | shardy: different countries different holidays))) | 14:32 |
| pasquier-s | skraynev, you mean January, right? | 14:32 |
| shardy | skraynev: I'm well aware of that, but a sufficient number of the core team will be absent, so we decided to cancel the meetings | 14:32 |
| skraynev | sdake: I have asked about meeting in 5 Jan) | 14:32 |
| therve | Well Jan 5th is a sunday, so it should be a problem | 14:32 |
| skraynev | pasquier-s: yes | 14:33 |
| therve | should not | 14:33 |
| skraynev | lol) I have looked to other mount in calendar. sorry) | 14:34 |
| shardy | andersonvom: hi | 14:35 |
| sdake | skraynev our next meeting is jan 8 2014 | 14:35 |
| openstackgerrit | Pavlo Shchelokovskyy proposed a change to openstack/heat: Fix comparison with singletons https://review.openstack.org/62852 | 14:35 |
| sdake | I edited the wiki based upon shardy's comments | 14:35 |
| skraynev | sdake: I got it. thank you. | 14:35 |
| *** jasond` has joined #heat | 14:37 | |
| skraynev | long period for the resting and celebration NY with family)) | 14:37 |
| *** jcru has joined #heat | 14:40 | |
| *** liang has quit IRC | 14:40 | |
| *** pfreund has joined #heat | 14:42 | |
| *** openstack has joined #heat | 14:49 | |
| *** sabeen has joined #heat | 14:49 | |
| *** amritanshu_RnD has quit IRC | 14:50 | |
| andersonvom | shardy: morning! | 14:51 |
| shardy | Hey andersonvom, just wanted to quickly chat about the policy additions | 14:52 |
| shardy | I dug out my patch adding policy enforcement to the native API, and it's nearly there | 14:52 |
| *** gokrokve has joined #heat | 14:52 | |
| andersonvom | shardy: sure. I just saw your comments. if I can help in any way, I have bandwidth to do it | 14:52 |
| shardy | So if you're happy, I can finish adding the tests this afternoon, and post it for you to use as a basis for the management-api additions? | 14:52 |
| andersonvom | shardy: sounds good! | 14:53 |
| shardy | andersonvom: Yeah, after I wrote the comment I realized what I had was nearly there sat on a branch, as always adding tests is taking a while.. :) | 14:53 |
| andersonvom | shardy: I feel ya ;) | 14:53 |
| *** arbylee has joined #heat | 14:54 | |
| shardy | andersonvom: Ok, cool, I should have something for you in the next few hours | 14:54 |
| andersonvom | \o/ | 14:54 |
| *** tims has joined #heat | 14:54 | |
| *** vijendar has joined #heat | 14:55 | |
| openstackgerrit | Pavlo Shchelokovskyy proposed a change to openstack/heat: Fix comparison with singletons https://review.openstack.org/62852 | 14:57 |
| openstackgerrit | Pavlo Shchelokovskyy proposed a change to openstack/heat: Fix comparison with singletons https://review.openstack.org/62852 | 14:57 |
| *** pshchelo has joined #heat | 14:59 | |
| tomek_adamczewsk | Hi, wanted to ask about https://wiki.openstack.org/wiki/Heat/Environments | 15:08 |
| *** blamar has joined #heat | 15:08 | |
| tomek_adamczewsk | example shows that you can modify resource content through environment, but current source code (if I understand it correctly) doesn't support these sections in environment model | 15:09 |
| tomek_adamczewsk | is it something that is expected to work, or just an outdated document/example? | 15:09 |
| tomek_adamczewsk | *) expected to work one day | 15:10 |
| *** jergerber has joined #heat | 15:11 | |
| *** topol has joined #heat | 15:12 | |
| *** mestery_ has joined #heat | 15:12 | |
| *** AndrewMichalik has joined #heat | 15:15 | |
| *** mestery has quit IRC | 15:16 | |
| *** blamar has quit IRC | 15:16 | |
| *** blamar has joined #heat | 15:17 | |
| *** jasond` has quit IRC | 15:17 | |
| *** IlyaE has joined #heat | 15:21 | |
| *** mestery has joined #heat | 15:22 | |
| *** mestery_ has quit IRC | 15:25 | |
| *** rcleere has joined #heat | 15:27 | |
| *** nkhare has joined #heat | 15:30 | |
| *** rcleere has quit IRC | 15:39 | |
| *** sjmc has joined #heat | 15:40 | |
| *** vkozhukalov has quit IRC | 15:40 | |
| *** rongze has joined #heat | 15:42 | |
| *** rongze_ has quit IRC | 15:45 | |
| *** jasond` has joined #heat | 15:46 | |
| *** john-n-seattle1 has joined #heat | 15:47 | |
| openstackgerrit | A change was merged to openstack/heat: Convert Neutron resources to new Schema format https://review.openstack.org/61823 | 15:55 |
| *** SergeyLukjanov has quit IRC | 15:56 | |
| *** swygue has joined #heat | 15:56 | |
| *** ccs has joined #heat | 16:02 | |
| *** rcleere has joined #heat | 16:04 | |
| *** SushilKM__ has quit IRC | 16:05 | |
| *** rongze has quit IRC | 16:15 | |
| *** rongze has joined #heat | 16:16 | |
| *** tomek_adamczewsk has quit IRC | 16:18 | |
| *** swygue has quit IRC | 16:18 | |
| *** SushilKM__ has joined #heat | 16:18 | |
| *** Seal has joined #heat | 16:18 | |
| *** sdake_ has joined #heat | 16:22 | |
| *** sdake_ has quit IRC | 16:22 | |
| *** sdake_ has joined #heat | 16:22 | |
| *** rongze has quit IRC | 16:22 | |
| sdake_ | wow target got hacked | 16:22 |
| sdake_ | 40 million credit cards compromised | 16:22 |
| *** rongze has joined #heat | 16:22 | |
| *** urulama_ has quit IRC | 16:22 | |
| sdake_ | that is going to cost a boatload of money to fix | 16:22 |
| *** dims has quit IRC | 16:26 | |
| openstackgerrit | Richard Lee proposed a change to openstack/heat: Add way to group exceptions in DependencyTaskGroup https://review.openstack.org/62479 | 16:27 |
| openstackgerrit | Richard Lee proposed a change to openstack/heat: Refactor Instance#_check_active https://review.openstack.org/62478 | 16:27 |
| *** swygue has joined #heat | 16:27 | |
| *** aignatov has quit IRC | 16:28 | |
| *** dims has joined #heat | 16:30 | |
| *** SergeyLukjanov has joined #heat | 16:38 | |
| *** rushiagr is now known as rushiagr_away | 16:39 | |
| *** sjm__ has joined #heat | 16:42 | |
| SpamapS | sdake_: they seem pretty inept. Wife ordered a couch that was supposed to be shipped from Ohio to California, and was instead shipped to Wisconsin. | 16:45 |
| *** pshchelo has quit IRC | 16:45 | |
| SpamapS | sdake_: and then it sat in their distribution warehouse for 3 weeks and nobody knew where it was. | 16:45 |
| *** dims_ has joined #heat | 16:46 | |
| *** arbylee has quit IRC | 16:46 | |
| SpamapS | grncdr: your PortCheck resource has me concerned. | 16:47 |
| *** jasond`` has joined #heat | 16:47 | |
| openstackgerrit | Pablo Andres Fuente proposed a change to openstack/heat: Deleting a stack with a deleted nested stack fails https://review.openstack.org/62752 | 16:47 |
| *** aignatov has joined #heat | 16:51 | |
| *** dims has quit IRC | 16:52 | |
| *** jasond` has quit IRC | 16:52 | |
| *** sjmc has quit IRC | 16:52 | |
| *** SushilKM__ has quit IRC | 16:56 | |
| *** Seal has quit IRC | 16:57 | |
| *** AndrewMichalik has quit IRC | 17:00 | |
| openstackgerrit | A change was merged to openstack/heat: Convert numeric values in instance metadata to strings https://review.openstack.org/61883 | 17:02 |
| *** SushilKM__ has joined #heat | 17:02 | |
| *** Seal has joined #heat | 17:02 | |
| *** arbylee has joined #heat | 17:03 | |
| *** SergeyLukjanov is now known as _SergeyLukjanov | 17:03 | |
| *** _SergeyLukjanov has quit IRC | 17:04 | |
| *** pablosan has joined #heat | 17:05 | |
| *** SergeyLukjanov has joined #heat | 17:07 | |
| *** sjm__ has quit IRC | 17:07 | |
| openstackgerrit | DennyZhang proposed a change to openstack/heat: add assertion to heat-keystone-setup script https://review.openstack.org/61905 | 17:07 |
| *** Seal has quit IRC | 17:08 | |
| *** rushiagr_away is now known as rushiagr | 17:15 | |
| *** aignatov has quit IRC | 17:17 | |
| *** pfreund has quit IRC | 17:18 | |
| openstackgerrit | Pablo Andres Fuente proposed a change to openstack/heat: Deleting a stack with a deleted nested stack fails https://review.openstack.org/62752 | 17:21 |
| *** blues-man has quit IRC | 17:21 | |
| *** SushilKM__ has left #heat | 17:39 | |
| *** rongze has quit IRC | 17:39 | |
| *** rongze has joined #heat | 17:40 | |
| *** rongze has quit IRC | 17:44 | |
| *** vkozhukalov has joined #heat | 17:46 | |
| *** nkhare has quit IRC | 17:47 | |
| *** dims_ is now known as dims | 17:50 | |
| *** AndrewMichalik has joined #heat | 17:51 | |
| openstackgerrit | Anderson Mesquita proposed a change to openstack/heat: Check resources are in expected states https://review.openstack.org/58641 | 17:55 |
| *** che-arne has joined #heat | 17:57 | |
| *** TonyBurn has quit IRC | 17:58 | |
| *** aignatov has joined #heat | 18:03 | |
| *** harlowja_away is now known as harlowja | 18:04 | |
| openstackgerrit | Jason Dunsmore proposed a change to openstack/heat: Allow an in-progress stack to be deleted https://review.openstack.org/63002 | 18:06 |
| grncdr | SpamapS: you around? | 18:07 |
| grncdr | ugh, also does anybody know how to put links into gerrit comments? | 18:09 |
| grncdr | they have commit hashes in them which gerrit seems determined to interfere with | 18:09 |
| jasond`` | grncdr: maybe put it on it's own line with space before it? | 18:10 |
| grncdr | hah nope that just does even more broken things :( | 18:12 |
| grncdr | wtf gerrit authors | 18:12 |
| openstackgerrit | Steven Hardy proposed a change to openstack/heat: Add policy enforcement to ReST API https://review.openstack.org/63185 | 18:13 |
| shardy | andersonvom: ^^ | 18:13 |
| shardy | andersonvom: Not quite finished, still need a little bit more work on the tests | 18:13 |
| shardy | but I'm going afk for a couple of hours so there's a WIP for you to review | 18:13 |
| andersonvom | shardy: sweet. I'll take a look at it right after lunch. thanks! | 18:14 |
| shardy | andersonvom: Cool, I'll be back later after some food also | 18:15 |
| clarkb | grncdr: its actually our comment link config doing that | 18:15 |
| SpamapS | grncdr: I am now | 18:15 |
| clarkb | grncdr: are these links to openstack project commits? you can just dump the hash in there | 18:15 |
| grncdr | clarkb: I want to link to specific lines | 18:15 |
| clarkb | though we should probably go back to simply linking change ids and not commits | 18:15 |
| SpamapS | clarkb: I think this is just for the sake of discussion. | 18:16 |
| grncdr | so I'm doing https://github.com/openstack/heat/blob/<hash>/heat/engine/somefile.py#L123 | 18:16 |
| grncdr | SpamapS: since you're here, I'll post the review comment without the hashes (which should stop the auto-linking) | 18:17 |
| grncdr | e.g. just pointed at master | 18:17 |
| clarkb | grncdr: you know a simple solution is probably ot wrap the regex in \b | 18:17 |
| grncdr | clarkb: who has permissions to do that? | 18:17 |
| * grncdr hopes it's you ;) | 18:18 | |
| clarkb | grncdr: anyone. the gerrit config is hosted in gerrit. I am poking at it now though | 18:18 |
| SpamapS | grncdr: did you just want to chat here about it? | 18:19 |
| grncdr | SpamapS: yeah, I did take the time to write this though so I'm going to post it first ;) | 18:19 |
| *** akuznetsov has joined #heat | 18:19 | |
| SpamapS | grncdr: This is for posterity.. so please.. be honest: http://4.bp.blogspot.com/-C9FfVjS-l_8/UXoMAjav1WI/AAAAAAAAB1w/EtbFBZyK10Y/s1600/CountRugen.jpg | 18:20 |
| *** alexheneveld has quit IRC | 18:21 | |
| grncdr | :P | 18:21 |
| clarkb | grncdr: https://review.openstack.org/63187 | 18:21 |
| *** urulama has joined #heat | 18:23 | |
| *** pvaneck has joined #heat | 18:23 | |
| openstackgerrit | A change was merged to openstack/heat: Simplify update_with_template() method https://review.openstack.org/62832 | 18:24 |
| grncdr | woo published finally | 18:24 |
| grncdr | also seriously wtf does everybody invent their own half-assed text markup? </rant> | 18:25 |
| grncdr | (not OpenStack/Heat, but gerrit) | 18:26 |
| *** urulama has quit IRC | 18:27 | |
| grncdr | SpamapS: the gist of my comment regarding exceptions is "heat already does this and it works" | 18:27 |
| *** IlyaE has quit IRC | 18:28 | |
| *** radix_ has joined #heat | 18:29 | |
| SpamapS | grncdr: Ah right, I forgot that we treat these things like plugins and thus we expect the authors to do crazy things like raise their own exceptions. :) | 18:29 |
| SpamapS | grncdr: On the "should we do this".. I think for a small scale this is fantastic and so probably "Yes we should do it". But I think we may want to document the dangers of enabling it.. you basically turn heat-engine into a DDoS machine. We might not be expecting heat-engine to do that the same way we expect tenant vms to. | 18:30 |
| *** IlyaE has joined #heat | 18:31 | |
| grncdr | SpamapS: fwiw I agree with you on both points. In my particular use case we're not running heat-api publically, we have an application in front of it that is responsible for generating the templates and instantiating them via the heat-api | 18:32 |
| grncdr | the DDoS aspect is worth considering though… is there currently a mechanism for whitelisting resource types that heat can instantiate? | 18:33 |
| SpamapS | grncdr: FWIW I like the idea a lot. | 18:33 |
| grncdr | for that matter, is there a mechanism for telling heat "also load resources from this other module?" (I never looked for that...) | 18:34 |
| *** yogesh has joined #heat | 18:34 | |
| SpamapS | #plugin_dirs=/usr/lib64/heat,/usr/lib/heat | 18:34 |
| *** ifarkas has quit IRC | 18:34 | |
| *** gokrokve has quit IRC | 18:36 | |
| *** PhilK has quit IRC | 18:36 | |
| SpamapS | grncdr: this is where I wonder if there isn't real monitoring as a service in development. because that is really what you want anyway. | 18:36 |
| *** PhilK has joined #heat | 18:37 | |
| *** rcleere has quit IRC | 18:37 | |
| *** bgorski has quit IRC | 18:37 | |
| *** fsargent has joined #heat | 18:38 | |
| grncdr | well, I specifically want it to interact with heat | 18:38 |
| grncdr | though I suppose that part would be easy | 18:38 |
| *** rongze has joined #heat | 18:40 | |
| *** rushiagr has quit IRC | 18:41 | |
| *** gokrokve has joined #heat | 18:41 | |
| *** dmueller_ has joined #heat | 18:42 | |
| SpamapS | grncdr: right if it was monitoring as a service you just feed monitoring as a service the waitcondition handle. | 18:42 |
| SpamapS | grncdr: you could basically create a little "under the covers" nested stack with a waitcondition and a monitoring as a service thing. | 18:43 |
| SpamapS | grncdr: I think for now we should just merge your thing but open a bug to replace its underbelly with something more scalable. | 18:43 |
| grncdr | makes sense to me | 18:44 |
| *** IlyaE has quit IRC | 18:45 | |
| *** yogesh has quit IRC | 18:45 | |
| *** rongze has quit IRC | 18:45 | |
| *** yogesh has joined #heat | 18:47 | |
| *** aignatov has quit IRC | 18:47 | |
| *** dmueller__ has joined #heat | 18:48 | |
| *** dmueller_ has quit IRC | 18:49 | |
| openstackgerrit | Stephen Sugden proposed a change to openstack/heat: Add PortCheck resource https://review.openstack.org/62078 | 18:50 |
| SpamapS | grncdr: responded in-line and added a few more comments | 18:51 |
| *** dmueller__ has quit IRC | 18:53 | |
| grncdr | regarding the delete_stack_after… reset_dummy_db will do the same? | 18:56 |
| grncdr | from reading the code that's really not clear... | 18:58 |
| SpamapS | grncdr: well reset should clear everything out. Let me verify that. | 18:58 |
| *** yogesh has quit IRC | 18:58 | |
| grncdr | it deletes the database, but stack_delete_after calls stack.delete() | 18:58 |
| *** yogesh has joined #heat | 18:58 | |
| grncdr | I'm not sure whether there are other resources left hanging around by creating a stack that wouldn't be in the database... | 18:59 |
| grncdr | actually, I think I was just confusing myself, since I'm used to be a client of heat and not working on heat itself | 18:59 |
| SpamapS | grncdr: this is also just in a "delete what you create so tests are isolated" thing. It isn't an integration test after all. | 19:00 |
| *** yogesh has quit IRC | 19:01 | |
| grncdr | yeah, it's that my brain wasn't used to thinking of "the db" as "heat engine state" | 19:01 |
| grncdr | because in my app tests "the db" is "app state" and "heat engine state" has to be cleaned up via the heat api | 19:02 |
| grncdr | anyways, brain fart deconstructed, I'll get rid of the suspenders | 19:02 |
| openstackgerrit | Stephen Sugden proposed a change to openstack/heat: Add PortCheck resource https://review.openstack.org/62078 | 19:03 |
| *** yogesh_ has joined #heat | 19:04 | |
| *** aignatov has joined #heat | 19:07 | |
| *** AndrewMichalik has quit IRC | 19:09 | |
| *** che-arne has quit IRC | 19:10 | |
| *** radez_g0n3 is now known as radez | 19:13 | |
| stevebaker | morning | 19:14 |
| *** yogesh_ has quit IRC | 19:15 | |
| *** yogesh has joined #heat | 19:17 | |
| andersonvom | morning! | 19:17 |
| stevebaker | I'm going to go ahead and assume that randellburt won't -1 bgorski's nomination | 19:19 |
| *** akuznetsov has quit IRC | 19:20 | |
| jasond`` | stevebaker: he +1ed it | 19:22 |
| jasond`` | stevebaker: http://dunsmor.com/shots/2013-12-19-132315_562x632.png | 19:23 |
| stevebaker | ah, zaneb hasn't | 19:23 |
| shardy | stevebaker: zaneb is on PTO today I think | 19:25 |
| *** AndrewMichalik has joined #heat | 19:25 | |
| *** lsmola_ has quit IRC | 19:25 | |
| openstackgerrit | Steven Hardy proposed a change to openstack/heat: Add policy enforcement to ReST API https://review.openstack.org/63185 | 19:25 |
| stevebaker | too late, I've added bgorski to core | 19:29 |
| *** andersonvom1 has joined #heat | 19:31 | |
| *** andersonvom is now known as Guest66053 | 19:31 | |
| *** andersonvom1 is now known as andersonvom | 19:31 | |
| *** Guest66053 has quit IRC | 19:32 | |
| andersonvom | shardy: just looked at your patch and I think I see where you're going now. I'm wondering how do you see the tenant_safe flag being used, though | 19:32 |
| *** AndrewMichalik has quit IRC | 19:42 | |
| *** swygue has quit IRC | 19:42 | |
| *** scroiset_ has joined #heat | 19:42 | |
| *** scroiset has quit IRC | 19:42 | |
| SpamapS | hm | 19:43 |
| SpamapS | updates don't wait for waitconditions | 19:44 |
| SpamapS | wondering if they should | 19:44 |
| SpamapS | or if we should have some way to make a waitcondition an update wait condition. | 19:44 |
| stevebaker | isn't a waitcondition IN_PROGRESS while waiting? | 19:52 |
| SpamapS | stevebaker: nothing sets it back to that during update | 19:55 |
| *** sdake has quit IRC | 19:55 | |
| SpamapS | stevebaker: I'm thinking that we just need an UpdateWaitCondition that works exactly like WaitCondition but, on update, clears the signals and waits for them again. | 19:55 |
| *** sdake has joined #heat | 19:56 | |
| *** sdake has quit IRC | 19:56 | |
| *** sdake has joined #heat | 19:56 | |
| *** swygue has joined #heat | 19:56 | |
| stevebaker | SpamapS: that is what SoftwareDeployment signalling will do (not that I've tried it) | 19:57 |
| SpamapS | stevebaker: yay, ok so I will defer trying to do mine and just hack a 1 second timeout wobble into our updates which will force replacement of the wait conditions. | 19:57 |
| SpamapS | lifeless: ^^ HOT to the rescue :) | 19:58 |
| stevebaker | there might need to be some way of saying to reapply the deployment even if nothing has changed | 19:59 |
| SpamapS | hm that won't work | 20:00 |
| SpamapS | sorry not your thing stevebaker | 20:00 |
| SpamapS | the wobble | 20:00 |
| SpamapS | I need the WaitConditionHandle to be replaced, not the waitcondition | 20:01 |
| *** pablosan has quit IRC | 20:06 | |
| *** fsargent has left #heat | 20:07 | |
| *** pablosan has joined #heat | 20:08 | |
| *** ccs has quit IRC | 20:08 | |
| *** swygue has quit IRC | 20:19 | |
| *** swygue has joined #heat | 20:22 | |
| *** urulama has joined #heat | 20:24 | |
| *** ccs has joined #heat | 20:25 | |
| openstackgerrit | Anderson Mesquita proposed a change to openstack/heat: Check resources are in expected states https://review.openstack.org/58641 | 20:26 |
| *** ccs has quit IRC | 20:26 | |
| *** ccs has joined #heat | 20:26 | |
| stevebaker | jasond``: any progress on https://bugs.launchpad.net/heat/+bug/1261433 ? | 20:27 |
| uvirtbot | Launchpad bug 1261433 in heat "heat-tempest-xx already has an action (DELETE) in progress" [High,Triaged] | 20:27 |
| *** IlyaE has joined #heat | 20:28 | |
| *** urulama has quit IRC | 20:29 | |
| arbylee | in regards to the etc/heat/policy.json, does anyone know the context around the role "heat_stack_user"? What does that role represent and why is it not allowed to do any cloudformation/cloudwatch actions? | 20:29 |
| arbylee | I'm trying to understand this upcoming change: https://review.openstack.org/#/c/63185/2/etc/heat/policy.json | 20:30 |
| arbylee | ^ shardy if you're around | 20:31 |
| stevebaker | arbylee: that role is just for instance users, they are only allowed to fetch information about their own resources | 20:33 |
| openstackgerrit | Stephen Sugden proposed a change to openstack/heat: Add PortCheck resource https://review.openstack.org/62078 | 20:33 |
| grncdr | ^ fixed my 80 character line :\ | 20:33 |
| jasond`` | stevebaker: i think this will fix it https://review.openstack.org/#/c/63002/ | 20:34 |
| jasond`` | I'm going to add that bug to the commit message on the next update | 20:35 |
| arbylee | stevebaker: thanks | 20:36 |
| stevebaker | jasond``: OK, have you also looked into stopping heat-api logging user exceptions with ERROR? | 20:39 |
| jasond`` | stevebaker: i'll start working on that now | 20:41 |
| *** rcleere has joined #heat | 20:43 | |
| *** Tross has quit IRC | 20:43 | |
| stevebaker | jasond``: OK, thanks. That would probably be the quickest fix. We're racking up quite a few rechecks | 20:44 |
| jasond`` | stevebaker: well, the quickest fix is https://review.openstack.org/#/c/62938/ | 20:46 |
| stevebaker | jasond``: that is not a fix, that is hiding the bug ;) | 20:47 |
| *** alexheneveld has joined #heat | 20:47 | |
| jasond`` | stevebaker: so the test should keep trying to delete the same stack twice? | 20:47 |
| *** aignatov has quit IRC | 20:48 | |
| stevebaker | jasond``: sure, it should work. A test should automatically clean up its resources, and it is desirable to also explicitly test doing a delete | 20:48 |
| stevebaker | in the short term we could add that error to tempest/etc/whitelist.yaml | 20:49 |
| jasond`` | okay. so the double delete should be done in a separate test, right? | 20:49 |
| jasond`` | ie. test_stack_double_delete | 20:51 |
| *** alexheneveld has quit IRC | 20:51 | |
| stevebaker | we could eventually have a test which exercises multi-engine, attempting actions while in_progress doing other actions | 20:51 |
| *** vkozhukalov has quit IRC | 20:51 | |
| jasond`` | sounds good | 20:52 |
| *** Seal has joined #heat | 20:56 | |
| *** jdob has joined #heat | 20:57 | |
| openstackgerrit | A change was merged to openstack/python-heatclient: Sync base and exceptions from oslo https://review.openstack.org/61991 | 20:58 |
| openstackgerrit | A change was merged to openstack/python-heatclient: Reuse BaseManager and Resource from oslo https://review.openstack.org/62348 | 20:58 |
| openstackgerrit | A change was merged to openstack/heat: Fix copy/paste errors in API docstrings https://review.openstack.org/62536 | 20:58 |
| *** Seal has quit IRC | 21:01 | |
| *** sabeen has quit IRC | 21:05 | |
| *** radez is now known as radez_g0n3 | 21:08 | |
| openstackgerrit | Steve Baker proposed a change to openstack/heat: A cloud-config resource based on SoftwareConfig https://review.openstack.org/63214 | 21:09 |
| openstackgerrit | Steve Baker proposed a change to openstack/heat: A multipart cloud-init resource based on SoftwareConfig https://review.openstack.org/63215 | 21:09 |
| openstackgerrit | Steve Baker proposed a change to openstack/heat: OS::Nova::Server support for software config https://review.openstack.org/58880 | 21:09 |
| openstackgerrit | Steve Baker proposed a change to openstack/heat: REST API for software config/deployment https://review.openstack.org/58878 | 21:09 |
| openstackgerrit | Steve Baker proposed a change to openstack/heat: Resource types for software config/deployment https://review.openstack.org/58879 | 21:09 |
| openstackgerrit | Steve Baker proposed a change to openstack/heat: Database model for software config/deployment https://review.openstack.org/58876 | 21:09 |
| openstackgerrit | Steve Baker proposed a change to openstack/heat: RPC service for software config/deployment https://review.openstack.org/58877 | 21:09 |
| jasond`` | shardy: still here? | 21:14 |
| asalkeld | morning | 21:17 |
| *** gokrokve has quit IRC | 21:17 | |
| *** topol has quit IRC | 21:17 | |
| *** gokrokve has joined #heat | 21:18 | |
| jasond`` | asalkeld: good morning | 21:19 |
| *** Tross has joined #heat | 21:21 | |
| *** gokrokve has quit IRC | 21:21 | |
| *** blomquisg has quit IRC | 21:23 | |
| *** swygue has quit IRC | 21:28 | |
| jasond`` | stevebaker: https://review.openstack.org/#/c/62080/3 fixes the API error logging issue | 21:30 |
| andersonvom | morning | 21:31 |
| asalkeld | ooo, we really need that jasond | 21:32 |
| *** swygue has joined #heat | 21:34 | |
| *** cody-somerville has joined #heat | 21:37 | |
| *** cody-somerville has quit IRC | 21:37 | |
| *** cody-somerville has joined #heat | 21:37 | |
| *** gokrokve has joined #heat | 21:38 | |
| SpamapS | weird | 21:41 |
| SpamapS | stack validation requires nova to be running | 21:41 |
| *** jasond has quit IRC | 21:42 | |
| *** jasond has joined #heat | 21:42 | |
| SpamapS | stevebaker: so for the event-list ordering.. I wonder if we should actually bring back the autoincrement ID, but just never expose it to the user. | 21:43 |
| SpamapS | the random order is really awful | 21:43 |
| stevebaker | why was the int id removed? I don't recall | 21:44 |
| *** slong has quit IRC | 21:44 | |
| SpamapS | stevebaker: leaks data | 21:44 |
| SpamapS | stevebaker: same reason nova ids are uuids. | 21:45 |
| asalkeld | brb: dropping kids off | 21:46 |
| *** SergeyLukjanov has quit IRC | 21:46 | |
| *** jdob has quit IRC | 21:46 | |
| stevebaker | another option, add a milliseconds column and sort by created_at, millis | 21:48 |
| SpamapS | stevebaker: indeed, and that will support a more distributed scale-out-per-resource story better.. assuming we have time synced ;) | 21:49 |
| stevebaker | mysql starts supporting sub-second datetime from version 5.6.4 | 21:50 |
| shardy | subsecond timestamps are truncated on mysql, ref bug #1246383 | 21:50 |
| uvirtbot | Launchpad bug 1246383 in keystone "Trusts expires_at field is rounded down, doesn't match docs" [Low,New] https://launchpad.net/bugs/1246383 | 21:50 |
| SpamapS | stevebaker: but letting the database handle ordering instead of the clock will produce better results. | 21:50 |
| shardy | ha, snap | 21:50 |
| lifeless | yeah but | 21:50 |
| lifeless | in nova the id is an int | 21:50 |
| lifeless | the uuid is a string | 21:50 |
| lifeless | the id doesn't get exported over the API | 21:51 |
| SpamapS | lifeless: right, we just don't tell the user what the id is. | 21:51 |
| SpamapS | I'm suggesting we do the same in the events table. | 21:51 |
| lifeless | heat seems to have forgotten the basic rule of id is INT | 21:51 |
| lifeless | and instead gone nutty :) | 21:51 |
| *** alexheneveld has joined #heat | 21:55 | |
| sdake_ | rht gangbusters in aftermarket | 21:56 |
| sdake_ | street must have liked our performance | 21:56 |
| *** aweiteka has quit IRC | 21:56 | |
| stevebaker | can we get some reviews on this potential gate race fix? https://review.openstack.org/#/c/62080/ | 21:57 |
| SpamapS | woooot | 22:02 |
| *** giulivo has quit IRC | 22:02 | |
| SpamapS | http://paste.openstack.org/show/55553 | 22:02 |
| SpamapS | ^^ update retrying a failed create | 22:03 |
| *** tomek_adamczewsk has joined #heat | 22:03 | |
| *** slong has joined #heat | 22:04 | |
| asalkeld | stevebaker, interested in removing your -2 from: https://review.openstack.org/#/c/59685/ | 22:10 |
| SpamapS | you guys missed an extra not-needed upcall to setUp | 22:10 |
| SpamapS | I'll do a follow-up patch | 22:10 |
| chmouel | hello, how much ram should i need on my vm to test the cfn/fedora image? | 22:11 |
| asalkeld | chmouel, m1.small | 22:12 |
| SpamapS | whats an m1? | 22:12 |
| asalkeld | flavor | 22:12 |
| asalkeld | | 2 | m1.small | 2048 | 20 | 0 | | 1 | 1.0 | True | | 22:13 |
| * SpamapS is trolling | 22:13 | |
| asalkeld | m1.tiny can work if you don't have cfn-init in there | 22:13 |
| *** tomek_adamczewsk has quit IRC | 22:13 | |
| asalkeld | cfn-init does a yum makecache that blows the disk | 22:14 |
| stevebaker | asalkeld: removed, I'll review if I get the chance | 22:14 |
| *** Tross has quit IRC | 22:14 | |
| asalkeld | sure (it's just discouraging others from reviewing) | 22:15 |
| openstackgerrit | Clint "SpamapS" Byrum proposed a change to openstack/heat: Remove unnecessary setUp call https://review.openstack.org/63225 | 22:16 |
| *** arbylee has quit IRC | 22:17 | |
| *** rcleere has quit IRC | 22:20 | |
| *** urulama has joined #heat | 22:25 | |
| SpamapS | asalkeld: did we find a way around the upgrade problem? | 22:25 |
| asalkeld | (other that changing the charset?) | 22:25 |
| asalkeld | than | 22:26 |
| *** rpothier has quit IRC | 22:26 | |
| asalkeld | I can re-test my old patch, one moment | 22:26 |
| chmouel | asalkeld:nice thank you, seem to work if i increase my vm size from 1gb to 3gb :) | 22:27 |
| *** swygue has quit IRC | 22:28 | |
| asalkeld | chmouel, cool | 22:28 |
| asalkeld | SpamapS, that old patch of mine does not fix the issue | 22:29 |
| asalkeld | just re-tested | 22:29 |
| chmouel | trying to fix this bug https://bugs.launchpad.net/heat/+bug/1261883 but i need some proper events first :) | 22:29 |
| uvirtbot | Launchpad bug 1261883 in heat "event-list is in random order" [High,Triaged] | 22:29 |
| *** gokrokve has quit IRC | 22:29 | |
| *** gokrokve has joined #heat | 22:30 | |
| *** urulama has quit IRC | 22:30 | |
| asalkeld | chmouel, you could just start an empty stack | 22:30 |
| *** swygue has joined #heat | 22:30 | |
| asalkeld | you really don't need a server | 22:30 |
| asalkeld | why not a stack with only a LaunchConfig | 22:31 |
| asalkeld | or a whole bunch of them | 22:31 |
| chmouel | humm yeah, i was trying to get as much real life as possible | 22:31 |
| chmouel | and i need to get experience on heat | 22:31 |
| asalkeld | that will generate events | 22:31 |
| chmouel | so that's a good excuxe :) | 22:31 |
| asalkeld | ok, cool | 22:31 |
| SpamapS | also its good to have events that are separated by a few seconds | 22:32 |
| SpamapS | chmouel: how are you proposing to fix it? | 22:32 |
| chmouel | SpamapS: something like this? http://pastie.org/pastes/8559490/text?key=kyye29jqy9q06b0tzudhdw | 22:32 |
| SpamapS | chmouel: we were just discussing adding back in an incrementing ID. | 22:33 |
| *** arbylee has joined #heat | 22:33 | |
| *** pablosan has quit IRC | 22:33 | |
| SpamapS | chmouel: that will not produce an accurate order | 22:34 |
| *** gokrokve has quit IRC | 22:34 | |
| SpamapS | chmouel: it will just limit the random ordering to events in the same single second | 22:34 |
| chmouel | SpamapS: ah :( | 22:34 |
| chmouel | i guess as you said the incrementing ID would make things work then | 22:34 |
| asalkeld | note: i have a patch series to send the events via the notification system | 22:35 |
| SpamapS | asalkeld: send them where? | 22:35 |
| asalkeld | so first send them via notify() | 22:35 |
| asalkeld | I'll post a wip (easier than discribing) | 22:36 |
| SpamapS | ohh | 22:37 |
| SpamapS | no I think I understand | 22:37 |
| openstackgerrit | Angus Salkeld proposed a change to openstack/heat: WIP: Send events via the notification system https://review.openstack.org/63230 | 22:37 |
| SpamapS | asalkeld: ok, so in oslo I see log, rpc, and noop | 22:37 |
| chmouel | asalkeld: nice | 22:37 |
| SpamapS | asalkeld: nothing user facing unfortunately. :( | 22:37 |
| asalkeld | cfg.CONF.set_default(name='notification_driver', | 22:37 |
| asalkeld | 29 default=['heat.engine.event']) | 22:37 |
| asalkeld | so SpamapS all this goes to ceilometer | 22:38 |
| asalkeld | and ceilometer is moving to an "event" style api | 22:38 |
| *** nati_ueno has joined #heat | 22:38 | |
| SpamapS | asalkeld: Aha! | 22:39 |
| asalkeld | so we could at some point change "heat event-list" to actually call the ceilometer api | 22:39 |
| asalkeld | and all this is soon to end up in horizon | 22:39 |
| *** vijendar has quit IRC | 22:39 | |
| *** jasond`` has quit IRC | 22:39 | |
| SpamapS | asalkeld: sounds good because then every set of notifications that we want to show to users and let users store is in ceilometer. | 22:39 |
| asalkeld | so you can see your "events/notifications" | 22:40 |
| asalkeld | yip | 22:40 |
| SpamapS | asalkeld: however, right now, I don't need ceilometer for Heat. Will this be an alternate backend or are we going to tie Heat to ceilometer in a more aggressive way? | 22:40 |
| SpamapS | because I do actually need ordered events now | 22:40 |
| asalkeld | so SpamapS atm that patch will not change anything | 22:41 |
| chmouel | SpamapS: is storing events can be optional? | 22:41 |
| SpamapS | note that I've taken to piping to sort -t\| -k6 | 22:41 |
| asalkeld | chmouel, exactly | 22:41 |
| asalkeld | by default store the events as normal | 22:41 |
| SpamapS | which is what chmouel's suggested fix would do | 22:41 |
| *** swygue has quit IRC | 22:41 | |
| asalkeld | but also send to ceiloometer | 22:41 |
| asalkeld | in the future when ceilometer and horizon is ready | 22:42 |
| SpamapS | asalkeld: ahh, ok cool. I think this is an alternative-backend-driver type of solution, so less "but also ceilometer" and more "store and retrieve in sql, or store and retrieve from ceilometer or sql" | 22:42 |
| asalkeld | we can just remove that default event store notification handler | 22:42 |
| SpamapS | I'd just let it alone. | 22:43 |
| SpamapS | standalone people will thank you :) | 22:43 |
| asalkeld | yeah, my patch just gives you more options | 22:43 |
| SpamapS | have it work like keystone token drivers | 22:43 |
| SpamapS | asalkeld: there is the question of how ceilometer knows what to do with your notifications, but I presume you're involved with making that "awesome" :) | 22:44 |
| asalkeld | well it just stores | 22:44 |
| asalkeld | per resource/project/user | 22:45 |
| asalkeld | so you can see all your notifications | 22:45 |
| asalkeld | so you could say: what are all the notifications for my stack | 22:45 |
| SpamapS | asalkeld: and there is something that makes them "heat+stackid" so that heat event-list can grab just those. | 22:45 |
| SpamapS | ? | 22:45 |
| SpamapS | asalkeld: sounds like ceilometer is going to be logging as a service | 22:46 |
| SpamapS | though I know these are less verbose than logs | 22:46 |
| asalkeld | SpamapS, well I am not driving that | 22:46 |
| SpamapS | it would actually make sense | 22:46 |
| asalkeld | but if that is the way, then I'll use it | 22:46 |
| SpamapS | like a log != an event.. but having the ability to shove verbose logs in during an event is useful. | 22:47 |
| asalkeld | it's notification-aas | 22:47 |
| SpamapS | asalkeld: anyway, neat-o | 22:47 |
| asalkeld | cool | 22:47 |
| *** gokrokve has joined #heat | 22:49 | |
| SpamapS | chmouel: sorry we got off on a tangent. Did you need help building a template for testing? | 22:52 |
| chmouel | SpamapS: oh no worries, i am fighting with my nova on my vm but all good new stuff to learn | 22:52 |
| chmouel | SpamapS: thanks | 22:52 |
| chmouel | SpamapS: and i thing it's getting late for me so will see that tomorrow | 22:53 |
| SpamapS | chmouel: cool. I really want this bug fixed so let me know if there is anything I can do to help. | 22:53 |
| chmouel | SpamapS: cool let me give a try tomo morning (EU time) and if i can't i may unassigned myself to someone who know the code better | 22:54 |
| *** gokrokve has quit IRC | 22:54 | |
| SpamapS | chmouel: ok, nothing like a good old fashioned bug fix to learn. :) | 22:56 |
| chmouel | SpamapS: i can probably help on the swift sides do you know if there is some features requests there? | 22:57 |
| chmouel | SpamapS: i am thinking to add tempurl key for a start since that may be useful for a heat use case | 22:57 |
| SpamapS | chmouel: No I haven't heard much about that but we do have swift resources so reviewing them / filling out their docs / features would be useful. | 22:58 |
| SpamapS | chmouel: I am positive tempurls would be very useful | 22:58 |
| chmouel | SpamapS: ok great, thanks again and that would be the time for me to logoff | 22:58 |
| stevebaker | its a shame ceph doesn't support tempurls | 23:08 |
| SpamapS | http://tracker.ceph.com/issues/3454 | 23:09 |
| *** yogesh has quit IRC | 23:10 | |
| *** harlowja is now known as harlowja_away | 23:12 | |
| stevebaker | good to see | 23:15 |
| *** arbylee has quit IRC | 23:18 | |
| *** andersonvom has quit IRC | 23:21 | |
| *** rpothier has joined #heat | 23:29 | |
| openstackgerrit | Angus Salkeld proposed a change to openstack/heat: Add a db check for CHARSET=utf8 https://review.openstack.org/63241 | 23:34 |
| *** nati_uen_ has joined #heat | 23:34 | |
| *** nati_uen_ has quit IRC | 23:34 | |
| *** nati_uen_ has joined #heat | 23:35 | |
| *** nkhare has joined #heat | 23:35 | |
| *** Tross has joined #heat | 23:36 | |
| *** nati_ueno has quit IRC | 23:37 | |
| *** ccs has quit IRC | 23:42 | |
| stevebaker | asalkeld: I get this quite frequently but have not yet found how to reproduce it http://paste.openstack.org/show/55566/ | 23:42 |
| asalkeld | that's with the current code? | 23:43 |
| asalkeld | (not my patch) | 23:43 |
| stevebaker | yep | 23:43 |
| *** rongze has joined #heat | 23:43 | |
| asalkeld | mmm, that's not good | 23:44 |
| asalkeld | stevebaker, not sure what to make of that :( | 23:46 |
| stevebaker | me neither | 23:47 |
| stevebaker | I'll raise a bug | 23:47 |
| *** rongze has quit IRC | 23:48 | |
| *** slong has quit IRC | 23:49 | |
| stevebaker | hang on, does devstack create a new encrypt secret on every run? | 23:50 |
| asalkeld | maybe? | 23:51 |
| stevebaker | that will be it. unstack.sh has stopped killing heat-engine so the old heat-engine gets load-balanced until I notice and kill it | 23:53 |
| *** harlowja_away is now known as harlowja | 23:57 | |
| *** gokrokve has joined #heat | 23:59 | |
| *** ccs has joined #heat | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
