Tuesday, 2016-07-12

*** toddjohn has joined #openstack-mistral00:21
*** toddjohn has quit IRC00:32
*** gyee has quit IRC00:38
*** bobh has joined #openstack-mistral00:40
*** toddjohn has joined #openstack-mistral00:51
*** toddjohn has quit IRC00:55
*** bobh has quit IRC01:02
*** cheneydc has joined #openstack-mistral01:04
*** toddjohn has joined #openstack-mistral01:07
*** toddjohn_ has joined #openstack-mistral01:11
*** toddjohn has quit IRC01:15
*** toddjohn has joined #openstack-mistral01:19
*** toddjohn_ has quit IRC01:22
*** toddjohn has quit IRC01:25
*** toddjohn has joined #openstack-mistral01:25
*** toddjohn has quit IRC01:36
*** toddjohn has joined #openstack-mistral01:40
*** toddjohn has quit IRC01:41
*** toddjohn has joined #openstack-mistral01:43
*** toddjohn has quit IRC02:03
*** cheneydc has quit IRC02:04
*** toddjohn has joined #openstack-mistral03:04
*** cheneydc has joined #openstack-mistral03:08
*** toddjohn has quit IRC03:09
*** ramishra has quit IRC03:24
*** ramishra has joined #openstack-mistral03:26
*** ramishra has quit IRC04:04
*** ramishra has joined #openstack-mistral04:07
*** Ravikiran_K has quit IRC04:13
*** ramishra has quit IRC04:19
*** ramishra has joined #openstack-mistral04:21
openstackgerritShaik Apsar proposed openstack/mistral: Fix for YaqlEvaluationException in std.create_instance workflow.  https://review.openstack.org/34042104:30
*** ramishra has quit IRC04:31
*** ramishra has joined #openstack-mistral04:31
rakhmerovhparekh: here? Can you pls review https://review.openstack.org/#/c/34042104:32
rakhmerovwe overlooked a nasty bug in a standard workflow04:32
hparekhrakhmerov, yeah ok sure04:33
rakhmerovit's super simple04:33
*** ramishra has quit IRC04:41
*** ramishra has joined #openstack-mistral04:43
rakhmerovddeja: hi, can you please look at https://bugs.launchpad.net/mistral/+bug/1602003 ?04:47
openstackLaunchpad bug 1602003 in Mistral "No such option in group DEFAULT: rpc_backend" [Critical,New] - Assigned to Dawid Deja (dawid-deja-0)04:47
rakhmerovmy guess is that it was introduced by one of your recent patches04:47
rakhmerovrpc_utils.py has the line:04:48
rakhmerovif CONF.rpc_backend in ['rabbit', 'fake']:04:48
openstackgerritRenat Akhmerov proposed openstack/mistral-specs: Fixing typos in Custom API spec  https://review.openstack.org/34028204:51
*** ramishra has quit IRC05:00
*** d0ugal has quit IRC05:01
*** ramishra has joined #openstack-mistral05:01
*** toddjohn has joined #openstack-mistral05:06
*** toddjohn has quit IRC05:10
openstackgerritMerged openstack/mistral: Fix for YaqlEvaluationException in std.create_instance workflow.  https://review.openstack.org/34042105:13
*** ramishra has quit IRC05:35
*** Ravikiran_K has joined #openstack-mistral05:54
*** d0ugal has joined #openstack-mistral06:04
*** shardy has joined #openstack-mistral06:17
*** cheneydc has quit IRC06:29
*** d0ugal has quit IRC06:43
Ravikiran_KHi06:44
Ravikiran_Ki have a question, what are the things that Mistral do where Heat can not?06:46
*** shardy has quit IRC07:01
rakhmerovRavikiran_K: workflow07:04
rakhmerovdistributed workflow07:04
Ravikiran_Kok.... can we replace Heat with Mistral?07:06
*** toddjohn has joined #openstack-mistral07:07
rakhmerovwho do you mean by "we"?07:08
rakhmerovRavikiran_K: please give us more background07:08
rakhmerovwhat are you trying to solve etc.07:09
rakhmerovHeat and Mistral are completely different technologies07:09
rakhmerovI just don't know what specifically you're interested in07:09
rakhmerovneed more info07:09
*** toddjohn has quit IRC07:12
*** FL1SK has quit IRC07:13
mfloboRavikiran_K, I'd say that the main difference would be the CRON concept that Mistral has07:20
mfloboRavikiran_K, I would say that Mistral is not a replacement of Heat07:21
mfloboRavikiran_K, but as rakhmerov said, if you provide more info about your use case maybe we can help more07:22
rakhmerovCron is only a small part of it07:23
Ravikiran_Ki am just trying to understand the difference between the heat and mistral07:23
Ravikiran_Kwhat are the things that heat is doing can achieve with mistral right?07:25
rakhmerovin Heat you describe resources, the state of the environment that you want to get at the end07:25
Ravikiran_Kok07:25
rakhmerovin Mistral you describe processes07:25
rakhmerovmultiple steps that go one after another07:25
rakhmerovor in parallel07:25
rakhmerovaccording to some predefined logic07:26
rakhmerovin some tasks you don't have even a notion of a certain final state so you can't use Heat07:26
rakhmerovthe process itself may be more important07:26
rakhmerovHeat is only a deployment tool, that's it07:27
rakhmerovand it provides some LCM (life-cycle management) capabilities07:27
rakhmerovwhereas Mistral is a more generic tool for managing distributed processes07:27
rakhmerovworkflows07:27
rakhmerovit's applicable when you need to automate something and it will be a long running process07:28
rakhmerovthat needs to be scalable, highly available07:28
rakhmerovwhich should have a state so you can observe it07:28
rakhmerovyou can pause/resume them, recover from errors manually and continue07:28
rakhmerovyou can use it for tons of different use cases: deployment (special case that Heat handles), monitoring, auto-scaling, auto-healing (reacting on certain monitoring events and automatically applying healing actions)07:30
rakhmerovcoordination between multiple VIMs (sites)07:30
rakhmerovDisaster Recovery, replication07:30
rakhmerovetc. etc.07:30
rakhmeroveven simpler: imaging that you need to run a process that will upgrade some software on all of the machines in you data centre where some hosts need to be updated only after some certain software on other hosts are upgraded, how would you approach it? Say it will take a week to do07:32
rakhmerovyou have a choice: write bash/python/java whatever07:32
rakhmerovif you do so then when you run it you won't know at this stage your process is07:33
rakhmerovand if the machine that is running this script dies, you'll loose everything, all process state07:33
rakhmerovalternatively you can write a workflow and let a workflow engine manage execution of this process07:34
rakhmerovwhich will guarantee that you'll be able to see the state of the process, if some of the components die it will still continue (HA)07:34
rakhmerovif an error happens you'll be able to manually fix the problem and continue from that exact place07:35
rakhmerovdoes that make sense?07:35
Ravikiran_Kyes07:35
rakhmerovok :)07:38
Ravikiran_Krakhmerov: thanks for the info..i am just exploring the mistral capabilities in cloud environment07:40
rakhmerovok, sure07:40
rakhmerovif you have some specific questions pls let us know07:40
Ravikiran_Ksure :)07:41
openstackgerritMarcos Fermín Lobo proposed openstack/mistral: Allow to use both name and id to access action definitions  https://review.openstack.org/32589408:00
therverakhmerov, I know where in mistral channel, but Heat is not "just a deployment tool" :)08:05
rakhmerovdeployment and LCM08:06
rakhmerovwhat else?08:06
therveThe scaling and healing part exists in Heat as well08:07
rakhmerovyes, that fits into LCM to me08:07
therveOkay08:08
rakhmerovscaling and healing of resources deployed by Heat08:08
rakhmerovtherve: don't get me wrong, I know that Heat can do a lot and do well08:08
*** toddjohn has joined #openstack-mistral08:08
rakhmerovI'm just trying to draw some boundaries08:09
rakhmerovscope where these technologies may be applied08:09
rakhmerovin the best way08:09
therveYeah I don't think they overlap very much08:09
rakhmerovtrue08:09
rakhmerovbut for some reason I find it difficult to explain sometimes08:10
rakhmerovnew people are inclined to confuse them, I really don't understand why08:10
therveWell because you can use one to do stuff that the other does as well08:11
therveansible VS puppet if you want :)08:11
rakhmerovyes, right08:12
rakhmerovhm.. yeah, we definitely need to do a better job when documenting all this08:13
*** toddjohn has quit IRC08:14
*** jpich has joined #openstack-mistral08:15
*** jistr has joined #openstack-mistral08:22
openstackgerritRenat Akhmerov proposed openstack/mistral-extra: Restructure mistral-extra repo  https://review.openstack.org/34033408:57
ddejarakhmerov: Hi. About https://bugs.launchpad.net/mistral/+bug/160200309:24
openstackLaunchpad bug 1602003 in Mistral "No such option in group DEFAULT: rpc_backend" [Critical,New] - Assigned to Dawid Deja (dawid-deja-0)09:24
ddejaI cannot reproduce this09:24
rakhmerovddeja: hi, did you try to run api and engine separately as he explained?09:24
ddejaI always run it that way09:25
ddejawhat is more09:25
rakhmerovook..09:25
ddejaoption rpc_backend is one of defult options that oslo automatically adds to config09:25
ddejahttps://wiki.openstack.org/wiki/Oslo/Messaging#Configuration09:25
rakhmerovhm.. then can you pls put it into "incomplete" and ask him for more info?09:25
rakhmerovyeah, I know09:25
rakhmerovright09:25
rakhmerovbut this is the only place that seemingly could be a reason09:26
rakhmerovok, no problem09:26
rakhmerovwe need to ask him for more info (config, stack trace etc.)09:26
ddejayup09:26
rakhmerovthanks Dawid09:28
ddejarakhmerov: No problem :)09:29
rakhmerovand one question here: https://review.openstack.org/#/c/340450/09:29
rakhmerovon one hand we kind of change the default here, but on the other hand "at-least-once" didn't work properly so far so it didn't really make sense09:30
rakhmerovanyway, I'm just trying to understand what the good default should be09:31
ddejarakhmerov: this don't change default09:31
rakhmerovwhy?09:31
ddejawith o.m. default option is to never rerun task09:31
rakhmerovwell, effectively yes, you're right09:31
rakhmerovyes09:31
rakhmerovright09:31
rakhmerovanyway, let's just think what it should be by default09:32
rakhmerovTrue or False09:32
rakhmerovI also think that we can configure this default in cfg09:32
rakhmerovit may depend on certain installation09:32
ddejarakhmerov: right. I was also thinking to do so09:32
rakhmerovremember we discussed it in Austin09:32
rakhmerovit can be your next patch )09:32
rakhmerovno need to add into this one09:33
ddejawhat is more, I recently find out that rabbit have one super feature09:33
rakhmerovwhich one?09:33
ddejaevery message have flag 'resended'09:33
rakhmerovoooh!09:33
ddeja(or something similar, not remember extact name)09:33
rakhmerovand it can be extracted by executor09:33
ddejayup09:33
rakhmerov:)09:33
ddejaso we can rerun action or fail it09:33
ddejaso no more workflows stuck in running state09:34
rakhmerovyeah, that is super awesome09:35
rakhmerovok, seems like you have even more interesting work here :))09:35
rakhmerovddeja: btw, I was recently thinking about "executor|api|engine -> engine" communications09:36
ddejarakhmerov: Yup. And some reviews to do :)09:36
ddejarakhmerov: hm?09:36
openstackgerritMerged openstack/mistral-specs: Fixing typos in Custom API spec  https://review.openstack.org/34028209:36
rakhmerovit has pretty much the same problem with acknowledgements, right?09:36
ddejayup09:37
rakhmerovlike, for example, executor sends a result over MQ to engine09:37
rakhmerovand engine dies after it polled the message but before TX is committed09:37
rakhmerovbut in this case it seems like we can always safely resend messages09:38
rakhmerovI'm not sure on 100% but seems like it's true09:38
ddejaI think so, but it should be double-checked09:38
rakhmerovbecause: 1) DB transactions guarantee data integrity 2) All our operations seem to be idempotent09:38
rakhmerovfor example, handling action result second time won't just do anything bad09:39
ddejaI guees so :)09:39
rakhmerovit's just something we should think about but I think we're good here..09:39
*** shardy has joined #openstack-mistral09:40
ddejatotaly agree09:41
*** d0ugal has joined #openstack-mistral10:03
*** toddjohn has joined #openstack-mistral10:11
*** toddjohn has quit IRC10:15
*** FL1SK has joined #openstack-mistral10:17
*** venkat has joined #openstack-mistral10:37
*** mflobo has quit IRC11:00
*** mflobo has joined #openstack-mistral11:02
*** dprince has joined #openstack-mistral12:27
*** toddjohn has joined #openstack-mistral12:45
*** toddjohn has quit IRC13:10
*** toddjohn has joined #openstack-mistral13:11
*** toddjohn has quit IRC13:15
*** tonytan4ever has joined #openstack-mistral13:27
*** jpich_ has joined #openstack-mistral13:29
*** d0ugal_ has joined #openstack-mistral13:30
*** jpich has quit IRC13:32
*** d0ugal has quit IRC13:33
*** d0ugal_ is now known as d0ugal13:35
*** d0ugal is now known as Guest6571513:36
*** Guest65715 has quit IRC13:36
*** d0ugal_ has joined #openstack-mistral13:36
*** d0ugal_ has quit IRC13:44
*** d0ugal has joined #openstack-mistral13:44
*** toddjohn has joined #openstack-mistral13:46
*** bobh has joined #openstack-mistral13:48
*** bobh has quit IRC13:54
*** tonytan_brb has joined #openstack-mistral14:17
*** tonytan4ever has quit IRC14:20
*** venkat has quit IRC14:23
*** toddjohn_ has joined #openstack-mistral14:31
*** toddjohn has quit IRC14:32
*** tonytan4ever has joined #openstack-mistral14:36
*** tonytan_brb has quit IRC14:39
*** Ravikiran_K has quit IRC15:08
*** vishwanathj has joined #openstack-mistral15:19
*** gyee has joined #openstack-mistral15:55
*** rrecio has joined #openstack-mistral16:01
*** dprince has quit IRC16:06
*** tonytan_brb has joined #openstack-mistral16:07
*** mflobo has quit IRC16:08
*** tonytan4ever has quit IRC16:09
*** toddjohn_ has quit IRC16:13
*** toddjohn has joined #openstack-mistral16:14
*** toddjohn has quit IRC16:14
*** toddjohn has joined #openstack-mistral16:21
*** d0ugal has quit IRC16:24
*** jpich_ has quit IRC16:24
*** toddjohn has quit IRC16:25
*** mflobo has joined #openstack-mistral16:26
*** ninag has joined #openstack-mistral16:27
*** catintheroof has joined #openstack-mistral16:28
*** ninag has quit IRC16:32
*** mgershen has quit IRC16:54
*** dprince has joined #openstack-mistral16:57
*** toddjohn has joined #openstack-mistral17:09
*** toddjohn has quit IRC17:10
*** toddjohn has joined #openstack-mistral17:10
*** dprince has quit IRC17:41
*** dprince has joined #openstack-mistral17:43
*** tonytan_brb has quit IRC17:54
*** toddjohn has quit IRC18:16
*** toddjohn has joined #openstack-mistral18:17
*** tonytan4ever has joined #openstack-mistral18:20
*** shardy is now known as shardy_afk18:38
*** toddjohn has quit IRC19:04
*** toddjohn has joined #openstack-mistral19:06
*** toddjohn has quit IRC19:06
*** toddjohn has joined #openstack-mistral19:06
*** vishwanathj has quit IRC19:12
*** vishwanathj has joined #openstack-mistral19:13
*** gyee has quit IRC19:18
*** tonytan4ever has quit IRC20:08
*** toddjohn has quit IRC20:08
*** shardy_afk is now known as shardy20:44
*** gyee has joined #openstack-mistral20:48
*** clenimar has joined #openstack-mistral20:55
*** tonytan4ever has joined #openstack-mistral21:20
*** tonytan4ever has quit IRC21:26
*** dprince has quit IRC21:36
*** shardy has quit IRC21:38
*** toddjohn has joined #openstack-mistral21:53
*** toddjohn has quit IRC21:58
*** rrecio has quit IRC22:34
*** rrecio has joined #openstack-mistral22:35
*** tonytan4ever has joined #openstack-mistral22:37
*** rrecio has quit IRC22:39
*** tonytan4ever has quit IRC22:42
*** vishwanathj has quit IRC22:57
*** vishwanathj has joined #openstack-mistral22:58
*** vishwanathj has quit IRC23:02
*** vishwanathj has joined #openstack-mistral23:03
*** rrecio has joined #openstack-mistral23:07
*** vishwanathj has quit IRC23:07
*** vishwanathj has joined #openstack-mistral23:08
*** vishwanathj has quit IRC23:08
*** vishwanathj has joined #openstack-mistral23:09
*** catintheroof has quit IRC23:09
*** bobh has joined #openstack-mistral23:24
*** vishwanathj has quit IRC23:30
*** vishwanathj has joined #openstack-mistral23:31
*** rrecio has quit IRC23:33
*** vishwanathj has quit IRC23:36
*** vishwanathj has joined #openstack-mistral23:36
*** bobh has quit IRC23:38
*** toddjohn has joined #openstack-mistral23:44

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!