Friday, 2016-07-01

*** thorst_ has quit IRC00:00
*** michaelgugino has quit IRC00:08
*** thorst_ has joined #openstack-watcher00:59
*** thorst_ has quit IRC01:07
*** thorst_ has joined #openstack-watcher01:23
*** thorst_ has quit IRC01:24
*** jinquan has joined #openstack-watcher01:36
*** thorst_ has joined #openstack-watcher02:19
*** thorst_ has quit IRC02:20
*** thorst_ has joined #openstack-watcher02:20
*** thorst_ has quit IRC02:29
*** diga has quit IRC02:46
*** jwcroppe has quit IRC02:55
*** thorst_ has joined #openstack-watcher03:35
*** thorst_ has quit IRC03:44
*** diga has joined #openstack-watcher03:46
openstackgerritEdwin Zhai proposed openstack/watcher: Enable strategy parameters
openstackgerritEdwin Zhai proposed openstack/watcher: Documentation for strategy parameters
openstackgerritEdwin Zhai proposed openstack/python-watcherclient: Enable strategy parameter
*** jwcroppe has joined #openstack-watcher04:08
*** jwcroppe has quit IRC04:10
digaHi acabot, jed5604:29
digaI am commiting the change but it is asking me password04:30
digafatal: Authentication failed for '
digadigambar@pt-acceng4567:~/watcher$ git review04:30
digaPassword for '':04:30
digafatal: Authentication failed for ''04:30
digabut I think there is no password needed for commit04:31
*** thorst_ has joined #openstack-watcher04:43
*** thorst_ has quit IRC04:49
digaHi All05:31
digahas anyone faced this problem, now password error has gone but its asking passphrase05:32
*** jwcroppe has joined #openstack-watcher05:40
*** thorst_ has joined #openstack-watcher05:46
*** thorst_ has quit IRC05:53
*** jwcroppe has quit IRC06:04
*** thorst_ has joined #openstack-watcher06:52
*** jed56 has joined #openstack-watcher06:54
*** thorst_ has quit IRC06:59
*** alexchadin has joined #openstack-watcher07:05
alexchadinjed56: hi :)07:05
alexchadinjed56: how are you?07:06
jed56hi alexchadin07:06
openstackgerritDigambar proposed openstack/watcher: Fix field type to audit_type
*** ostroverkhov has joined #openstack-watcher07:06
jed56very well thanks and you ?07:06
alexchadinjed56: I'm fine07:07
alexchadinI had discussion with ostroverkhov about period field in watcher/w-client/w-dashboard07:08
openstackgerritMuzammil proposed openstack/watcher: removed column parameters from the following methods in both files     def get_audit_template_list(self, context, columns=None, filters=None,                                 limit=None, marker=None, sort_key=None,                                 sort_dir=
alexchadinThe Idea is that ONESHOT audits don't have to get period set07:09
alexchadinShould be None07:09
alexchadinFor continuous audits this field should be set or 3600 as default07:10
alexchadinCan't we change audit types after it is created?07:10
jed56We can't change the audit type after his creation07:16
alexchadinwhen we create ONESHOT audit07:17
alexchadinperiod field should be None07:17
jed56yes for ONESHOT the period should be None07:17
alexchadinand in DB this field also should be None07:17
jed56sadly, we can't use inheritance with the watcher objects07:18
jed56if it was the case, we could create Continuous Audit which extends Audit in the db07:18
alexchadinIf we set period to 007:21
alexchadincould it be translated as None to API?07:21
alexchadincause ostroverkhov have some problems with IntegerField in w-dashboard07:22
alexchadinIt can't be set to None, only integer values07:22
*** dtardivel has joined #openstack-watcher07:24
dtardivelostroverkhov: Hi, I would like to review the changeset "Add support continuously-optimizatiion" on watcher-dashbard.07:25
dtardivelostroverkhov: you replied Done on last comments from vincentfrancoise, but you did not push another PS. Are you working on it ?07:25
jed56alexchadin:  The period can be set as None => 'period': obj_utils.int_or_none,07:26
ostroverkhovdtardivel: Hi, yes I'm working on it now.07:27
dtardivelostroverkhov: ok, great !!!07:27
alexchadinjed56: It can in watcher object07:27
alexchadinjed56: take a look
alexchadin55 line07:28
alexchadindtardivel: hi:)07:29
alexchadindtardivel: how are you?07:29
dtardivelalexchadin: fine :)07:29
jed56alexchadin: I'm not sure for the dashboard07:29
dtardivelHI Alex07:29
alexchadindtardivel: could you please also review it?
alexchadindtardivel: if you have a time, of course07:30
jed56alexchadin: you can't set None ? forms.IntegerField(initial=None,07:31
*** vincentfrancoise has joined #openstack-watcher07:33
dtardivelalexchadin: I would like to review this feature globally with the dashboard changeset :) But It is in my plan ....07:33
*** openstackgerrit has quit IRC07:33
*** openstackgerrit has joined #openstack-watcher07:33
alexchadindtardivel: okay :)07:34
ostroverkhovjed56: I have already fixed it07:38
jed56ostroverkhov: great07:39
*** alexchadin has quit IRC07:42
*** ostroverkhov has quit IRC07:45
*** thorst_ has joined #openstack-watcher07:57
*** thorst_ has quit IRC08:04
openstackgerritDigambar proposed openstack/watcher: Fix field type to audit_type
digajed56: Hi08:08
digacan you check my above comming08:08
*** openstackgerrit has quit IRC08:18
*** openstackgerrit has joined #openstack-watcher08:18
dtardivelzigo: Hi, I have a bug on installing of the  watcher-api Debian experimental package. I have started a Docker container based on the debian:experimental image, and I tried to install into this container the watcher-api.08:35
zigodtardivel: What is the problem exactly?08:35
dtardivelzigo: In the configuration step, I have this error:08:35
zigodtardivel: Can you add a "set -x" in /var/lib/dpkg/info/watcher-api.postinst, and run "apt-get -f install", so that we see what happens?08:37
zigoOh, sorry.08:37
zigoGot it above.08:37
zigoWell, it needs /sbin/route08:37
dtardivelzigo: /sbin/route binary is not installed on Debian experimental container08:37
zigoapt-get install net-tools08:37
zigoI'll add it as a hard dependency.08:38
dtardivelzigo: great, thanks08:38
zigodtardivel: Fix uploaded, however I just missed the dak run, so it will take a few hours until it reaches Debian.08:52
*** thorst_ has joined #openstack-watcher09:02
*** vtech has joined #openstack-watcher09:06
*** jwcroppe has joined #openstack-watcher09:08
*** thorst_ has quit IRC09:09
*** jwcroppe has quit IRC09:13
*** jwcroppe has joined #openstack-watcher09:18
openstackgerritTomasz Kaczynski proposed openstack/watcher: Add scoring engines to database and API layers
digajed56: py27 jenkin job is failing, dont know how this file is checked in earlier09:44
digabecause the error it is showing are there in the file before this patch09:45
vincentfrancoisehi diga09:45
digavincentfrancoise: hey09:45
digavincentfrancoise: can go through this patch -
vincentfrancoisedid you try to run the tests locally?09:47
vincentfrancoise(to see if there are any discrepancies between you running it locally I the failing results we see on the gates)09:48
vincentfrancoisediga: I'm looking at the error you got09:51
digaactually those failed errors alraedy existed in the file09:52
digabecause when I downloaded the code, I run some pep8 tests. I was failing there09:53
digaso I dont know who has commited that09:53
vincentfrancoiseHow did you run pep8?09:53
digaI try with flake8 then try pep809:54
vincentfrancoisecan you use the 'tox -e pep8' command as explained in ?09:54
vincentfrancoisebecause we may have gotten errors coming from some error codes we purposefully deactivated09:56
vincentfrancoises/we may/you may09:56
*** jwcroppe has quit IRC09:57
digaI will run it10:03
*** thorst_ has joined #openstack-watcher10:09
*** thorst_ has quit IRC10:14
*** vtech has quit IRC10:31
*** vtech has joined #openstack-watcher10:33
*** jed56 has quit IRC10:35
*** vtech has quit IRC10:38
digavincentfrancoise: Hi11:39
digavincentfrancoise: we need to change the type attribute from audit table & for that we need to add sqlalchemy migration11:40
digabefore going to add it, we need to file seperate dependant for that11:41
digaacabot: vincentfrancoise : anyone there ?11:43
*** jwcroppe has joined #openstack-watcher11:46
vincentfrancoisehi diga11:58
digavincentfrancoise: hey11:58
vincentfrancoiseas for SQL migration11:59
digavincentfrancoise: this change requires db field migration also for audit table11:59
vincentfrancoisewe currently do not do them since we did not reach a 1.0 version yet11:59
vincentfrancoisewe expect DB migrations to be handled around the end of the Newton cycle12:00
digathen If I add/update field name in the table, I need to do db-manage command again to reflect the column in audit table12:01
vincentfrancoisealthough this is to be taken with a grain of salt12:01
vincentfrancoiseyes you will have to recreate your DB12:01
digagot it12:02
digavincentfrancoise: thanks for your immediate reply12:02
vincentfrancoiseyou're welcome ;)12:03
digaI appreciate12:03
vincentfrancoisefeel free to drop questions again :)12:03
digayep :)12:03
digayour geographical region12:03
vincentfrancoiseFrance so UTC+212:04
digaI am from India12:04
*** edleafe is now known as figleaf12:04
vincentfrancoisecool :)12:06
*** jed56 has joined #openstack-watcher12:09
*** jwcroppe has quit IRC12:15
*** jwcroppe has joined #openstack-watcher12:17
*** diga has quit IRC12:18
*** jwcroppe has quit IRC12:21
dtardivelzigo, I have another issue with the watcher-api debian experimental package. It seems the 'Authentication server hostname' (Keystone) is not correctly reported into the Keystone URL use to query token while Watcher endpoints creation.12:25
dtardivelzigo: ^^12:27
tkaczynskihi guys, do you have any idea why my change is failing test in multinode environment? the logs are here:
tkaczynskiit doesn't seem to be related to my changes whatsoever12:34
vincentfrancoisethere are 2 or 3 tests that are currently failing in the multinode env12:34
vincentfrancoisea bug was created12:35
vincentfrancoisebut no one fixed it yet:
openstackLaunchpad bug 1595534 in watcher "Fix failing Watcher Tempest tests" [Medium,Triaged]12:35
tkaczynskishouldn't they be disabled then until the bug is fixed?12:35
vincentfrancoisethe fix is very minimal, the thing is that whoever gets close to it will fix it fairly quickly12:36
tkaczynskiok. so I shouldn't worry about my change then :)12:36
vincentfrancoiseit's a known issue so yeah, no worries :)12:38
tkaczynskicheers :)12:40
vincentfrancoisetkaczynski: since you are here12:45
vincentfrancoiseI have a question regarding your code12:45
vincentfrancoisewhy did you make the ID column be a String(50) for the ScoringEngine model?12:46
tkaczynskiwhy is is a string? or why it has 50 chars?12:46
vincentfrancoisekind of both actually: is it meant to be a normal string or a UUID one? if so the others are String(36)12:47
tkaczynskibasically, the idea for the scoring engines is that they are identified by a unique string, but I didn't need the uuid. it should be short, readable by human. the size 50 I chose arbitrarily. not too long, not too short12:49
tkaczynskiI thought that string(50) is used in some other places now, but I did that some time ago and I might not remember correctly. any particular reason to use String(36)?12:50
vincentfrancoiseI'm asking because a couple of months ago I tried to do the same thing and I had many comments on it12:52
vincentfrancoisehere is the review
tkaczynskiso what was the conclusion?12:52
vincentfrancoise to be more precise12:52
vincentfrancoisebasically, people feel that an ID has to be an Integer somehow12:53
*** alexchadin has joined #openstack-watcher12:53
*** vtech has joined #openstack-watcher12:55
tkaczynskiI don't think the world assumed that id in database means integer :)12:55
jed56ah ah ah : )12:55
jed56also agreed but this is the openstack way : )12:56
jed56thanks to oslo_db12:56
tkaczynskiI was thinking about it actually and also about UUID. for my use cases, UUID was completely unnecessary. zero benefits, just work to handle it around. but I've noticed that there are few places, which assume that UUID is present12:57
vincentfrancoisetkaczynski: yes that's the issue12:57
tkaczynskibut I managed to overcame it, everything works for me without UUID12:58
tkaczynskiso what should I do with this id being integer then? should this id be auto-generated by watcher when discovering new scoring engine plugin? should strategy developers rely on this id then?13:00
tkaczynskimy idea was that scoring engine plugin developer picks his unique id and then it can be used everywhere, strategies could rely on it etc.13:00
tkaczynskiand id being a string allows to introduce some naming convention, so there will not be id conflicts etc13:01
vincentfrancoiseI agree, I am simply playing the devil's advocate13:02
tkaczynskiI fully understand and see where you are coming from13:03
vincentfrancoisethe thing is that back then13:03
vincentfrancoiseI had to change the whole thing to ad IDs everywhere, so I don't want you to go through the same experience :p13:04
tkaczynskiquestion is what to do now: do I need to change it? I would really prefer to work on the next part of the scoring module than refactoring the whole thing for no benefits really13:04
vincentfrancoisejed56: what do you think of it?13:05
tkaczynskibtw: is the id field currently autoincremented?13:06
vincentfrancoiseyes it is13:06
tkaczynskiand what is the story with uuid everywhere? if id is so great, do you really need uuid?13:07
jed56we don't really need uuid for most of the use cases13:09
vincentfrancoisethe uuid is because it is a security issue to display it via the API13:09
jed56in restFul Api13:09
vincentfrancoisebecause it is sequential13:10
jed56this is not recommanded that we could guest the uuid13:10
vincentfrancoisetkaczynski: BTW, you have both a Unique constraint and a primary key constraint13:12
vincentfrancoisetkaczynski: I guess you can leave it as-is for now, if someone else complains in the meantime, we will reconsider it then13:13
tkaczynskivincentfrancoise: sounds good. waiting for the review then, I hope to get it merged, because I have quite big change coming next, it's getting harder and harder to maintain 2 branches13:15
vincentfrancoisetkaczynski: I'm well aware of that, I had 15 chained changesets at some point waiting to get merged13:16
vincentfrancoiseit takes one hell of a time to rebase the lot13:17
tkaczynskihonestly, I can't even imagine that :) I would rather implement a bot for these merges, than do it myself ;)13:18
*** alexchadin has quit IRC13:19
*** alexchadin has joined #openstack-watcher13:20
*** jwcroppe has joined #openstack-watcher13:21
*** jwcroppe has quit IRC13:25
jed56vincentfrancoise: you love to rebase patchsets ;)13:26
*** alexchadin has quit IRC13:33
*** alexchadin has joined #openstack-watcher13:34
zigodtardivel: If you don't have the keystone server in localhost, you have to dpkg-reconfigure debconf to set a lower priority, so that it will ask you the IP of your keystone server.13:38
zigodtardivel: Otherwise, just setup the endpoint by yourself, by hand ...13:38
zigoThat works too.13:39
zigoThere's nothing mandatory with the debconf stuff.13:39
dtardivelzigo: ok, thanks, This is what I did (declare the endpoint by myself)13:40
*** vtech has quit IRC14:01
*** danpawlik has quit IRC15:05
*** alexchadin has quit IRC15:06
*** jwcroppe has joined #openstack-watcher15:23
*** jwcroppe has quit IRC15:28
*** brunograz has quit IRC15:54
*** vincentfrancoise has quit IRC15:56
*** jwcroppe has joined #openstack-watcher17:26
*** jwcroppe has quit IRC17:31
*** jed56 has quit IRC17:45
*** dtardivel has quit IRC17:47
*** jwcroppe has joined #openstack-watcher18:25
*** figleaf has quit IRC19:40
*** edleafe has joined #openstack-watcher19:48
*** harlowja has quit IRC21:53
*** harlowja has joined #openstack-watcher21:53
*** harlowja_ has joined #openstack-watcher22:55
*** harlowja has quit IRC22:59
*** dronshaw has quit IRC23:01
*** jwcroppe has quit IRC23:42
*** jwcroppe has joined #openstack-watcher23:49
*** jwcroppe has quit IRC23:53

Generated by 2.14.0 by Marius Gedminas - find it at!