Perl 6 - the future is here, just unevenly distributed

IRC log for #openstack-rally, 2014-10-02

| Channels | #openstack-rally index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:07 marcoemorais joined #openstack-rally
00:17 openstackgerrit Kui Shi proposed a change to stackforge/rally: Add volume option to boot server from volume and run script  https://review.openstack.org/122409
01:31 vkmc joined #openstack-rally
01:31 vkmc joined #openstack-rally
02:38 klindgren joined #openstack-rally
03:10 openstackgerrit joined #openstack-rally
03:18 openstackgerrit joined #openstack-rally
03:54 rvcehimanshu joined #openstack-rally
03:59 rvcehimanshu joined #openstack-rally
04:03 openstackgerrit Oleh Anufriiev proposed a change to stackforge/rally: Add detailed description for rally commands  https://review.openstack.org/124910
04:23 oanufriev joined #openstack-rally
06:23 ekarlso joined #openstack-rally
06:23 ekarlso oii
06:26 ekarlso boris-42: ? :)
06:28 boris-42 ekarlso so
06:28 boris-42 ekarlso gates doesn't work
06:29 ekarlso boris-42: why ?
06:29 boris-42 ekarlso https://review.openstack.org/#/c/123442/
06:29 boris-42 ekarlso gate-rally-dsvm-designate-rally
06:30 boris-42 ekarlso http://logs.openstack.org/42/123442/2/experimental/gate-rally-dsvm-designate-rally/50a6d2c/console.html
06:31 boris-42 ekarlso 2014-10-01 10:38:36.314 | 2014-10-01 10:38:36.254 | 2014-10-01 10:38:36.203 29704 TRACE rally     self.clients.designate().quotas.update(tenant_id, kwargs)
06:31 ekarlso yeah, I see
06:31 boris-42 ekarlso so for some reason service exist designate
06:31 ekarlso boris-42: I think it's due to that designate doesn't have the quotas extension enabled :D
06:32 boris-42 ekarlso oh
06:33 boris-42 ekarlso https://review.openstack.org/#/c/123421/1/modules/openstack_project/files/jenkins_job_builder/config/rally.yaml
06:33 boris-42 ekarlso just don't say to me that we need to change localrc???
06:33 openstackgerrit A change was merged to stackforge/rally: Allow user specified endpoint_type in config  https://review.openstack.org/125012
06:33 boris-42 rmk ^ so public api can be used now
06:33 ekarlso boris-42: :P
06:33 boris-42 ekarlso maybe guys you can enable it by default?)
06:34 boris-42 ekarlso it will be simpler then to make it in gates..
06:34 ekarlso boris-42: i'll check with kiall
06:34 ekarlso when he comes on
06:34 boris-42 ekarlso thanks
06:34 boris-42 ekarlso otherwise it will take a lot of time
06:34 ekarlso boris-42: how so ? :D
06:35 boris-42 ekarlso you need to merge to devstack-gate really shitty code for that
06:35 boris-42 ekarlso there is no way to pass localrc file
06:36 ekarlso :D
06:36 boris-42 ekarlso ya that is wired shit-)
06:36 boris-42 ekarlso I don't want to touch it
06:43 ekarlso boris-42: anytihng cool in rally as of late ?
06:43 boris-42 ekarlso ?
06:43 ekarlso boris-42: any new features or so ;)
06:44 boris-42 ekarlso so we improved a bit atomic actions
06:44 boris-42 ekarlso now they are counted properly
06:45 boris-42 ekarlso and they are simpler
06:45 ekarlso cool :)
06:45 boris-42 ekarlso the second thing is we improved a bit existing cloud config
06:45 boris-42 ekarlso fixed as well bug https://review.openstack.org/#/c/125012/
06:46 boris-42 ekarlso so now you can use any of those APIs public/internal/admin
06:46 boris-42 ekarlso and it will even work=)
06:46 boris-42 ekarlso we moved SLA checks
06:47 boris-42 ekarlso so results is stored in DB now
06:47 ekarlso oh
06:47 boris-42 ekarlso so when you make task results you are getting them => simplifies stuff a lot
06:47 boris-42 ekarlso we fixed bug in nova
06:48 boris-42 ekarlso https://review.openstack.org/#/c/122291/
06:48 boris-42 ekarlso so we don't have anymore races in rally gate jobs in rally=)
06:48 boris-42 ekarlso I am so happy =)
06:49 boris-42 ekarlso and I am working now on few things
06:49 boris-42 ekarlso unification of validation for scenarios & context
06:50 boris-42 ekarlso cleanuping benchmark engine, to prepare it for collecting more information
06:50 boris-42 ekarlso like information about context, graceful catch some exceptions in internal logic, and so on
06:50 boris-42 ekarlso as well finishing support of LDAP -)
06:51 boris-42 ekarlso so soon there should be a bunch of features=)
06:51 ekarlso why soon boris-42 ? :D
06:51 boris-42 ekarlso ?
06:52 boris-42 ekarlso cause I expect to finish them soon=)
06:52 ekarlso ah ok
06:52 boris-42 ekarlso as well this should be cool stuff https://review.openstack.org/#/c/125119/
06:52 boris-42 ekarlso new benchmark plot2html
06:56 boris-42 ekarlso and this improves https://review.openstack.org/#/c/125245/ scalability of rally
06:56 boris-42 ekarlso storing and working with iterations data as chunks so we can run any amount of iterations and everything will be just ifne
06:57 boris-42 ekarlso so a lot of cool stuff is comming=)
07:02 rvcehimanshu joined #openstack-rally
08:07 amaretskiy joined #openstack-rally
08:11 andreykurilin_ joined #openstack-rally
08:48 cdent joined #openstack-rally
08:48 fandi joined #openstack-rally
08:52 ekarlso boris-42: I was thinking of adding dns query stuff also to rally
08:53 ekarlso so that we can benchmark our MDNS service also using rally
09:12 mkoderer_cloud joined #openstack-rally
09:18 ekarlso boris-42: getting this when running the rally command: ImportError: Unable to import backend 'sqlalchemy'
09:19 ekarlso http://paste.openstack.org/show/117688/
09:24 fandi joined #openstack-rally
09:29 ekarlso ~/rally/rally/db/__init__.py#16 I think is a bug, it tries to from rally.db.api import *  # noqa, but rally.db.api tries to from rally.db.sqlalchemy import models which causes it to break like I pasted
09:53 ekarlso nevermind that, dunno what went wrong in creating my env
10:16 andreykurilin :)
10:18 ekarlso andreykurilin: seen this before? 2014-10-02 12:15:05.987 32159 TRACE rally Conflict: Conflict occurred attempting to store project.
10:19 andreykurilin ekarlso: can you share full log?
10:19 ekarlso andreykurilin: sure
10:20 ekarlso http://paste.openstack.org/show/117699/
10:23 andreykurilin ekarlso: hm...looks like a bug, please report it to https://bugs.launchpad.net/rally/
10:24 ekarlso doh :(
10:25 ekarlso submitted andreykurilin https://bugs.launchpad.net/rally/+bug/1376662
10:27 openstackgerrit Andrey Kurilin proposed a change to stackforge/rally: test 1376662  https://review.openstack.org/125587
10:29 boris-42 ekarlso what cloud did you use
10:29 boris-42 ekarlso and what rally version? latest?
10:31 ekarlso boris-42: existing cloud, latest trunk of rally
10:33 boris-42 ekarlso impassible
10:33 boris-42 ekarlso okay I'll take a look
10:33 boris-42 ekarlso but what I can say is it's very very strange
10:33 boris-42 ekarlso cause user context is quite well tested in gates
10:33 ekarlso boris-42: well: (.venv)ubuntu@rally:~/rally$ git show HEAD | grep commit
10:34 ekarlso commit b3bfe485bc511e5b1d1436e85def58de53fccff0
10:34 ekarlso latest commit as far as I can see
10:34 boris-42 ekarlso hm did you recarete deployment?
10:34 ekarlso boris-42: it's a fresh install in a vm
10:34 boris-42 ekarlso I mean in rallt
10:34 boris-42 ekarlso in rally*
10:34 boris-42 ekarlso or you mean it's fresh installation of rally?
10:35 ekarlso boris-42: yes
10:35 ekarlso http://paste.openstack.org/show/117709/
10:35 boris-42 ekarlso okay
10:35 ekarlso is my deployment crds
11:03 boris-42 ekarlso will need some deep dive
11:03 ekarlso boris-42: k
11:03 boris-42 ekarlso cause it's very strange=)
11:03 ekarlso you reproduced or ?
11:03 boris-42 ekarlso nope
11:03 boris-42 ekarlso I mean we are running benchmarks in gates
11:03 boris-42 ekarlso a lot of benchmakrs
11:03 boris-42 ekarlso they are all using this context
11:04 boris-42 ekarlso btw what db backed are u using?
11:09 ekarlso boris-42: mysql
11:12 jjmb joined #openstack-rally
11:47 stannie joined #openstack-rally
11:54 pboros joined #openstack-rally
11:54 ekarlso boris-42: found anything ?
11:55 boris-42 ekarlso ilyashakhat has same issues
11:56 openstackgerrit Mikhail Dubov proposed a change to stackforge/rally: Fix messages in rally exceptions  https://review.openstack.org/125345
11:56 ilyashakhat confirm: http://paste.openstack.org/show/117731/
11:59 boris-42 ilyashakhat could you share deploy config
11:59 boris-42 ilyashakhat and task config
12:01 ilyashakhat boris-42: http://paste.openstack.org/show/117734/
12:01 boris-42 ilyashakhat hmmmm
12:01 boris-42 ilyashakhat     "use_public_urls": true,
12:01 boris-42
12:01 boris-42 ilyashakhat this shouldn't work at all
12:01 ilyashakhat boris-42: it worked a couple days ago :)
12:02 boris-42 ilyashakhat yes but it was couple days ago
12:02 boris-42 ilyashakhat and yesterday we merged patch
12:02 boris-42 ilyashakhat I need to add to schema restriction
12:02 ilyashakhat but still, the config is accepted successfully and even 'deployment check' succeeds
12:02 boris-42 ilyashakhat to not allow extra fields
12:05 boris-42 ilyashakhat i will check with that config
12:05 ilyashakhat boris-42: ok
12:12 dpaterson joined #openstack-rally
12:16 cdent joined #openstack-rally
12:17 boris-42 ekarlso btw are you using devstack installation
12:17 boris-42 ekarlso ?
12:18 ekarlso boris-42: nope
12:18 openstackgerrit Mikhail Dubov proposed a change to stackforge/rally: Fix messages in rally exceptions  https://review.openstack.org/125345
12:18 boris-42 ekarlso ah so
12:18 boris-42 ekarlso I know probably the fix
12:18 vkmc joined #openstack-rally
12:18 vkmc joined #openstack-rally
12:18 boris-42 ekarlso change  "endpoint_type": "internal", to  "endpoint_type": "public",
12:19 boris-42 ekarlso in your deployment config
12:20 davpat2112 joined #openstack-rally
12:21 boris-42 cdent hmm
12:21 cdent hmmm indeed
12:25 ekarlso boris-42: uhm
12:25 ekarlso what should auth_url bwe set to ?
12:26 boris-42 ekarlso as usually
12:26 boris-42 cdent ya I mean in mysql it works well in mariadb
12:26 boris-42 cdent doesn't work
12:26 ekarlso http://x.com:5000/v2.0
12:26 ekarlso ?
12:26 boris-42 ekarlso ??
12:27 boris-42 ekarlso ya if is keystone
12:27 ekarlso boris-42: i'm just struggling to get my deployment file correctly: /
12:27 boris-42 ekarlso hm
12:27 boris-42 ekarlso how did you use rally before?
12:27 boris-42 ekarlso in authurl it's url of your keystone
12:27 boris-42 ekarlso rest of arguments are ok
12:28 boris-42 ekarlso I mean clear*
12:28 boris-42 cdent sooo I believe I know where is the issue
12:28 boris-42 cdent seems like you catch the sisue)
12:28 boris-42 cdent woot
12:28 cdent if you have a potential patch, let me know and I can test it for you
12:29 boris-42 cdent can you change few lines in rally code?
12:29 cdent certainly
12:29 boris-42 cdent take a look here https://github.com/stackforge/rally/blob/master/rally/db/sqlalchemy/types.py#L34-L38
12:29 boris-42 cdent I believe that mariadb won't have mysql name
12:29 boris-42 cdent but it has to do the same
12:29 boris-42 cdent so this should be like https://github.com/stackforge/rally/blob/master/rally/db/sqlalchemy/types.py#L35
12:30 boris-42 cdent if dialect.name in ["mysql", "mariadb"]:
12:30 boris-42 cdent I am not sure about what is exactly the name of dialect for mariadb
12:30 boris-42 cdent so it makes sense to print it
12:30 boris-42 cdent could you do it/
12:30 boris-42 ?
12:30 cdent okay, i'll figure that out
12:30 cdent and then try it
12:30 boris-42 cdent so fist time just print dialect.name
12:31 boris-42 cdent then change if
12:31 boris-42 cdent and test again
12:32 * cdent nods
12:35 boris-42 cdent btw one more thing
12:35 boris-42 cdent before testing you'll need to recreate DB
12:35 * cdent nods
12:35 boris-42 cdent so change that if, then run "rally-manage db recreate"
12:35 boris-42 cdent and I hope it will work
12:37 cdent unfortunately mariadb is reporting dialect name as 'mysql'
12:37 dpaterson joined #openstack-rally
12:52 vkmc joined #openstack-rally
12:52 vkmc joined #openstack-rally
12:55 openstackgerrit Mikhail Dubov proposed a change to stackforge/rally: Fix messages in rally exceptions  https://review.openstack.org/125345
13:13 boris-42 cdent heh
13:13 boris-42 cdent soo
13:13 boris-42 cdent I was wrong then
13:13 boris-42 cdent so it fails to safe quite small amount of data to DB..
13:14 cdent yes
13:14 cdent usually what that error means is that a connection in the pool was tried but is no longer connected
13:14 boris-42 cdent and you are sure that everything is ok with your DB backend?
13:15 cdent a) it works for everything else b) its unmodified from whatever the default install is
13:15 boris-42 cdent heh
13:15 boris-42 cdent why it produce such issues in rally..
13:20 cdent boris-42 does rally establish the db connection at startup and then use that same connection later when processing the atom stuff
13:21 boris-42 cdent it uses connection pool as far as I know
13:21 boris-42 cdent it uses oslo.db as other project does
13:21 boris-42 cdent https://github.com/openstack/oslo.db/blob/master/oslo/db/sqlalchemy/session.py#L366
13:22 boris-42 cdent and we are using few transaction/sessions
13:22 boris-42 cdent in one of them we are creating task result
13:22 boris-42 cdent that has all this info
13:23 boris-42 cdent maybe you have some timeout in mariadb
13:25 * cdent looks at my.cnf
13:25 boris-42 cdent http://stackoverflow.com/questions/12444272/error-2006-mysql-server-has-gone-away-using-python-bottle-microframework-and
13:25 boris-42 cdent read the points
13:26 boris-42 cdent  If you need big queries (for example, if you are working with big BLOB columns), you can increase the query limit by setting the server's max_allowed_packet
13:26 boris-42 cdent which has a default value of 1MB.
13:26 boris-42 cdent http://dev.mysql.com/doc/refman/5.5/en/packet-too-large.html
13:26 boris-42 cdent so seems like devatack mysql is setting this up
13:26 * cdent nods
13:27 boris-42 cdent interesting
13:31 boris-42 cdent hm
13:31 boris-42 cdent how to make this simple to end user and why it is not shown in gates..
13:38 rediskin joined #openstack-rally
13:45 cdent boris-42: I've made some changes to my.cnf and am trying again, will let you know how it goes
13:46 cdent however, I've just discovered that it seems to be the normal for openstack stuff to not retry connections:
13:46 cdent i restarted mariadb and keystone and nova both no longer worked
13:46 cdent this seems, uh, bad
14:11 rook joined #openstack-rally
14:24 rvcehimanshu joined #openstack-rally
14:26 dpaterson joined #openstack-rally
14:26 cdent boris-42: good news
14:26 cdent my oslo.db was out of date
14:26 dpaterson Bit of a newb question, how can I run a single unit test at the command line without running them all via run_test.sh
14:27 dpaterson boris-42: I have made all the changes you requested and I am adding the unit tests, which are taking me longer than the implementation thus far :)
14:35 openstackgerrit Sergey Skripnick proposed a change to stackforge/rally: Add config file for rally-designate job  https://review.openstack.org/123442
14:54 dpaterson joined #openstack-rally
15:04 openstackgerrit Oleh Anufriiev proposed a change to stackforge/rally: Add detailed description for rally commands  https://review.openstack.org/124910
16:06 rediskin left #openstack-rally
16:31 rvcehimanshu joined #openstack-rally
16:48 openstackgerrit Alexander Maretskiy proposed a change to stackforge/rally: (WIP) Improvements for task plot2html  https://review.openstack.org/125119
16:57 marcoemorais joined #openstack-rally
18:00 rook joined #openstack-rally
18:39 marcoemorais joined #openstack-rally
18:40 marcoemorais joined #openstack-rally
18:41 marcoemorais joined #openstack-rally
19:03 marcoemorais joined #openstack-rally
19:04 marcoemorais joined #openstack-rally
19:04 marcoemorais joined #openstack-rally
19:04 oanufriev joined #openstack-rally
19:05 marcoemorais joined #openstack-rally
19:05 oanufriev joined #openstack-rally
19:06 marcoemorais joined #openstack-rally
20:17 mwagner_lap joined #openstack-rally
20:31 marcoemorais1 joined #openstack-rally
20:32 marcoemorais1 joined #openstack-rally
21:02 rook joined #openstack-rally
21:02 marcoemorais joined #openstack-rally
21:13 andreykurilin_ joined #openstack-rally
21:36 marcoemorais joined #openstack-rally
21:38 marcoemorais joined #openstack-rally
21:38 marcoemorais joined #openstack-rally
21:58 openstackgerrit Kui Shi proposed a change to stackforge/rally: Add volume option to boot server from volume and run script  https://review.openstack.org/122409
23:13 jjmb joined #openstack-rally
23:48 jjmb joined #openstack-rally

| Channels | #openstack-rally index | Today | | Search | Google Search | Plain-Text | summary