Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel-dev, 2014-06-25

| Channels | #fuel-dev index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:07 e0ne joined #fuel-dev
00:15 xarses angdraug: rmoe ^
00:15 xarses new rev
00:50 rmoe joined #fuel-dev
01:07 e0ne joined #fuel-dev
01:29 xarses joined #fuel-dev
03:02 e0ne joined #fuel-dev
03:07 e0ne joined #fuel-dev
04:07 e0ne joined #fuel-dev
05:09 dshulyak_ joined #fuel-dev
05:35 vkozhukalov joined #fuel-dev
05:35 xarses joined #fuel-dev
05:58 sambork joined #fuel-dev
06:24 vkozhukalov joined #fuel-dev
06:31 tzn joined #fuel-dev
06:57 guillaume_ joined #fuel-dev
07:00 sbartel joined #fuel-dev
07:00 al_ex6 joined #fuel-dev
07:05 sambork joined #fuel-dev
07:26 salmon_ joined #fuel-dev
07:29 Guest4877 hi
07:29 mirrorbox joined #fuel-dev
07:29 mirrorbox when I run fuel-web tests, I get this:
07:29 mirrorbox from sqlalchemy.dialects.postgresql import JSON
07:29 mirrorbox ImportError: cannot import name JSON
07:29 mirrorbox am I missing some package?
07:32 mirrorbox also I did 'install -U -r nailgun/test-requirements.txt' and it didn't help
07:32 mirrorbox I run tests with ./run_tests.sh -n
07:39 ddmitriev joined #fuel-dev
07:40 e0ne joined #fuel-dev
07:59 al_ex6 joined #fuel-dev
08:00 dshulyak mirrorbox: have you tried to install all dependencies in blank virtualenv?
08:00 e0ne joined #fuel-dev
08:01 mirrorbox dshulyak: does fuel-web/nailgun provide an automated way of running tests in venv?
08:01 mirrorbox anyways, I think I've found the reason
08:01 IvanBerezovskiy joined #fuel-dev
08:01 mirrorbox I'm actually trying this commit
08:01 mirrorbox https://review.openstack.org/#/c/97827/
08:02 mirrorbox it started using the JSON type
08:02 dshulyak mirrorbox: but it is not merged yet
08:02 mirrorbox but it didn't update the sqlalchemy
08:02 mirrorbox and currently sqlalchemy in requirements.txt is 0.7.9
08:02 mirrorbox and support for JSON added in 0.9.something
08:03 mirrorbox 0.9.0
08:03 dshulyak mirrorbox: yes, afaik we are working on packages for this versions, and it will be updated soon
08:03 mirrorbox dshulyak: yes, I can see that it's not merged
08:04 dshulyak ok, you are just interested in plugins )
08:06 mirrorbox that is true
08:22 brain461 joined #fuel-dev
08:31 al_ex6 joined #fuel-dev
08:32 guillaume_ joined #fuel-dev
08:32 ykotko joined #fuel-dev
08:35 mattymo1 joined #fuel-dev
08:35 evgeniyl mattymo: https://review.openstack.org/#/c/102464/
08:39 mattymo1 evgeniyl, we merged it with flake8 bugs? :(
08:40 evgeniyl mattymo1: yep, there was +1 from Fuel CI for some reasons.
08:41 mattymo1 evgeniyl, +1. looks great. I don't think we need any strict reviews here
08:41 evgeniyl mattymo1: thank
08:41 evgeniyl s
08:47 mattymo evgeniyl, maybe we updated flake8 since the previous commit?
08:48 bogdando hi all.
08:48 mattymo hey bogdando !
08:48 bogdando need your opinion and comments on https://blueprints.launchpad.net/fuel/+spec/deployment-stage-galera
08:48 evgeniyl mattymo: there was a bug in run_tests.sh script with wrong exit code handling, it was fixed, but your patch had previous version of the script and it is the reason why there was +1 from fuel ci
08:48 bogdando There are 4 related bugs already
08:49 bogdando perhaps we should address it asap
08:49 mattymo evgeniyl, good you know the root cause
08:49 bogdando I see the solution in a 3 steps, but I'm not sure about the step 3
08:50 mattymo bogdando, why are we splitting galera from granular deployment?
08:50 mattymo we've beaten around the bush about granular deployment for a whole year now
08:50 mattymo either we're doing it or we're not
08:50 mattymo I say we just hack it up (move to a stage like you say) until we can really do granular deployment
08:51 mattymo service tweaks are time consuming and I wouldn't try to rely too heavily on that if we can avoid it
09:00 bogdando mattymo, perhaps because the granular deployment is being deferred consecutively and we need to address this issue asap
09:01 mattymo bogdando, with mihgen and aglarendil on PTO, it's hard to reprioritize what's critical for the next release
09:01 bogdando mattymo, so we could create a separate deployment stage for it
09:01 bogdando for now
09:02 bogdando and put it in between network_config and main
09:02 mattymo yes
09:02 guillaume_ Hi guys. Quick question about building the fuel ISO. I have some problem with permission with npm install on package.json.  I run the make iso as a normal user and to be able to continue the build I have to chown -R ~/.npm.
09:02 mattymo it would be nice if network was even more granular: ifaces, vips, then firewall, then corosync base layer
09:02 mattymo guillaume_, type ls -ld $HOME
09:03 guillaume_ drwxr-xr-x 11 guillaume guillaume 4096 Jun 25 10:36 /home/guillaume
09:03 mattymo and  ls -ld $HOME/.npm
09:04 mattymo if you ran sudo make, your $HOME/.npm will be root owned
09:05 guillaume_ I done a chown -R guillaume:guillaume ~/.npm
09:05 guillaume_ drwxr-xr-x 199 guillaume guillaume 20480 Jun 25 09:28 .npm/
09:05 mattymo ok now try make iso again
09:05 guillaume_ some directories in .npm were created and owned by root before I ran the chown
09:06 guillaume_ and the make iso works now but if I didn't do the chown I got some errors with permissions.
09:07 e0ne joined #fuel-dev
09:08 e0ne_ joined #fuel-dev
09:09 guillaume_ I can try to remove the $HOME/.npm and rerun  make iso to see if I have the problem again.
09:26 mattymo guillaume_, go ahead
09:26 alex_didenko devops gyus, could you please re-trigger this CI job https://fuel-jenkins.mirantis.com/job/master_fuellib_review_systest_ubuntu/1256/ ?
09:30 teran joined #fuel-dev
09:36 bookwar alex_didenko: done
09:37 alex_didenko bookwar: thanks
09:37 * bookwar adds word 'devops' to the list of highlighted messages
09:39 omelchek joined #fuel-dev
09:42 evgeniyl mattymo: I looked at the problem with permissions in nginx container and it looks like chmod -R doesn't follow symlinks, https://github.com/stackforge/fuel-main/blob/master/docker/nginx/start.sh#L6 is it ok if I will rewrite chmod to `find /var/www/nailgun -exec chmod 755 {} \;` ?
09:43 mattymo evgeniyl, write the change request, but I believe we should explicitly follow links in the find command
09:43 mattymo I can test and make sure
09:45 omelchek left #fuel-dev
09:45 evgeniyl mattymo: yeah, your are right, I need to use -follow key
09:45 mattymo :)
09:53 teran joined #fuel-dev
10:02 guillaume_ So I tried to removed .npm and rerun the make iso and I didn't have the issue with persmissions again. I try again from a fresh fuel-main to see if it happens again...
10:15 al_ex6 joined #fuel-dev
10:25 sambork joined #fuel-dev
10:38 evgeniyl mattymo: https://review.openstack.org/#/c/102488/
10:38 brain461 joined #fuel-dev
10:50 mattymo evgeniyl, chmod: changing permissions of `/var/www/nailgun/ubuntu/fuelweb/x86_64/pool/TRANS.TBL': Read-only file system
10:56 guillaume_ mattymo, from a fresh fuel-main I have the error with permission:
10:56 guillaume_ npm ERR! Error: EACCES, open '/home/guillaume/.npm/findup-sync/0.1.3/package/package.json'
10:56 mattymo guillaume_, did you run make iso or sudo make iso?
10:56 guillaume_ make iso (I didn't see that it must be run as root)
10:57 guillaume_ and I saw some sudo command I think.
10:57 mattymo there are many sudo commands, but not for npm
10:57 mattymo guillaume_, so ls -lR /home/guillaume/.npm/findup-sync and pastebin the output?
10:58 guillaume_ some files are owned by root and some others by me (user guillaume). I'll pastebin the output.
10:59 guillaume_ http://pastebin.com/XqgK73ZV
11:00 al_ex6 joined #fuel-dev
11:00 guillaume_ note that /home/guillaume/.npm/findup-sync is owned by root.
11:01 alex_didenko ping devops guys :) https://fuel-jenkins.mirantis.com/job/master_fuellib_review_systest_ubuntu/1266/ failed due to node-3 deployment timeout :(
11:01 guillaume_ but /home/guillaume/.npm/forever-agent is owned by me...
11:01 alex_didenko cat node-3.test.domain.local/commands/uptime.txt
11:01 alex_didenko 10:41:42 up 53 min,  1 user,  load average: 29.08, 26.03, 22.42
11:02 mattymo guillaume_, something's wrong here. on our build server, these files all stay owned by the build user (jenkins)
11:02 guillaume_ Here is the complete listing: http://pastebin.com/Z6hP6wLL
11:03 mattymo guillaume_, if I were you, I'd chown -R guillaume:guillaume ~/.npm and try again
11:04 guillaume_ I did that the first time and it solved the problem that's ok but I was just surprised by this problem  :)
11:05 mattymo something's strange on your end
11:07 guillaume_ I know that scroiset_ has the same issue
11:07 guillaume_ But anyway it's not blocking us.
11:09 guillaume_ We tried with ubuntu 12.04.4 and 12.04.3.
11:12 evgeniyl mattymo: ok, updated the patch https://review.openstack.org/#/c/102488/ or maybe you have better idea how to do it?
11:13 mattymo evgeniyl, I'm still thinking
11:13 mattymo but I believe the best route is just chmod 755 /var/www/nailgun and chmod -R 755/var/www/nailgun/* 2>/dev/null
11:14 mattymo there's no links inside to follow (only 1 level)
11:15 evgeniyl Ok, I'll try it
11:24 bogdando strange issue with Ubuntu puppet logs in snapshot updated https://bugs.launchpad.net/fuel/+bug/1308684
11:28 evgeniyl mattymo: works for me, updated the patch https://review.openstack.org/#/c/102488/
11:30 al_ex6 joined #fuel-dev
11:30 ikalnitsky evgeniyl: hm.. i thought the issue was that a dump dir has incorrect owner, not in chmod
11:33 evgeniyl ikalnitsky: it should not have nginx as an owner, because it's shared directory, we just need set read permission for it.
11:36 ikalnitsky evgeniyl: I'm not sure about it. When I met this problem yesterday the chmod wasn't work for me.
11:40 evgeniyl ikalnitsky: I checked it several minutes ago, do you have env after upgrade right now?
11:41 ikalnitsky evgeniyl: unfortunately, no.
11:51 ikalnitsky evgeniyl: it's a kind of magic. i've just tested your fix and it works fine. :) so I'm ok with it.
11:52 evgeniyl ikalnitsky: set +1 for the patch, please)
12:10 dilyin joined #fuel-dev
12:13 salmon_ joined #fuel-dev
12:22 bdudko joined #fuel-dev
12:27 mattymo bookwar, I'm going to enable parallel docker builds :)
12:27 bookwar cool!
12:27 mattymo mkdir -p ~/tmp/docker; sudo docker -d -g ~/tmp/docker -p ~/tmp/docker.pid -H=unix://~/tmp/docker.sock
12:28 mattymo then docker -H=unix://~/tmp/docker.sock build, pull, save, etc..
12:31 brain461 joined #fuel-dev
12:41 evgeniyl mattymo: did you have a chance to look at https://review.openstack.org/#/c/102488/ ?
12:41 mattymo evgeniyl, +1
12:41 evgeniyl mattymo: thanks
12:43 dshulyak_ joined #fuel-dev
12:57 salmon_ joined #fuel-dev
13:00 enikanorov_ joined #fuel-dev
13:16 bookwar alex_didenko: your ubuntu test fails on different slaves, i'd say it is not the environment fault
13:17 alex_didenko bookwar: it fails with timeout after 1 hr deployment and load average is 22+ on the slave
13:17 alex_didenko do we have performance issues on CI? do we need more hardware?
13:17 spiderik joined #fuel-dev
13:18 alex_didenko the same system test (group "deploy_pr_ha") works fine on my local PC for the same pull request
13:18 bookwar alex_didenko: we run at most two environments per server, and it worked pretty well in last month
13:19 alex_didenko then we need to increase timeout, CI fails not because of errors in manifests, it fails because of deployment timeout and load average is above 20 on virtual node
13:24 nzrf joined #fuel-dev
13:28 salmon_1 joined #fuel-dev
13:31 bookwar alex_didenko: maybe your patch is affected by our environment, which iso you use for local tests?
13:32 alex_didenko bookwar: I've tested this patch on #263, I think
13:33 alex_didenko I'm looking at CI failure diagnostic snapshot, and I see nothing wrong with puppet run except the fact it's taking way to much time to apply:Wed Jun 25 09:49:42
13:33 alex_didenko started at Wed Jun 25 09:49:42, interrupted at Wed Jun 25 10:40:30
13:39 alex_didenko bookwar: here's an example of slowness seen in puppet log
13:39 alex_didenko Wed Jun 25 10:37:32 +0000 2014 Exec[concat_/etc/swift/proxy-server.conf](provider=posix) (debug): Executing '/var/lib/puppet/concat/bin/concatfragments.sh -o "/var/lib/puppet/concat/_etc_swift_proxy-server.conf/fragments.concat.out" -d "/var/lib/puppet/concat/_etc_swift_proxy-server.conf"'
13:39 alex_didenko Wed Jun 25 10:37:51 +0000 2014 /Stage[main]/Swift::Proxy/Concat[/etc/swift/proxy-server.conf]/Exec[concat_/etc/swift/proxy-server.conf]/returns (notice): executed successfully
13:39 alex_didenko 19 seconds for exec
13:48 nzrf joined #fuel-dev
13:49 bookwar alex_didenko: i'll research a bit
13:51 alex_didenko bookwar: thanks
13:58 brain461 joined #fuel-dev
14:06 monester team, do we have any good cisco admin?
14:18 salmon_ joined #fuel-dev
14:18 tuve_ joined #fuel-dev
14:28 IvanBerezovskiy left #fuel-dev
14:30 salmon_ joined #fuel-dev
14:33 al_ex6 joined #fuel-dev
14:36 omelchek joined #fuel-dev
14:52 ikalnitsky joined #fuel-dev
15:09 aleksandr_null monester: I could help with administering Cisco equipment
15:24 artem_panchenko joined #fuel-dev
15:46 brain461 joined #fuel-dev
15:54 tuve_ joined #fuel-dev
16:03 rmoe joined #fuel-dev
16:08 monester aleksandr_null, thanks for reply but I think I find a reason of my troubles
16:11 angdraug joined #fuel-dev
16:12 e0ne joined #fuel-dev
16:14 bookwar WARNING currently fuel-web tests are disabled, we are working on the bug here https://bugs.launchpad.net/fuel/+bug/1334349 and at #fuel-python
16:34 jobewan joined #fuel-dev
17:07 aleksandr_null monester: Great! Feel free to ping me if there is any questions about networking equipment will appears.
17:09 dshulyak_ joined #fuel-dev
17:17 xarses joined #fuel-dev
17:17 jaypipes joined #fuel-dev
17:30 artem_panchenko joined #fuel-dev
17:31 xarses angdraug: https://review.openstack.org/#/c/95764/
17:34 xarses angdraug: https://review.openstack.org/#/c/101242/
17:35 e0ne joined #fuel-dev
17:40 e0ne joined #fuel-dev
17:41 e0ne joined #fuel-dev
17:44 e0ne joined #fuel-dev
17:44 jaypipes joined #fuel-dev
17:50 e0ne joined #fuel-dev
18:02 bookwar which tag we use for fuel-related stuff in openstack-dev@? I see [Fuel-dev], [fuel], [Fuel] and even [FUEL] variants. Our wiki suggests [Fuel] here https://wiki.openstack.org/wiki/Fuel#Subscribe_to_mailing_lists, is it up-to-date?
18:13 e0ne joined #fuel-dev
18:19 angdraug [Fuel]
18:20 bookwar angdraug: ok, thanks
18:24 xarses bookwar: was there a fix for the fuel-web tests?
18:24 bookwar yes, it is fixed but we need to rebase all the recent patches
18:24 xarses bookwar: also why did we disable the tests? Last time we did that we ended up with bad code merged
18:27 bookwar i didn't actually disable them, just set them in a queue for non-existent server, so builds were triggered but then we aborted them altogether. They look like this in Gerrit: https://review.openstack.org/#/c/102559/
18:39 e0ne joined #fuel-dev
19:02 vkozhukalov joined #fuel-dev
19:46 e0ne joined #fuel-dev
19:52 e0ne joined #fuel-dev
20:11 angdraug rmoe: https://review.openstack.org/102486
20:36 teran joined #fuel-dev
20:39 e0ne joined #fuel-dev
20:51 salmon_ joined #fuel-dev
20:54 salmon_ joined #fuel-dev
21:39 e0ne joined #fuel-dev
22:01 brain461 joined #fuel-dev
22:05 nzrf joined #fuel-dev
22:29 teran joined #fuel-dev
22:30 teran_ joined #fuel-dev
22:33 teran__ joined #fuel-dev
22:39 e0ne joined #fuel-dev
22:52 brain461 joined #fuel-dev
22:54 brain4611 joined #fuel-dev
23:21 teran joined #fuel-dev
23:39 e0ne joined #fuel-dev
23:48 brain461 joined #fuel-dev

| Channels | #fuel-dev index | Today | | Search | Google Search | Plain-Text | summary