20:00:40 <Rockyg> #startmeeting log_wg
20:00:41 <openstack> Meeting started Wed Sep 30 20:00:40 2015 UTC and is due to finish in 60 minutes.  The chair is Rockyg. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:42 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:00:43 <jokke_> o/
20:00:45 <openstack> The meeting name has been set to 'log_wg'
20:00:46 * bknudson lurks
20:01:03 <Rockyg> jokke_, bknudson others?  I need to creat a list...
20:01:14 * jokke_ bangs the bushes to scare bknudson out ;)
20:02:15 <Rockyg> grrr.  I'm fighting git and gerrit right now, too.
20:02:25 <jokke_> you :o
20:02:52 <jokke_> sounds serious
20:02:57 <bknudson> what's the problem?
20:03:56 <Rockyg> trying to file my candidacy for tc and the gerrit remote can't be found and can't be removed, either
20:04:35 <jokke_> Rockyg: .git/config
20:05:39 <bknudson> I'd start over with a fresh git clone
20:05:48 <bknudson> I'm lazy that way
20:05:52 <Rockyg> yeah.  not happening.  I'm causing python to throw a traceback.
20:05:59 <Rockyg> That's what I did!
20:06:43 <jokke_> ;)
20:07:08 <Rockyg> yeah.  not happening.  I'm causing python to throw a traceback.
20:07:36 <Rockyg> So I want to thank bknudson and dhellmann  for their work on getting more doc strings into config options
20:08:04 <bknudson> maybe we'll have regular doc days
20:08:13 <bknudson> like once a month
20:08:53 <jokke_> so did you guys have productive doc day? I didn't see anything ongoing on the etherpad after dumping the current texts there
20:08:54 <Rockyg> Now I need to go to the docs meeting and ask for a section for log config
20:09:20 <bknudson> most of the doc work this time was just making the doc structure more consistent
20:09:31 <bknudson> so that they were all doing autogeneration of docs
20:09:42 <bknudson> and all autogenerating entrypoints
20:09:51 <jokke_> good stuff
20:09:55 <bknudson> and all generating change history
20:10:35 <bknudson> I did pick up a oslo.log doc bug and fix that so now the config options say that most of them are ignored if you switch to logging config file from oslo.log config options
20:10:49 <jokke_> so I should get better & more consistent output if I pulled genconfig again?
20:11:06 <bknudson> you'll see it on the developer doc pages.
20:11:23 <bknudson> off of here: http://docs.openstack.org/developer/openstack-projects.html
20:11:24 <jokke_> ah, even better
20:11:32 <Rockyg> yay!
20:11:48 <bknudson> so for example oslo.db was missing a bunch of modules because they weren't auto-docs: http://docs.openstack.org/developer/oslo.db/
20:11:58 <bknudson> now they're all there: http://docs.openstack.org/developer/oslo.db/#api-documentation
20:12:32 <bknudson> we'll need a release of oslo.log before the config option doc changes are available for genconfig
20:12:45 <jokke_> sounds like it was productive then
20:13:12 <bknudson> y, but also pretty low-level stuff for now.
20:13:26 <Rockyg> what's the chance we have a release for Liberty?
20:13:51 <bknudson> liberty was several weeks ago.
20:13:59 <bknudson> for the oslo libs
20:14:10 <Rockyg> Yeah, I know.
20:14:21 <jokke_> yeah ... but the bugfix at least would be backportable
20:14:56 <jokke_> improvements not maybe so
20:15:04 <Rockyg> But, I guess I can get the docs folks to take the output and incorporate in user docs
20:16:54 <Rockyg> So, anything else to talk about?  Other than I'm going to run for TC if I can get gerrit to work before 5:59 UTC 10/1
20:17:23 <Rockyg> Unless I already missted the deadline and it was this morning :-(
20:18:35 <bknudson> hopefully I can make some real headway on keystone logging this release.
20:18:51 <bknudson> I've got some help now but it'll take a while to bring him up to speed.
20:19:49 <Rockyg> any chance you can grep the code base and see how many places oslo.log is called with ERROR?  I'd love to know the scope of the error code work
20:20:09 <Rockyg> I'm still learning what I can and can't do with git
20:20:26 <bknudson> $ find keystone -name "*.py" | xargs grep LOG.error | wc -l
20:20:30 <bknudson> 27
20:21:06 <Rockyg> Nuthin!  That could easily be changed by hand.  I bet nova is a lot more, though.
20:21:21 <bknudson> 168
20:21:48 <Rockyg> OK!  jokke_ , see, ECs could be fixed by hand.
20:22:05 <bknudson> 2182 for all of openstack repos
20:22:26 <jokke_> devstack-02 :: ~ % find glance -name "*.py" | xargs egrep -i "LOG.error|logging.error" | wc -l
20:22:29 <jokke_> 780
20:22:46 <Rockyg> OK.  these are great numbers.
20:23:20 <bknudson> 2571 is for all openstack with that grep
20:23:27 <Rockyg> jokke_, so what is the difference?  Oh, I see, both log.error and logging.error
20:23:53 <jokke_> so we're about 3rd on glance ... Duh!
20:24:05 <Rockyg> Wow!.
20:24:18 <Rockyg> But, that may mean better error handling in glance.
20:24:26 <jokke_> Rockyg: and case insensitive
20:24:34 <bknudson> I only get 114 for glance
20:24:47 <Rockyg> I wonder if all those different error paths in conductor lead back to the same error generating code for "no valid host"
20:25:18 <bknudson> make sure you don't have a bunch of .tox es
20:26:20 <jokke_> 113 ... looks more sensible, yeah
20:26:43 <jokke_> I was one step too high and had indeed the tox venvs there
20:27:28 <Rockyg> Which means all that talk about needing to auto generate is nonsense.
20:27:42 <Rockyg> It could all get done in a sprint for each project.
20:28:15 <jokke_> I think that's what I was saying a year ago :P
20:28:20 <Rockyg> Maybe we can get agreement on putting the string in and do a POC with keystone
20:28:25 <bknudson> so what are you thinking of doing?
20:28:34 <jokke_> until I was convinced otherwise
20:28:46 <bknudson> have an extra parameter to LOG.error with the message ID?
20:28:47 <Rockyg> https://review.openstack.org/#/c/172552/
20:29:08 <Rockyg> need to tweak it a bit, then convince the devs we really need it.
20:29:10 <Rockyg> Which we do.
20:29:29 <Rockyg> yup
20:30:08 <Rockyg> I'll suck the numbers out of this meeting and add them to the spec
20:30:14 <Rockyg> That should help some.
20:31:10 <Rockyg> How about numbers for CRITICAL and WARNING?  Those would increase the load, but we could do it as a "while you're in there.
20:31:12 <bknudson> we don't have that many errors in keystone... if it's caused by the client that would be an info.
20:31:39 <bknudson> there's 44 warnings
20:32:15 <bknudson> 4 critical
20:32:28 <Rockyg> Kewl
20:32:35 <jokke_> devstack-02 :: ~/glance/glance ‹master› % find . -name "*.py" | xargs egrep -i "LOG.error|logging.error|LOG.critical|logging.critical|LOG.warn|logging.warn" | wc -l
20:32:37 <bknudson> actually only 2 critical
20:32:40 <jokke_> 261
20:32:42 <jokke_> devstack-02 :: ~/glance/glance ‹master› % find . -name "*.py" | xargs egrep -i "LOG.error|logging.error|LOG.critical|logging.critical" | wc -l
20:32:45 <jokke_> 115
20:32:59 <bknudson> 2 of the critical log calls are in hacking tests.
20:33:54 <Rockyg> nice.  A couple of interns could do it over a summer, no problem.
20:34:26 <bknudson> they could write a script to do it.
20:34:33 <bknudson> and spend the summer partying
20:34:34 <jokke_> and learn hell of a lot of openstack :)
20:34:41 <Rockyg> And any one project only needs 4 places in the number portion
20:34:53 <Rockyg> yeah!
20:34:56 <jokke_> bknudson: I'd like to just give them the task and see how they prefer to solve it :P
20:35:08 <Rockyg> even better.
20:35:40 <jokke_> then after they have done it by hand ask if they thought about scripting it :P
20:37:07 <Rockyg> Depends on the intern.  I knew a guy who interned between sophmore and junior years.  He was supposed to right a bunch of code in fortran, but didn't know it.  so, he wrote a fortran interpreter in LISP and wrote all the code in LISP.
20:37:16 <Rockyg> Had it done in two months.
20:38:13 <jokke_> I'm blessed, haven't really touched either :)
20:38:57 <Rockyg> Guys, take a look at the proposed cross project sessions and make some comments: http://odsreg.openstack.org/
20:39:42 <Rockyg> I ended up with three sessions proposed, because I somehow managed to submit error codes three times and *THERE IS NO WAY TO DELEATE A PROPOSAL*
20:39:55 <Rockyg> sorry for yelling.
20:41:38 <jokke_> Rockyg: we have the log-wg session dedicated for that already
20:41:56 <bknudson> we have 1 sesion
20:41:58 <bknudson> ?
20:42:00 <Rockyg> Ah.  OK.  Hmm.
20:42:00 <bknudson> for log wg
20:42:01 <jokke_> yes
20:42:04 <bknudson> great
20:42:42 <jokke_> http://sched.co/4Fjf
20:42:43 <Rockyg> make a comment in the prosal on odsreg then, please?  Thanks!
20:42:59 <Rockyg> It'll get tossed from that then.
20:45:11 <jokke_> I did
20:45:49 <jokke_> Rockyg: you told me that this was the topic you wanted working session for so I earmarked it for that ;)
20:46:21 <Rockyg> :-)
20:47:39 <jokke_> anything else?
20:48:07 <Rockyg> nope.  I think that covers it.
20:48:09 <Rockyg> We done?
20:49:01 <jokke_> I think we are for now around this topic
20:49:20 <Rockyg> any other topics?
20:49:45 <jokke_> not from me
20:50:18 <Rockyg> Once I get the candidacy thing squared away, I'll gather all the related logging and standards reviews together.  I'm just a bit distracted at the moment;-)
20:50:34 <jokke_> understandable
20:51:31 <Rockyg> Well, let's end this thing, then!
20:51:38 <Rockyg> #endmeeting