Open Source Integrated Library System

Evergreen on IRC

#evergreen Logs for Tuesday, January 4th, 2011

< Monday, January 3rd, 2011Raw Log FileWednesday, January 5th, 2011 >
#TimeNickMessage
#00:04:33tildeequals has joined #evergreen
#00:10:14dbs has quit IRC
#03:44:17tildeequals has quit IRC
#03:44:17denials has quit IRC
#03:44:17atz_ has quit IRC
#03:44:18shadowspar has quit IRC
#03:44:18finnapz has quit IRC
#03:44:18kbeswick_ has quit IRC
#03:44:19csharp has quit IRC
#03:44:40phasefx has quit IRC
#03:44:41gmcharlt has quit IRC
#03:44:41eeevil has quit IRC
#03:44:41wjr has quit IRC
#03:44:42brendan2 has quit IRC
#03:44:43berick has quit IRC
#03:45:29mtate has quit IRC
#03:45:29moodaepo has quit IRC
#03:45:29mrpeters-isl has quit IRC
#03:45:29Callender has quit IRC
#03:45:45dbwells has quit IRC
#03:48:08brendan2 has joined #evergreen
#03:48:08wjr has joined #evergreen
#03:48:08eeevil has joined #evergreen
#03:48:08gmcharlt has joined #evergreen
#03:48:08phasefx has joined #evergreen
#03:48:08berick has joined #evergreen
#03:48:38mtate has joined #evergreen
#03:48:38moodaepo has joined #evergreen
#03:48:38mrpeters-isl has joined #evergreen
#03:48:38Callender has joined #evergreen
#03:49:07dbwells has joined #evergreen
#03:49:48tildeequals has joined #evergreen
#03:49:48denials has joined #evergreen
#03:49:48atz_ has joined #evergreen
#03:49:48shadowspar has joined #evergreen
#03:49:48finnapz has joined #evergreen
#03:49:48kbeswick_ has joined #evergreen
#03:49:48csharp has joined #evergreen
#06:07:41gett has joined #evergreen
#06:07:47getthi
#06:08:04gettI need help for opensrf
#06:09:24gettI run maverick. When I installing i have error in command make -f src/extras/Makefile.install <distribution> where <distribution> is maverick. Any help?
#06:27:46gett has quit IRC
#06:36:17gett has joined #evergreen
#06:46:56gett has quit IRC
#07:26:17alxp has joined #evergreen
#08:26:29Dyrcona has joined #evergreen
#08:33:16mjgiarlo@later tell gett I don't think maverick is a valid distribution in opensrf trunk yet. I've been using "lucid" as the distribution on maverick, though, and it seems to work. YMMV.
#08:33:16pinesolmjgiarlo: The operation succeeded.
#08:44:05Dyrcona@later tell gett As a followup to what mjgiarlo said about maverick, I'd recommend sticking to long term support releases, particularly if you're talking about a production environment.
#08:44:05pinesolDyrcona: The operation succeeded.
#08:53:05mjgiarlopeople put things in production? that's what I have been missing.
#08:55:04kmlussier has joined #evergreen
#08:57:32Dyrcona:)
#08:57:39DyrconaWe'll get there, eventually. :)
#09:04:34r123 has joined #evergreen
#09:20:20Callender has quit IRC
#09:21:32tspindler has joined #evergreen
#09:25:56Callender has joined #evergreen
#09:32:05Dyrcona has quit IRC
#09:35:33dbs has joined #evergreen
#09:40:25jenny has joined #evergreen
#09:42:57tspindler has quit IRC
#09:43:44tspindler has joined #evergreen
#09:50:52rsoulliere has joined #evergreen
#09:53:30tsbere wonders if eeevil is around or still off without a "real" computer
#09:54:30eeeviltsbere: on a call, but as soon as it's over I'll be committing that patch from Dyrcona
#09:55:12tsbereeeevil: Was more in a "poke you about my patch as we are after RC1, last I checked" mood today, but Dyrcona's is good too.
#09:55:14eeeviltsbere: if that's what you're referring to, re "real computer" ;)
#09:57:05eeevilahh... I'm prepping for ALA midwinter, but I will try ... swore off computers for the most part during my time off :)
#09:58:32tsbereAs for "real computer" I was referring to your comment on bug 695510 where you said "not near a real computer ATM" ;)
#10:11:31atheos has joined #evergreen
#10:15:31rsoulliere has quit IRC
#10:18:17bshum has joined #evergreen
#10:20:43bshumHas anyone here heard of SIF specification for schools?
#10:21:40gmcharltbshum: sifinfo.org ?
#10:21:53gmcharltI don't know anything about it other than that it exists
#10:21:57bshumThat seems to be the one.
#10:22:10bshumSomeone here was wondering if Evergreen could work with something like that.
#10:22:17bshumI didn't recall reading anything like that.
#10:22:23bshumBut thought I would poke around
#10:22:34tsbere gets SQL errors from MSSql upon opening the page
#10:22:45tsbereI don't think Evergreen works well with MSSql errors
#10:24:28tsbereMore errors on the specification page. I already don't trust the specification if the site promoting it can't even work properly for more than 1 page load.
#10:24:39bshumtsbere: Heh :S
#10:26:14gmcharltbshum: well, the general answer is yes - SIF just appears to be a set of XML schemas and an SOA framework for data exchange
#10:26:59gmcharltmain question, really, is whether there are implementation guidelines for whatever a school would be using it for to interchange with an ILS that are more nailed down than (say) NCIP
#10:28:24bshumgmcharlt: I'll ask for some more details. Thanks for peeking at it for us.
#10:29:14kbeswick_ has quit IRC
#10:29:42gmcharltbshum: http://specification.sifinfo.org/implementation/2.0/LibraryAutomationWorkingGroup.html
#10:29:46kmlussier has quit IRC
#10:30:01gmcharltso it appears that at least some thought has been spent on what an ILS would do with it
#10:30:20bshumHmm
#10:30:59kbeswick has joined #evergreen
#10:36:23kmlussier has joined #evergreen
#10:58:17djfiander has joined #evergreen
#11:01:49eeevilgrabbing 0474
#11:03:45agJohnThere's a minor error in the install instructions for Evergreen. I'm a little reluctant to just change it but it persists from version to version.
#11:03:47agJohnThe instructions say to edit the user running apache2 by changing /etc/apache2/apache2.conf when on Ubuntu Hardy & Debian Etch.
#11:03:49agJohnI don't have a copy of Etch running at the moment, but I have a bunch of Hardy installs and they require changing /etc/apache2/envvars (as the instructions direct for newer versions).
#11:04:22agJohnSo, anyone got Etch handy so they can check the config file that contains the www-user setting? (Or I can just change the instructions for Hardy.)
#11:05:20tsbereLucid I believe is configured the same way
#11:05:40bshum has quit IRC
#11:07:04agJohnYeah, the directions for Lucid are correct as they stand; it's only the older version(s?) that have the wrong file for making the change listed. I know it's not correct for Hardy. Don't have Etch available to check it...
#11:07:43agJohnAnd, the entry is: User www-date changed to User opensrf -- sorry for the incorrect note above.
#11:07:51tsbereI wonder if it is more of a "version of Apache" difference
#11:08:48agJohntsbere, that's almost certainly the case. I think when the directions were originally created, that the aptitude-available version of apache was older on the older OS's.
#11:10:53agJohnHmmm. I'll see if the apache docs suggest a version number. But, if the current version for Debian Etch is the same as what I get on Ubuntu Hardy (which seems likely), then an adjustment would be a help, I think. (Minor, but I know the intent has always been that if you follow the directions carefully, the end result works.)
#11:35:08agJohnOK. Well, seeing no one jump in with info on Etch, and finding tsbere's note about the Apache version being the real issue, I've changed the directions to basically suggest editing the apache.conf file only if the envvars file is not there to edit. http://open-ils.org/dokuwiki/doku.php?id=server:1.6.1:install (step 8); I modified the 2.0 instructions to remove the potentially-confusing...
#11:35:10agJohn...reference to Debian Squeeze at the corresponding spot (since no other step has that kind of for-this-OS-version-do-it-this-way kind of entry--the directions do say they're tested on Deb. Squeeze). http://evergreen-ils.org/dokuwiki/doku.php?id=server:2.0:install (step 8)
#11:35:45agJohnStep 8 (w/ or w/o the sunglasses)
#11:36:39agJohnOther than the Random Magic Spells section in the wiki, is there a spot for SQL tips and tricks? Should we have a separate spot for this?
#11:39:00eeevilagJohn: that's fine for now, and yes, it's probably time to refactor the magic spells page into separate sections ... just a tuit issue, I suppose
#11:39:36dbs+1 to refactoring magic spells
#11:40:22dbsagJohn: maybe link to http://coffeecode.net/archives/212-Introduction-to-SQL-for-Evergreen-administrators.html while you're at it?
#11:40:44agJohnVery good.
#11:40:59dbs needs to fix the link to the HTML version (or more accurately the apache rewrite)
#11:44:40granitize has joined #evergreen
#11:46:47demiankatz has joined #evergreen
#11:47:17demiankatzHello. Anyone have a moment to help a newbie with a 2.0RC1 problem?
#11:47:47dbsdemiankatz: fire away!
#11:48:00dbs(and welcome, nice to see you in these parts)
#11:48:08demiankatzThanks! I've installed the software successfully on a brand new Ubuntu 10.04 installation...
#11:48:20demiankatzThe web client seems to be working correctly (accessed via browser)...
#11:48:35demiankatzbut when I try to log in with the staff client (using the default admin / open-ils login), it gives me a nasty error message.
#11:48:48dbsdemiankatz: can you log in via srfsh?
#11:49:17demiankatzYes. OpenSRF seems to be up and running. In fact, the staff client even works to the extent that if I give it a bad password, it fails to authenticate...
#11:49:20dbsdemiankatz: and also, did you restart apache after running the 'osrf_ctl.sh' command?
#11:49:25demiankatzThe thing that's going wrong is when I enter GOOD credentials, I get the error.
#11:49:43dbsdemiankatz: hmm - and memcached is running?
#11:49:52jenny has quit IRC
#11:50:05demiankatzMemcached is definitely running... and I believe I restarted Apache, though it's easy enough to try that again as a sanity check.
#11:50:07tsbereAny chance of pasting the error message?
#11:50:13demiankatzFIXME: If you encounter this alert, please inform your IT/ILS helpdesk staff or your friendly Evergreen developers. Tue Jan 04 2011 11:48:43 GMT-0500 (Eastern Standard Time) Error during login sequence. The client will logout after this dialog.
#11:50:36demiankatz...and more detailed debug info:
#11:50:42demiankatz"message":"robj is null",
#11:50:48dbshttp://paste.lisp.org/new/evergreen is available for pastes
#11:50:49demiankatz"fileName":"http://hector.library.villanova.edu/xul/rel_2_0_rc1/server/OpenILS/data.js", "lineNumber":540, "stack":"()@http://hector.library.villanova.edu/xul/rel_2_0_rc1/server/OpenILS/data.js:540\n(0)@http://hector.library.villanova.edu/xul/rel_2_0_rc1/server/util/exec.js:76\n", "name":"TypeError"
#11:51:15demiankatzAhh, good to know. I'll use that next time!
#11:51:42dbsthe paste bot isn't in channel, so you'll need to copy the link for us
#11:51:51demiankatzAnyway, one other possibly pertinent detail: during installation, everything seemed to run smoothly... except I noticed a number of errors scrolling by during the initialization of the Postgres database. Is that normal?
#11:52:10tsbereA few errors do go by the first time (dropping things that aren't there)
#11:52:14dbsdemiankatz: depends on the errors; it tries to drop schemas
#11:52:31tsbereI think it is all functions that it throws a hissy on the first time, though.
#11:53:11demiankatzDefinitely looked like a lot of function-related errors... so hopefully that's just the normal noise. In any case, since things appear to be mostly working, it certainly wasn't a total failure.
#11:53:32tsbereWere any of them errors on a CREATE FUNCTION compared to a DROP FUNCTION?
#11:54:51demiankatzUnfortunately, I didn't have the foresight to redirect the output to a log... so I don't have a record of exactly what happened.
#11:55:00dbsdemiankatz: might be in the PostgreSQL log
#11:55:03eeevildemiankatz: just to be certain, you ran autogen.sh -u?
#11:55:38demiankatzI believe so -- that sounds familiar, and I went straight from the published installation docs for 2.0RC1.
#11:55:44demiankatzIn any case, let me see if I can dig up some Postgres logs....
#11:55:50dbsdemiankatz: also, if you could poke at /openils/var/log/osrfsys.log for "ERR", that might be helpful
#11:56:05demiankatzwill do
#11:57:45demiankatzI'm noticing this error repeatedly in my postgres logs:
#11:57:46demiankatzrelation "config.circ_modifier" does not exist at character 407
#11:58:03demiankatzIt doesn't seem to be directly tied to my login problem, though (that is, I don't get a new log entry every time I reproduce the error).
#11:58:03dbsdemiankatz: oh, that sounds like a problem
#11:58:54dbs wonders if rerunning the eg_db_config.pl script will resolve that issue - and if there's an error in creating the schema on perfectly clean databases
#11:59:19demiankatzI'll give it a try.
#12:00:47Dyrcona has joined #evergreen
#12:01:03Dyrcona is in another meeting.
#12:01:34demiankatzRerunning the script definitely cleared the database (my workstation registration had to be re-entered) but it didn't change the underlying problem.
#12:01:50demiankatzLots of messages about the "container" and "asset" schemas not existing.
#12:01:59slipscomb has joined #evergreen
#12:02:14demiankatzAlso "action" and "acq" for that matter.
#12:03:20tsberedemiankatz: Sounds like some backend modules weren't installed in the DB to me, like the XML or tsearch pieces. Did you get those steps done?
#12:03:22demiankatz(oh, and to address the earlier point, I checked my bash history and confirmed that I definitely ran autogen.sh -u)
#12:04:06ChanServ changes topic to "Dev meeting agenda: http://ur1.ca/2qsta | Welcome to the #Evergreen library system channel! | We are publicly logged. | Large pastes at http://paste.lisp.org/new/evergreen"
#12:04:14agJohnIf schemas do not exist, then something rather earlier in the process is the difficulty.
#12:04:31agJohnThat is, than autogen.sh
#12:04:35demiankatzAre you referring to 2.II. in this document? http://evergreen-ils.org/dokuwiki/doku.php?id=server:2.0:install
#12:04:56dbsdemiankatz: redirecting stdout and stderr for eg_db_config.pl and pasting the schema creationg errors would be really helpful at this point :)
#12:04:57demiankatzI skipped that step since I'm using a single-server setup -- was that a mistake?
#12:05:10demiankatzOkay, will do.... one moment.
#12:05:25tsberedemiankatz: 3.I, the three psql -f statements is what I was thinking.
#12:05:32dbsdemiankatz: heh
#12:05:35dbs"You must install the PostgreSQL server yourself prior to running the prerequisite installer Makefile ('Makefile.install'); either on the same system as Evergreen itself, or on another system on the network"
#12:06:13dbsoh wait, yes, you have postgresql installed
#12:06:27dbssorry, trying to focus on the meeting which is supposed to be held right now
#12:06:35jenny has joined #evergreen
#12:06:48demiankatzRight, I installed with "make -f Open-ILS/src/extras/Makefile.install install_pgsql_server_debs_84" but I skipped the second part of the section because it says "If PostgreSQL is running on the same server as the rest of Evergreen, these prerequisites will automatically be available to PostgreSQL."
#12:06:54dbsdemiankatz: so yes, what tsbere said: step 3. I
#12:07:06djfianderwonders if anybody would fund development of an OpenSRF binding for elisp
#12:07:36demiankatzYes, I definitely completed that step with no apparent errors.
#12:07:38Dyrconadjfiander: opensrf in emacs?
#12:07:52dbsdemiankatz: cool, let's see the pastes then :)
#12:08:38djfianderDyrcona: there's already a jabber client, so how hard could it be?
#12:09:04Dyrconadjfiander: easy peasy, lemon sqeezy. :)
#12:09:18demiankatzIt's about 90k -- is that okay in a paste, or would it be better to email it to you?
#12:10:22dbsdemiankatz: that probably won't make it into a paste
#12:10:28dbspost it somewhere with a link?
#12:10:37demiankatzSure, that should be possible... one moment.
#12:10:42dbsbtw, anybody interested in the dev meeting?
#12:10:57phasefx looks at the agenda
#12:10:59djfianderdbs: it's why I'm here
#12:11:01Dyrconame, me, me! i'm interested in the dev meeting!
#12:11:24demiankatzhttp://vufind.org/docs/dblog.txt
#12:11:28tsberedev meetings are good.
#12:11:48berickmtg++
#12:12:17tsberedemiankatz: Can't locate Library/CallNumber/LC.pm - Some of your perl is missing.
#12:12:29demiankatzaha.
#12:12:51demiankatzCome to think of it, I think I noticed an error about that during the CPAN installation and then promptly forgot all about it!
#12:13:00dbsheh
#12:13:15demiankatzLet me see if I can fix it by reinstalling the general dependencies.
#12:13:18dbsokay, let's roll; I'll run the meeting / minutize if there are no other volunteers
#12:13:48dbsAction items from previous meeting: I still haven't done either of mine
#12:14:11dbsI don't think eeevil has written up his release notes for 1.6.1.x
#12:14:26dbsphasefx: any action on acq.open-ils.org?
#12:14:34phasefx updated acq.open-ils.org this morning. brianf also made a 2.0rc1 server available. both are listed on the wiki
#12:14:42dbsphasefx++ brianf++
#12:15:13dbscsharp is going to get back to us on mailing lists in general real soon now (per email this morning I think?)
#12:15:13demiankatzI don't want to disrupt your meeting -- should we continue this in a little while?
#12:15:28b_bonner has joined #evergreen
#12:15:41Dyrconademainkatz: yes, please.
#12:15:44dbsdemiankatz: we'll hopefully wrap up in less than half an hour, if you could get back to us that would be great
#12:15:54demiankatzwill do. Thanks for your help!
#12:15:56dbs(and you're welcome to hang out, we'll try to suck you in)
#12:16:24djfianderdemiankatz: flee while you still can!!!
#12:16:28phasefxdemiankatz: if you're curious, re: meeting, http://www.open-ils.org/dokuwiki/doku.php?id=dev:meetings:2011-01-04
#12:16:41dbseeevil: I don't think you got the release checklist proposal out to the mailing list, although it was discussed at the community meeting
#12:17:00jamesrf has joined #evergreen
#12:17:02dbs(also, general statement, http://evergreen-ils.org/dokuwiki/doku.php?id=dev:evergreen:release_checklist is available for extending / munging)
#12:17:49dbsAny other outstanding action items from previous meetings?
#12:18:11eeevildbs: I didn't
#12:18:47dbseeevil: okay, it's a new year, time for a fresh start :)
#12:18:53eeevilheh
#12:19:11phasefxwxpython for staff client
#12:19:31dbs1.6.1.5 was released ages ago, we need to create a 1.6.1.6 milestone and "Fix release" the 1.6.1.5 bugs in launchpad
#12:19:47dbsergo, there are no bugs targeted to 1.6.1.6 currently :)
#12:20:05Dyrconai nominate bshum for bug wrangler for 1.6.1.x.
#12:20:33dbsDyrcona: heh, pick on the guy who's not here eh?
#12:20:35eeevilDyrcona: :)
#12:20:48Dyrconaoh, sorry, didn't know he wasn't here.
#12:20:49dbs will do the 1.6.1.x launchpad munging
#12:20:54dbsat least for today :)
#12:20:59gmcharltbshum was in earlier this morning, that's close enough ;)
#12:21:17dbshttps://launchpad.net/evergreen/+milestone/1.6.2.1 shows only 1.6.2.1 bug, but eeevil mentioned two others in his "sugar plum" email
#12:21:37dbs(probably not targeted because I hadn't created the 1.6.2.1 milestone until recently)
#12:21:47eeeviljust let me know when to cut 1.6.1.6, 1.6.2.1 (well, I'll just cut that asap) and 2.0-whatever (RC2?)
#12:22:19dbseeevil++
#12:22:32dbsare there any known bugs in 1.6.1.5 that require a 1.6.1.6?
#12:23:06eeevilthere are a couple post-1.6.1.5 fixes in svn ... I don't recall them exactly
#12:23:13eeevilexcept for the one I committed today
#12:23:28eeevilfrom Robert@Mohawk
#12:23:38eeevilwhich was just the class::dbi version check
#12:23:40dbsokay, https://launchpad.net/evergreen/+milestone/1.6.1.6 exists now
#12:26:03dbs1.6.2.0 - I've got a longish (sigh) email on that drafted, but in general we're pretending it doesn't exist except for those libraries that really, really want it?
#12:26:33dbsI am curious about what 1.6.2.0 has that 2.0 doesn't, though
#12:27:12eeevildbs: as far as the community is concerned, I'm fine with it being treated as, say, 1.6.1 + unsupported features ... and not advertised
#12:27:24tsberedbs: It may be the opposite, what did 2.0 change that 1.6.2 hasn't?
#12:27:28dbseeevil: right
#12:27:45eeevilIOW, I think applying fixes to the 1.6.2 branch that come from 1.6.1 bug reports/patches is correct
#12:27:58agJohnOthers are curious about what 1.6.2.0 has that 2.0 does not have...
#12:28:08eeevilbut I wouldn't expect much community effort on the only-in-1.6.2 stuff
#12:28:27eeevildbs: a big one is the batch marc update via staff client UI
#12:28:34eeevil(non-2.0 feature)
#12:28:39dbsand hopefully there won't be too many conflicts in those bug fixes
#12:28:47dbsah
#12:29:19eeevilthat's purely data and UI, but there's something db-related that I totally can't recall OTTOMH right now
#12:29:38dbscoll
#12:29:40dbscool
#12:29:42phasefxrecord merge interface has been tweaked too
#12:29:58phasefxallows marc editor use, for example
#12:30:00dbsphasefx: and that's not in 2.0 either?
#12:30:03phasefxright
#12:30:35dbsall queued up for 2.1 though
#12:31:01dbsokay, moving on to 2.0 (if that's okay): https://launchpad.net/evergreen/+milestone/2.0rc2 shows no bugs, but again, that milestone didn't exist in launchpad until recently
#12:31:30phasefxif 2.1 is getting cut from trunk, guaranteed. If not, then we'll have to remember to backport
#12:31:32dbs will move one of the https://launchpad.net/evergreen/+milestone/2.0rc1 bugs - about see / see from tracings - to 2.0rc2
#12:32:28eeevilI should create one ... for moving custom index defs around during upgrade to make room for the pinned-id ones added in stock 2.0
#12:32:41berickphasefx: 2.1 will from trunk
#12:32:47phasefxroger that
#12:32:49dbsI've managed to put a UI / API-adding but not API-changing fix together in trunk that will help bring a level of see/see from authority support into existence
#12:32:51eeeviljust dreading doing it is all
#12:33:15eeevildbs++ #authority stuffs ... rad work, sir
#12:33:17dbsfixing an arguable regression from 1.6 in terms of the tracings
#12:33:17phasefxdbs++
#12:33:40jasonb has joined #evergreen
#12:33:53dbseeevil: yayz for the custom index defs, if you can get to that before I can, awesome
#12:34:18Meliss has joined #evergreen
#12:34:55dbshttps://bugs.launchpad.net/evergreen/+bugs?orderby=-datecreated looks pretty good (look at all those "Fix committed" at the top!), anything else to note on 2.0 or bugs in general?
#12:35:12eeevildbs: be warned, I'm running away to ALA tomorrow and really need to get to tsbere's patches soon ... :(
#12:35:34eeevil(re index def upgrade)
#12:36:09dbseeevil: no problem
#12:36:48eeevilIMO, it's all cleanup at this point ... perhaps I should cut rc2 now-ish for testing of recent commits and we can work on polishing (upgrade script, etc) while rc2 is tested?
#12:36:48dbsphasefx: a number of those bugs look like UI bits (hold notes, "boo")
#12:37:07jamesrfwe at sitka are upgrading our test server to rc1 (rc2?) this week and then doing some testing so we might find some bugs
#12:37:13eeevil(s/cleanup/cleanup and minor UI stuff)
#12:37:34dbseeevil: lemme backport that authority tracing stuff - I'll do it right after the meeting
#12:37:46eeeviljamesrf: there are a few fixed post-rc1, so ...
#12:38:02eeevil lets
#12:38:19dbssitka++ for the apache process spinning out of control detailed bug report: https://bugs.launchpad.net/evergreen/+bug/690910
#12:38:33jamesrfyeah it's actually rel_2_0@19082
#12:40:42dbsOkay - so not getting much on the outstanding bugs - shall we move on to post-2.0 planning?
#12:41:24eeevil+1
#12:42:38eeevilI'm not going to raise my hand for dojo 1.5, but I'm all for someone doing it :)
#12:42:40dbsI put the Dojo + DVCS items as brainstorming ideas more than anything else, although my email about Git at the EG conference is an attempt to shore that up somewhat
#12:42:52djfiandereeevil: really.
#12:42:53djfiander?
#12:43:42gmcharltwell, there's at least one dojo 1.3 bug we're currently working around in 2.0
#12:43:44dbsIf there are other major infrastructure changes to consider, now would be a good time to get them onto the docket
#12:43:52eeevilre git ... many of you know my opinion on dvcs in general and git in particular ... however, if we're going to discuss dvcs I think we should toss bzr into the mix
#12:44:57eeevilfor a couple reasons: it's much simpler (as in, it doesn't twist the brain of those used to non-d vcs) and it's got lauchpad integration (potentially a big bonus for us)
#12:45:19dbseeevil: sure, I would support that (discussing which dvcs); it's one of the reasons I put the link to the python developers' fairly thorough ruminations on hg vs. bzr vs. git into my email
#12:45:22eeevilthe first is obviously subjective, but the second is less so
#12:46:02Ghidorah has joined #evergreen
#12:46:22jamesrfi think due diligence is warranted
#12:46:29eeevilbiggest drawback with bzr (that I've run into in practice) is it's crap support on older debians ... but with squeeze out now, that's not so much a concern
#12:46:44eeevilor very nearly out ... or whatever it is
#12:47:09eeevil(there are other drawbacks, like "it's not git" ;), of course)
#12:47:24GhidorahHello everyone.
#12:47:31eeevilanyway, I'll respond to the initial on-list discussion
#12:47:45dbsHi Ghidorah: we're just wrapping up a developer meeting - should be done in less than 10 minutes :)
#12:48:06Ghidorahdbs: Ok, I won't interrupt .
#12:48:17eeevildbs: re "other major infrastructure changes to consider" ... I have one
#12:48:25dbsGhidorah: you're welcome to hang out, for sure!
#12:48:36dbseeevil: sure, fire away / put it on the list
#12:49:05eeevilso ... I'll be sending to the list soon (I hope ... working out a couple things) a proposal that will require postgres 9.0+
#12:49:27dbsI'll put a message on the mailing list to encourage others to propose and discuss
#12:49:30eeevil(or, lots of 8.4-compat code ... which is an option, but not my pref)
#12:50:17eeevilso ... just to plant the seed ... when shall we start thinking about bumping the PG version to 9.0 ;) (says they guy that fought to keep support for 8.1 in 1.6.x ;) )
#12:50:55dbseeevil: sure - obviously lack of current distro out-of-the-box packages for 9.0 is the major concern, but otherwise you know I'm all for keeping postgresql current
#12:51:17dbs(says the guy who didn't want any pre-8.3 support in 1.6.x :)
#12:52:02djfianderget a room you two
#12:52:03dbs(lack of distro support probably isn't a big hurdle given backport repos)
#12:52:12dbsOh, oh!
#12:52:28dbsTotally off the agenda, but - Evergreen conference - I was going to book flights
#12:52:35eeeviland, all tsbere's in-db stuff needs to be taken back up
#12:53:02berick chuckles at squeeze backport before squeeze is even released
#12:53:15dbsand wondering if I should tack an extra day or two at the end for some focused dev time (less "hack fest" and more "IRC meeting / coding in real life")
#12:53:55dbs will post something to the mailing list about Evergreen conference and dev team meeting possibility
#12:54:02eeevildbs: I'm for it
#12:54:25dbseeevil: "taken back up" - forward ported to 9.0? (sorry, not that familiar with the in-db stuff)
#12:56:36dbsanyhoo - wrapping up - next meeting, same time, same place? we'll hopefully had 6 days of 2.0rc2 testing under our belts and be getting closer to the 2.0 GA?
#12:56:49eeevildbs: just reviewed and committed to trunk
#12:57:17eeevil+1 +1
#12:57:36berick(+1 +1)++
#12:58:12djfiandernot an rvalue
#12:58:22djfianders/r/l
#12:58:48dbsokay. yay. Welcome to 2011, everyone :)
#12:59:24ChanServ changes topic to "Welcome to the #Evergreen library system channel! | We are publicly logged. | Large pastes at http://paste.lisp.org/new/evergreen"
#12:59:30berickindeed
#12:59:53dbsGhidorah / demiankatz: I think the meeting is done
#13:00:24demiankatzOkay... Good news on this end: everything is working now.
#13:00:33demiankatzThe problem is that I can't work CPAN properly. ;-)
#13:00:37bericklet's see if djfiander is an lvalue.. djfiander-- Hey, it worked ;)
#13:00:57demiankatzThe OpenSRF installation docs suggest that "You can say “no” to the initial CPAN configuration prompt to allow it to automatically configure itself to download and install Perl modules from CPAN."
#13:01:06berickdjfiander++ # karma restoration ;)
#13:01:22demiankatzI followed this advice, made some bad decisions, and messed things up... but restoring everything to defaults and letting the auto-configure do its thing allowed everything to work properly.
#13:01:32demiankatzThen I just had to blow away and rebuild the database, and all was well.
#13:01:48demiankatzA couple of suggestions to prevent others from falling into this trap:
#13:02:11demiankatz1.) Perhaps suggest that while users *can* say "no" to the CPAN auto-configure prompt, they really shouldn't unless they know that they need to.
#13:02:37demiankatz2.) Add better detection of missing dependencies to the database generation script so that it does not appear to succeed when it is really unable to continue.
#13:02:39Ghidorah has quit IRC
#13:03:12demiankatzHowever, in spite of this one problem, the installation process was a lot less painful than I remember it being the last time I tried this. Things definitely seem to be moving in the right direction!
#13:03:16dbsdemiankatz: heh, I think the CPAN "yes/no" dialogue changed between releases
#13:03:25demiankatzAhhh, that would explain it!
#13:03:50dbsat some point, CPAN said "say 'no' if you want it to auto-configure", and now I think it says "say "yes" if you want it to autoconfigure"
#13:03:52Ghidorah has joined #evergreen
#13:04:19dbsdemiankatz: good suggestion on 2
#13:04:47demiankatzIn that case, the documentation should probably be reworded to avoid mentioning a specific UI response. I let my sense of the literal overwhelm my common sense!
#13:04:57dbsdemiankatz: good suggestion
#13:05:02dbsdemiankatz++
#13:05:40demiankatzGlad to be of some small assistance. Thanks for all of your help in getting me to this point. I'm sure I'll be back with more questions as I get deeper into the setup process. Hopefully I'll be able to get a bit more involved in the community sooner or later.
#13:06:13demiankatzI'm definitely looking forward to switching to open source here -- hopefully we can pull it off in the not-too-distant future!
#13:07:24rjackson-isl has joined #evergreen
#13:09:05b_bonner has quit IRC
#13:09:30dbsdemiankatz: http://evergreen-ils.org/dokuwiki/doku.php?id=opensrf:1.6:install&rev=1292534895&do=diff look okay?
#13:09:52dbsdemiankatz: we'd be lucky to have you :)
#13:10:20demiankatzLooks good to me!
#13:10:26demiankatzThanks for the fast fix.
#13:10:39demiankatzNow I need to go show off our new Evergreen installation to the rest of the team. ;-)
#13:10:51dbsbetter to do it now, lest I forget :)
#13:10:58demiankatzI know what you mean!
#13:11:41jasonb has quit IRC
#13:12:17dbsGhidorah: fire away
#13:12:46Ghidorahdbs: Thanks! I was wondering if there is anything special I need to do
#13:12:59slipscomb has left #evergreen
#13:13:04Ghidorahto go from a server running 1_6_0_3 to the latest 1.6.1 branch
#13:15:50jenny has quit IRC
#13:17:00dbsGhidorah: http://evergreen-ils.org/dokuwiki/doku.php?id=upgrading:evergreen suggests that you should do all of the 1.6.0.x->1.6.0.y upgrades (at least the database schema upgrades) before jumping from 1.6.0.x to 1.6.1.x
#13:17:15demiankatz has quit IRC
#13:17:21dbsalthough the page looks a little out of date :/
#13:18:09GhidorahSo if I were going from server A (running older evergreen) to server B, I should upgrade server A all the way?
#13:18:20Dyrcona has quit IRC
#13:18:51Dyrcona has joined #evergreen
#13:30:58b_bonner has joined #evergreen
#13:38:48tsbereGhidorah: You should only need to upgrade the database portion, not the actual software install on the server. So if you are using the same database for both you don't have to do much on A.
#13:47:07Ghidorah has quit IRC
#13:50:52bshum has joined #evergreen
#14:04:59bshum1 has joined #evergreen
#14:05:16bshum has quit IRC
#14:05:18bshum1 is now known as bshum
#14:06:14jenny has joined #evergreen
#14:13:02bshum1 has joined #evergreen
#14:13:10bshum has quit IRC
#14:13:14bshum1 is now known as bshum
#14:18:02tildeequals has quit IRC
#14:20:35moodaepoagJohn++ # SQL stuff
#14:21:29bshumIndeed! agJohn++
#14:23:28DyrconaIs there a way to do OpenSRF::Multisession from JavaScript?
#14:23:40dbseeevil: my backporting is done
#14:25:19eeevildbs++
#14:25:30eeevilDyrcona: not without porting it ;)
#14:29:07Dyrconaeeevil: not good. :(
#14:30:11Dyrconaeeevil: I'm trying to fix some crashes we get when doing "Browse Hold Shelf" and it appears that the requests are firing "too fast" for our server or something, so I was hoping to make it work with OpenSRF::Multisession.
#14:32:23DyrconaI see in the gateway.log that the requests for open-ils.circ.hold.details.retrieve.authoritative are coming from the client, so to use something like multisession around the results of open-ils.circ.captured_holds.id_list.on_shelf.retrieve, I'd need to do multisession in the javascript.
#14:32:24eeevilDyrcona: not sure what you mean ... what requests are firing too fast? the ones to retrieve data to flesh out the rows in the grid?
#14:32:32Dyrconayes.
#14:32:53Dyrconacstore gives us timed out errors and the client goes bonkers with error windows.
#14:33:00eeeviltoo many happening at once?
#14:33:16eeevilcstore give the timeout, or something calling cstore times out on the cstore call
#14:34:11eeevilDyrcona: if you're seeing a really long running query (as in, SELECT * FROM pg_stat_activity; ), then perhaps we just need an index
#14:34:51eeevilif you're seeing too many cstores get eaten, then increase your cstore max backends (and pg max_connections, likely) ... no?
#14:35:23Dyrconai'll have the error message for you in a second. it takes a minute or so.
#14:35:28eeevilIOW, how long does one open-ils.circ.hold.details.retrieve.authoritative take
#14:36:02Dyrconaanother problem is that the error dialogs pile up and some go "blank" along the way, that's probably a xulrunner issue, though.
#14:36:51Dyrconaso, in other words, I need a more powerful database server...
#14:37:25dbsDyrcona: or, as eeevil noted, an index (if it's just a long-running query)
#14:37:33eeevilDyrcona: so it's the queries behind open-ils.circ.hold.details.retrieve.authoritative that are timing out?
#14:38:36eeevilif so, which one(s) is (are) running too long? pg_stat_activity can tell you that if you look at it while you're trying to load the browse hold shelf UI
#14:39:08Dyrconamy client seems to have frozen or something...
#14:39:26eeevilDyrcona: select now()-query_start as duration,current_query from pg_stat_activity where current_query <> '<IDLE>' order by 1 desc limit 10;
#14:40:18Dyrconaum, i think the query has ended....
#14:40:50Dyrconai also don't think it is 1 query. i think it is the number of queries for hold request information because 1 of them runs fairly fast.
#14:41:26Dyrconaand my client is useless, now, so I can't get the error information that was coming up.
#14:42:04eeevilDyrcona: if they're timing out then they're slow and you'll see them in pg_stat_activity ... otherwise something else is going on
#14:42:23eeeviland regardless, we may just need an index to make dogpiles faster
#14:42:40eeevil(fwiw, I don't think we've seen this particular issue before)
#14:44:02rsoulliere has joined #evergreen
#14:45:19dbsretrieve_hold_queue_status_impl in O:A:Circ:Holds makes a number of newish-looking queries, perhaps it's an indexing issue with one of those once you hit a certain threshold of # of holds?
#14:46:23DyrconaThe org_unit that I'm seeing this at has 296 items on the hold shelf.
#14:47:51dbsfirst query is ordered by action.hold_request.cut_in_line, which has no index
#14:48:54tildeequals has joined #evergreen
#14:48:59Dyrconabrb
#14:49:48jeff___ is now known as jeff
#14:49:53jeff has joined #evergreen
#14:50:07dbswith a subquery against action.hold_copy_map with a where using the hold column - but there is a composite unique index that includes the hold column
#14:52:19eeevilDyrcona: fwiw, the hold shelf is in use in an org that has (last I looked) 1.5k+ holds on the shelf
#14:52:42dbseeevil: still on the shelf because the client is blocking? :)
#14:53:49eeevildbs+- #funny ...
#14:54:57dbshmm - does acp.circ_modifier need an index for that last query in retrieve_hold_queue_status_impl where acp is left-joined against ccm?
#14:55:31eeevil looks
#14:56:32Dyrcona is going to rejoin from other computer.
#14:56:39Dyrcona has quit IRC
#14:57:14Dyrcona has joined #evergreen
#14:58:46eeevildbs: I believe the index would be on ccm, which wouldn't be used on a table that small
#15:01:00dbseeevil: that would make sense, and hopefully postgresql planner would work that out, I just try not to assume
#15:01:35dbsDyrcona: how many entries does action.hold_request have in your db?
#15:03:02Dyrcona70,084 rows in ahr.
#15:03:41eeevilDyrcona: real data, or all holds are for one usr or on one title or something like that
#15:03:48eeevil?
#15:04:04Dyrconareal data migrated from our old system.
#15:04:52granitize has left #evergreen
#15:04:58eeevilk
#15:05:21eeevilI'd still really like to know the query times while you're pulling the list up
#15:05:40Dyrconamost of them lack a current_copy 'cause I'm not doing something right to get the targeter to work on them.
#15:06:12Dyrconahowever, when this was first reported to us by a member last month, we deleted all of the rows where current_copy is null, and the problem still occurred.
#15:06:59eeevilthese are all id or (type+target) driven queries, so either you really do need a bigger server (doubtful) or something needs indexing, or the data is borked ...
#15:07:49eeevilDyrcona: if you have more holds per title than copies on the shelf (available/reshelving) you should expect null current_copy values ...
#15:07:55dbsThere's no index on action.hold_request.hold_type, which seems strange
#15:08:01eeevilif not, then yeah, something else is going on
#15:08:29eeevildbs: but there is on target
#15:08:34dbsI would hope that postgresql would use the target index for the (type+target) where clause, and just scan the rows returned for target
#15:08:36dbsright
#15:08:45eeeviland that's very selective even without type
#15:08:50rickd_ has joined #evergreen
#15:08:53Dyrconaok. i'm looking through the code to see if the queries are using a field that I haven't filled in.
#15:09:11Dyrconai think i can ignore cut_in_line, though, 'cause that appears to use a coalesce.
#15:09:42eeevilsorting 300-ish rows isn't worth the index anyway
#15:10:15eeevilsorting 10k rows probably wouldn't be if it's not used in the where clause too (it's not)
#15:11:55Dyrconawould a blank avg_wait_time in config.circ_modifier be the problem?
#15:12:35Dyrconai'll set them all to 1 day to see if that makes a difference.
#15:12:39eeevilDyrcona: shouldn't be: OpenSRF::Utils::interval_to_seconds($wait_data->{avg_wait_time} || $default_wait);
#15:12:58Dyrconaah. ok.
#15:13:37Dyrconaoh. i see its a left join anyway.
#15:14:59granitize has joined #evergreen
#15:21:24Dyrconai'm going to try again. any logs in particular that I should tail?
#15:23:13eeevilDyrcona: just watch the query times while you try to use the UI
#15:27:51dbsmight be useful seeing what max_connections in postgresql.conf is set to, along with the open-ils.cstore opensrf.xml config section
#15:28:35Dyrconaok. i tried with a library that have a few things on their hold shelf. it runs rather quickly.
#15:28:54Dyrconai'll try the one with 296 now and watch pg_stat_activity again.
#15:31:45b_bonner has quit IRC
#15:35:43mrpeters-isl has quit IRC
#15:35:48mrpeters-isl has joined #evergreen
#15:36:02mrpeters-isl has joined #evergreen
#15:37:11Dyrconahmm. i'm more confused.
#15:37:21Dyrconanothing really seemed to run that long, then boom!
#15:38:42Dyrconaalso. i did a query to guesstimate how many items should be on 1 library's hold shelf and get 82, but the client only shows 25. I guess I need to understand this stuff better. it is possibly something that I'm not doing when loading requests.
#15:40:47mjg_ has joined #evergreen
#15:41:04mjg_ has left #evergreen
#15:43:13alxp has quit IRC
#15:44:51DyrconaI do get a few of these: open-ils.cstore 2011-01-04 15:36:40 [WARN:4214:osrf_application.c:839:1294173303
#15:44:53Dyrcona74860] Returning method exception with message: An unknown server error occurred
#15:45:22dbsDyrcona: anything in the postgresql logs?
#15:47:38Dyrconano. not near that timestamp, and I noted that the clock on the postgres server is about 16 minutes off from the evergreen server.
#15:58:49Dyrconawell, i have a demo server to set up, so I think I'll skip trying to load holds on it until i have this figured out.
#15:59:04rsoulliere has quit IRC
#16:00:37Meliss has quit IRC
#16:04:12tspindler has quit IRC
#16:09:45jamesrf has quit IRC
#16:15:58b_bonner has joined #evergreen
#16:20:20dbsMinutes from today's meeting at http://evergreen-ils.org/dokuwiki/doku.php?id=dev:meetings:2011-01-04
#16:20:25dbsplease adjust
#16:20:39granitize has quit IRC
#16:26:18b_bonner has joined #evergreen
#16:26:51rjackson-isl has quit IRC
#16:27:21b_bonner has joined #evergreen
#16:28:08dbs2011-01-11 agenda at http://evergreen-ils.org/dokuwiki/doku.php?id=dev:meetings:2011-01-11
#16:28:20Dyrconadbs++
#16:29:45berickdbs++
#16:33:14kmlussier has quit IRC
#16:41:08jenny1 has joined #evergreen
#16:42:33jenny1 has left #evergreen
#16:43:02jenny has quit IRC
#16:44:22dbwells has left #evergreen
#16:54:52b_bonner has joined #evergreen
#16:59:42bshum has quit IRC
#17:04:25b_bonner has quit IRC
#17:22:32djfiander has quit IRC
#17:29:01Dyrcona has quit IRC
#17:35:41tildeequals_ has joined #evergreen
#17:36:17tildeequals has quit IRC
#18:17:39Dmagick-home has joined #evergreen
#18:29:40b_bonner has joined #evergreen
#18:46:27Maryll has joined #evergreen
#18:50:57Maryll has quit IRC
#19:11:28jamesrf has joined #evergreen
#19:20:06b_bonner has quit IRC
#19:50:35jamesrf has quit IRC
#20:24:14tildeequals_ has quit IRC
#20:59:49vickir has joined #evergreen
#21:02:29vickirWe hope to have our Evergreen ILS ready to go within a week or two. Does anyone know if there is a way to set closed dates by a formula? For example, closing every Sunday?
#21:05:42eeevilvickir: hours of operation in the Organizational Unit editor
#21:06:11vickirthank you
#21:06:15eeevilnp
#21:09:57vickir has quit IRC
#21:22:26jamesrf has joined #evergreen
#21:32:27rjackson-isl has joined #evergreen
#21:43:24corretico has joined #evergreen
#21:45:37rjackson-isl has quit IRC
#21:46:31jamesrf has quit IRC
#21:57:29eeevildbs: the authority nls stuff may have not been completely backported ...
#21:57:30eeevilcp: cannot stat `locale/ru-RU/authority.js': No such file or directory
#21:57:31eeevilmake: *** [install] Error 1
#21:57:48eeevildbs: so .. I'm commenting that out in the makefile for rc2
#22:16:48Anon507 has joined #evergreen
#22:16:58Anon507wats up
#22:17:38Anon507 has quit IRC
#22:42:59eeevilWOOO HOOO ... 1.6.1.6, 1.6.2.1 and 2.0-RC2 all uploaded! ... off to pack for ALA mid-winter. if any of you will be there, swing by the ESI booth.
#22:43:02eeevil out
#22:56:43dbseeevil: argh. okay, I'll see what I can do for rc3
#22:56:48dbshave a good ALA
#23:26:21dbseeevil: fixed now (whee)
#23:26:32dbseeevil++ # mr. release man!
< Monday, January 3rd, 2011Raw Log FileWednesday, January 5th, 2011 >