2011-01-10T00:29:29 *** youdonotexist has quit IRC 2011-01-10T01:13:19 anybody here? 2011-01-10T01:17:03 *** bshum has joined #evergreen 2011-01-10T01:18:48 trying to figure out exactly what integer asset.copy wants for age_protect... it's usually null. I'm guessing 1 = 3 months? 2011-01-10T01:19:23 I don't have my pg tables open at the moment (firing up a laptop) 2011-01-10T01:19:32 But I think that sounds right. It's related to another table I think. 2011-01-10T01:21:33 Okay, I'm not seeing a direct correlation 2011-01-10T01:21:45 But there's a table called config.rule_age_hold_protect 2011-01-10T01:21:48 That has ids in it 2011-01-10T01:21:52 By default 2011-01-10T01:21:57 1 = 3 months 2011-01-10T01:22:01 2 = 6 months 2011-01-10T01:22:43 I'm assuming that if one were to enable the age item protection in the copy editor, it would insert a 1/2 based on the values it found in that table. 2011-01-10T01:23:34 That was what I was thinking, but wanted to be certain before I changed 14K records. :) 2011-01-10T01:23:39 Thanks! 2011-01-10T01:23:43 But I don't see a constraint enforcing that behavior in asset.copy. 2011-01-10T01:24:19 Dmagick-home: Good luck! Late-night record changing is fun times indeed. 2011-01-10T01:24:54 From what I am seeing around my open-ils searches, that all seems to be handled on the software side based on the DB, but not chained to it. Be safe with the snow! 2011-01-10T01:27:09 Dmagick-home: Thanks! Off to attempt some sleep. 2011-01-10T01:27:13 Good luck 2011-01-10T01:27:30 *** bshum has quit IRC 2011-01-10T01:27:31 Thanks! Sleep well! 2011-01-10T05:20:59 *** jennam has joined #evergreen 2011-01-10T07:32:16 *** ebyr has quit IRC 2011-01-10T08:00:57 *** collum has joined #evergreen 2011-01-10T08:30:21 *** AaronZ has joined #evergreen 2011-01-10T08:45:12 anyone object to me updating http://www.open-ils.org/dokuwiki/doku.php?id=evergreen-admin:sip with some information on logging via a remote syslog-ng server? 2011-01-10T08:45:28 http://pastie.org/1445118 2011-01-10T08:52:47 http://www.open-ils.org/dokuwiki/doku.php?id=evergreen-admin:sip#logging - hope thats ok 2011-01-10T08:53:25 *** collum has quit IRC 2011-01-10T09:00:20 *** Dyrcona has joined #evergreen 2011-01-10T09:02:17 *** Ghidorah has joined #evergreen 2011-01-10T09:02:40 Good morning everyone. 2011-01-10T09:02:58 Does anyone know if reports are store on the server or client side? 2011-01-10T09:07:33 *** dbs has joined #evergreen 2011-01-10T09:07:33 *** dbs has joined #evergreen 2011-01-10T09:09:02 Ghidorah: reports are stored on the server 2011-01-10T09:09:53 *** atz__ has quit IRC 2011-01-10T09:10:18 *** atz__ has joined #evergreen 2011-01-10T09:10:19 Ghidorah: thus, you can create shared folders of report templates + report output to share with other users in the system 2011-01-10T09:10:22 *** bshum has joined #evergreen 2011-01-10T09:10:31 *** Meliss has joined #evergreen 2011-01-10T09:10:45 mrpeters-isl: you should always feel free to update the wiki; worst-case scenario, we can roll back to a previous revision 2011-01-10T09:11:16 * dbs wonders if the doc team is subscribed to wiki changes so that updates can also flow into the official docs... 2011-01-10T09:11:55 *** kmlussier has joined #evergreen 2011-01-10T09:12:19 ok! 2011-01-10T09:14:12 dbs: dang. I was hoping this was a client side issue then. I just upgraded to the 1.6.1.6. I'm unable to do anything with reports. 2011-01-10T09:15:17 is there a setup of some sort the canned reports? 2011-01-10T09:15:21 Ghidorah: some explanation of your problem would probably help 2011-01-10T09:15:49 Ghidorah: there are no canned reports out of the box; Equinox makes some available though 2011-01-10T09:15:59 dbs: I navigate to to admin > local administration > reports 2011-01-10T09:16:11 Ghidorah: http://blog.esilibrary.com/2010/04/21/canned-reports-for-evergreen-160/ 2011-01-10T09:16:15 ok 2011-01-10T09:16:53 From there I see only Templates, Reports and Output 2011-01-10T09:17:10 but I have no options other than to create a sub folder 2011-01-10T09:18:56 mrpeters-isl++ # Nice SIP logging 2011-01-10T09:20:02 Ghidorah: that's step 1; you have to create a subfolder 2011-01-10T09:20:16 one each, actually, for templates, reports, and output 2011-01-10T09:20:34 I wonder if I goofed something from the upgrade then. 2011-01-10T09:20:41 The reports were there prior to the upgrade. 2011-01-10T09:20:57 http://docs.evergreen-ils.org/1.6/draft/html/reports.html 2011-01-10T09:21:00 oh. hmm 2011-01-10T09:21:23 Ghidorah: is the rest of your data still in the system? 2011-01-10T09:21:32 I'm unable to create shared folders. 2011-01-10T09:23:04 Rather IYes sir. 2011-01-10T09:23:06 err 2011-01-10T09:23:07 yes sir. 2011-01-10T09:24:06 okay. Can you poke around in "Admin -> For developers only -> JavaScript Console" and see if there are any obvious errors or warnings associated with the reports UI? 2011-01-10T09:25:57 Might be worth checking to see if reporter.report and reporter.report_folder on the database still has anything in them. 2011-01-10T09:26:09 Ghidorah: also, I'm assuming you're logging in with the 'admin' user or someone with reports permissions 2011-01-10T09:26:33 bshum: if there was nothing in them, he should still be able to create a folder 2011-01-10T09:27:43 Error: not well-formed Source File: file:///C:/Program%20Files/Evergreen%20Staff%20Client/build/chrome/content/OpenILS/data.js Line: 1, Column: 5 Source Code: dump('entering OpenILS/data.js\n'); 2011-01-10T09:28:29 Also I feel more dumb by the minute. 2011-01-10T09:28:43 I was clicking on the actual folder and not the arrow to the left of it 2011-01-10T09:28:51 so I wasn't getting a drop down 2011-01-10T09:29:08 Ghidorah: well, that's arguably a user interface usability problem, you're not alone 2011-01-10T09:29:30 Ghidorah: you can ignore the "dump()" messages 2011-01-10T09:29:46 so now that you're clicking on the arrows, are you seeing what you should be? 2011-01-10T09:30:01 *shrug* I see stuff that the librians created... 2011-01-10T09:30:09 I'm not the user of Evergreen 2011-01-10T09:30:21 *** phasebb has joined #evergreen 2011-01-10T09:30:24 I'm just the guy who was told "Hey get this server up and going" 2011-01-10T09:31:50 dbs: sorry to waste your time. 2011-01-10T09:31:54 Thank you for your help once again. 2011-01-10T09:32:54 Ghidorah: happy that it's working! 2011-01-10T09:43:13 *** jenny has joined #evergreen 2011-01-10T09:44:55 *** atz__ has quit IRC 2011-01-10T09:45:20 *** atz__ has joined #evergreen 2011-01-10T09:50:09 *** ebyr has joined #evergreen 2011-01-10T09:50:37 *** jennam has quit IRC 2011-01-10T09:53:43 where would I look to learn about super pages, etc. with search results? 2011-01-10T09:54:05 QueryParser.pm 2011-01-10T09:54:09 thanks Rubber Duck 2011-01-10T09:55:16 *** jennam has joined #evergreen 2011-01-10T09:55:42 *** jennam is now known as Guest82635 2011-01-10T09:56:53 *** atz__ has quit IRC 2011-01-10T09:57:19 *** atz__ has joined #evergreen 2011-01-10T10:06:03 *** tate-home has joined #evergreen 2011-01-10T10:08:22 phasefx: heh, I've been using trunk as my rubber duck for SuperCat experiments; you're doing better than me :) 2011-01-10T10:08:58 *** tate-home is now known as tater 2011-01-10T10:09:05 *** ebyr has quit IRC 2011-01-10T10:11:48 dbs: better duck than glove, I think 2011-01-10T10:13:07 * mjgiarlo snorts 2011-01-10T10:15:18 phasefx: if you're working with my code, you'll want the glove 2011-01-10T10:16:18 mmm 2011-01-10T10:42:58 eww, got my first osrfsys.log:[2011-01-10 10:54:05] org_tree_js.pl [ERR :27501:EX.pm:66:] Exception: OpenSRF::EX::Session 2011-01-10T10:54:05 OpenSRF::Transport /usr/local/share/perl/5.10.0/OpenSRF/Transport.pm:83 Session Error: router@private.localhost/open-ils.cstore IS NOT CONNECTED TO THE NETWORK!!! 2011-01-10T10:43:24 during /openils/bin/autogen.s 2011-01-10T10:49:52 *** atz__ has quit IRC 2011-01-10T10:50:17 *** atz__ has joined #evergreen 2011-01-10T10:51:26 *** corretico has joined #evergreen 2011-01-10T10:52:54 phasefx: i have seen those before as well. usually just re-running autogen works fine 2011-01-10T10:53:26 only ever happens w/ the first run of autogen, then it goes away. not sure what's up. 2011-01-10T11:08:19 race condition if cstore hasn't finished parsing fm_IDL? 2011-01-10T11:09:14 definitely reproducible for me, doesn't go away on subsequent runs of autogen 2011-01-10T11:09:55 I have pristine logs if anyone is interested in digging 2011-01-10T11:10:09 phasefx: is this after a rebuild of trunk / 2.0? 2011-01-10T11:10:20 yeap 2011-01-10T11:10:23 phasefx: is cstore running? 2011-01-10T11:10:33 not anymore (doh) 2011-01-10T11:10:38 * phasefx stopped everything 2011-01-10T11:11:02 you might need to rerun the first autogen.sh, make clean, and make to get the right library versioning happening 2011-01-10T11:11:36 * phasefx had ran the build autogen 2011-01-10T11:11:37 * dbs added library versioning to trunk & 2.0 last night 2011-01-10T11:11:58 alright, error happens again after a restart 2011-01-10T11:12:11 so whatever i have I have it good :) 2011-01-10T11:12:26 no sign of cstore 2011-01-10T11:12:56 so what do the logs say - should see something obvious in osrfsys.log 2011-01-10T11:13:10 nothing as obvious as [ERROR except what I pasted before 2011-01-10T11:13:14 (assuming it's the library versioning) 2011-01-10T11:13:29 * phasefx has logging turned up, so less obvious may be harder for him to find, looking 2011-01-10T11:16:18 *** atz_ has joined #evergreen 2011-01-10T11:19:25 *** atz__ has quit IRC 2011-01-10T11:19:45 after starting c services, only new things that showed up were math and dbmath. hrmmm :) 2011-01-10T11:20:22 phasebb, 2011-01-10T11:20:50 phasefx: so the logs should definitely have "open-ils.cstore failed to find liboils_utils.so.2" or something like that 2011-01-10T11:21:15 phasefx: I just updated about 10 minutes ago and I see open-ils.cstore running after a restart. 2011-01-10T11:21:33 this is trunk, right? 2011-01-10T11:21:43 dbs: yes, it does have such a thing 2011-01-10T11:21:55 Failed to dlopen library file oils_cstore.so: /openils/lib/liboils_utils.so.0: undefined symbol: oilsIDL_ntop 2011-01-10T11:22:10 * phasefx adds "Failed" to his common grep repetoire 2011-01-10T11:22:11 liboils_utils.so.0? 2011-01-10T11:22:26 * dbs cleans out /openils/lib and remakes 2011-01-10T11:23:08 such a file exists for me 2011-01-10T11:24:20 phasefx: do you have liboils_utils.so.0.0.0 2011-01-10T11:24:26 yeap 2011-01-10T11:24:36 phasefx: working here with a clean rebuild 2011-01-10T11:24:52 is liboils_utils.so.0 linking to liboils_utils.so.0.0.0? 2011-01-10T11:24:54 phasefx: do you have /openils/lib in your ldpath? 2011-01-10T11:25:10 *** ebyr has joined #evergreen 2011-01-10T11:25:15 also... I do not have liboils_utils.so.0* 2011-01-10T11:25:23 I have liboils_utils.so.2* 2011-01-10T11:25:45 so I suspect something is fouled up in your build 2011-01-10T11:25:48 oils_auth, oils_rstore, and oils_pcrud are similarly afflicted 2011-01-10T11:26:21 phasefx: please rm -f /openils/lib/oils* and liboils*, then make clean / make / make install 2011-01-10T11:27:13 also - have to duck out to french class. later 2011-01-10T11:29:32 now have .2 instead of .0 2011-01-10T11:30:05 dbs++ rebuilding_over_and_over_in_the_same_directory-- 2011-01-10T11:30:19 ooh.. make clean seems to do the trick. 2011-01-10T11:32:10 for the record: i had half of your problem, phasefx. I didn't have the 2.0 libraries, but things still seemed to be working for me. 2011-01-10T11:41:50 *** ebyr has quit IRC 2011-01-10T11:41:57 *** ebyr has joined #evergreen 2011-01-10T11:42:05 *** ebyr_ has joined #evergreen 2011-01-10T11:44:35 * Dyrcona is still baffled by the Browse Hold Shelf crashes on his server. 2011-01-10T11:45:13 Dyrcona: I haven't looked at it on our 2.0 rc1 server, but I remember having that problem repeatedly with our alpha tests. 2011-01-10T11:45:39 I've seen it in RC1+ and in trunk, now. 2011-01-10T11:45:56 I get some different errors, though eventually I start seeing this: 2011-01-10T11:46:10 *** ebyr has quit IRC 2011-01-10T11:46:10 *** ebyr_ is now known as ebyr 2011-01-10T11:46:17 openils/lib/perl5/OpenILS/Utils/CStoreEditor.pm:745 System ERROR: CStore connection timed out - transaction cannot continue\n", 2011-01-10T11:47:09 *** jamesrf has joined #evergreen 2011-01-10T11:47:12 I think it might be related to what I'm doing to "migrate" our holds from our legacy system to Evergreen, but I'm not really sure. 2011-01-10T11:47:30 I'll have to dig into the code some more after lunch. 2011-01-10T11:48:07 I'll take a look at our server. The holds there are migrated from 1.6.1 to 2.0, so if I see anything I'll let you know. 2011-01-10T11:49:39 What is frustrating is that as the smaller dialogs open up, the large one that opens first goes blank, at least with the client on Linux. 2011-01-10T11:50:07 Eventually, I kill the client from the command line to get control of my screen back. 2011-01-10T11:50:12 Whoa, network error 2011-01-10T11:50:14 Fun times 2011-01-10T11:50:53 *** atz_ has quit IRC 2011-01-10T11:51:02 *** agJohn has quit IRC 2011-01-10T11:51:17 *** atz_ has joined #evergreen 2011-01-10T11:51:35 guess i'll go to lunch and look at this after. 2011-01-10T11:54:03 *** b_bonner has joined #evergreen 2011-01-10T11:57:20 *** jamesrf has quit IRC 2011-01-10T11:58:53 *** rickd_ has joined #evergreen 2011-01-10T12:01:19 *** phasebb has quit IRC 2011-01-10T12:02:44 *** jamesrf has joined #evergreen 2011-01-10T12:19:22 *** ebyr has quit IRC 2011-01-10T12:25:35 *** jenny has left #evergreen 2011-01-10T12:25:43 *** jenny has joined #evergreen 2011-01-10T12:37:55 *** rsoulliere has joined #evergreen 2011-01-10T12:38:37 *** ebyr has joined #evergreen 2011-01-10T12:39:19 *** AaronZ has quit IRC 2011-01-10T12:48:34 *** tjspindler has joined #evergreen 2011-01-10T13:06:37 Dyrcona: The error I receive for when I try to check our hold shelf in 2.0: http://paste.lisp.org/display/118498 2011-01-10T13:06:53 Not sure if it's similar to what you see. 2011-01-10T13:07:17 Yep. I get that, plus lots more. 2011-01-10T13:07:45 Initially, I seemed to pull up the list without troubles. 2011-01-10T13:08:02 On subsequent reloads, it dies horribly with the errors I linked 2011-01-10T13:08:13 So it's sporadic 2011-01-10T13:08:50 Eventually, this shows up in the debug output: http://paste.lisp.org/display/118498#2 2011-01-10T13:09:11 Hmm 2011-01-10T13:09:36 *** ebyr_ has joined #evergreen 2011-01-10T13:09:36 *** ebyr has quit IRC 2011-01-10T13:09:37 *** ebyr_ is now known as ebyr 2011-01-10T13:09:40 Maybe we don't have any notes associated with our holds in 1.6 2011-01-10T13:10:52 Yeah, nothing in action.hold_request_note 2011-01-10T13:11:35 Oh, obviously 2011-01-10T13:11:40 It doesn't exist in 1.6.1 :( 2011-01-10T13:11:56 We don't have any notes, either, but I'm not sure that's the problem. 2011-01-10T13:12:17 If this is a code issue, I can't imagine that KCLS hasn't seen this. 2011-01-10T13:12:32 Indeed. 2011-01-10T13:27:19 Dyrcona: Maybe we should start a bug ticket on LP to track this one, and ask other people to check their hold shelf when they test 2.0 2011-01-10T13:29:55 bshum: maybe. i'm doing some other things right now, but I'll get back to looking at this in a few minutes. 2011-01-10T13:44:57 *** tate-home has joined #evergreen 2011-01-10T13:46:07 *** tater has quit IRC 2011-01-10T13:46:44 *** brian_f has joined #evergreen 2011-01-10T13:53:57 *** brian_f has quit IRC 2011-01-10T13:59:41 *** jenny1 has joined #evergreen 2011-01-10T13:59:51 *** brian_f has joined #evergreen 2011-01-10T14:02:35 *** jenny has quit IRC 2011-01-10T14:03:58 *** ebyr has quit IRC 2011-01-10T14:03:58 *** ebyr_ has joined #evergreen 2011-01-10T14:03:59 *** ebyr_ is now known as ebyr 2011-01-10T14:05:24 bshum: we've been looking at this problem off and on for a while here. 2011-01-10T14:06:09 there's a simple fix for this error: Can't use an undefined value as an ARRAY reference at /openils/lib/perl5/OpenILS/Application/Circ/Holds.pm line 1085. 2011-01-10T14:06:23 but it isn't the real problem. 2011-01-10T14:09:03 *** ebyr_ has joined #evergreen 2011-01-10T14:09:03 *** ebyr has quit IRC 2011-01-10T14:09:04 *** ebyr_ is now known as ebyr 2011-01-10T14:15:56 Ah. 2011-01-10T14:16:00 Well darn :) 2011-01-10T14:16:07 Though I enjoy simple fixes 2011-01-10T14:33:46 I will add the patch for that one in an annotation to your paste at lisp.org. 2011-01-10T14:37:00 does anyone know if Evergreen's SRU can use arbitrary supercat formats? Semi-related, I tried getting OAI DC to work with EG's opensearch by modifying Feed.pm, and merely get an empty dcCollection element back 2011-01-10T14:38:33 SRU? sru_search() in WWW::SuperCat is pretty much hardcoded for MARCXML 2011-01-10T14:38:51 ah 2011-01-10T14:39:34 it doesn't provide any mechanism for MARCXML2* transforms currently, at least none that I'm aware of 2011-01-10T14:39:55 roger that, gracias 2011-01-10T14:41:57 *** Dyrcona has quit IRC 2011-01-10T14:45:19 anyone experienced a problem where sip proceses never "close" off? 2011-01-10T14:45:37 i have probably 60 stale SIP processes 2011-01-10T14:45:51 and, thus of course coming close to my max_servers of 150 2011-01-10T14:47:35 mrpeters-isl: no problems here, but I'm using the SIP code from Sourceforge, not atz's fork on github, so your mileage may vary. 2011-01-10T14:50:16 heh, i think its the library 2011-01-10T14:50:22 67 connections from one library 2011-01-10T14:50:42 and they just keep growing 2011-01-10T15:05:12 *** Callender_ has joined #evergreen 2011-01-10T15:05:44 *** tate-home is now known as tater-home 2011-01-10T15:06:18 *** Callender has quit IRC 2011-01-10T15:06:20 *** Callender_ is now known as Callender 2011-01-10T15:06:31 *** mrpeters-isl has quit IRC 2011-01-10T15:06:39 *** mrpeters-isl has joined #evergreen 2011-01-10T15:10:58 *** Dyrcona has joined #evergreen 2011-01-10T15:11:19 * Dyrcona just found a way to lock up X11 via wine. 2011-01-10T15:12:56 * Dyrcona checks the logs to see where we were. Long support calls are a distraction. 2011-01-10T15:20:12 bshum: http://paste.lisp.org/display/118498#3 2011-01-10T15:20:40 However, I think if $q_holds is not defined, we have bigger problems. :) 2011-01-10T15:21:08 and the above patch applies equally to rel_2_0 and to trunk. 2011-01-10T15:22:39 Applying it and seeing what happens next. 2011-01-10T15:22:45 i wonder, too, if it wouldn't be better to use unless with && instead of two negatives with ||, but that line isn't the real problem. 2011-01-10T15:23:59 unless-- 2011-01-10T15:24:52 * dbs goes with http://search.cpan.org/dist/Perl-Critic/lib/Perl/Critic/Policy/ControlStructures/ProhibitUnlessBlocks.pm 2011-01-10T15:25:06 Dyrcona: Hmm, and now it doesn't like line 1130 2011-01-10T15:25:31 bshum: that's what we found in early-mid December when we first encountered this. 2011-01-10T15:25:42 That the problem keeps shifting place? 2011-01-10T15:25:54 So it must be some deeper problem 2011-01-10T15:26:04 yes, but the problem is cstore timing out/crashing. 2011-01-10T15:26:35 Aha 2011-01-10T15:26:41 it only happens with a large number of items on the holds shelf, say 100+. 2011-01-10T15:27:09 I can reliably reproduce with 100+ items on the holds shelf, but it never happens when there are only 25, for instance. 2011-01-10T15:28:04 eeevil and dbs were thinking that something needs an index last week, and we were looking through the code, but I think that work got interrupted. 2011-01-10T15:28:06 Gah.... I see what you mean. 2011-01-10T15:28:23 I was using one of our moderate libraries. 2011-01-10T15:28:29 So many more errors when opening the biggest one 2011-01-10T15:29:03 If there's no bug for this in Launchpad, it sounds like it needs one stat as a blocker for a 2.0 release 2011-01-10T15:29:27 dbs: OK. 2011-01-10T15:29:28 Indeed 2011-01-10T15:29:38 bshum: you want to add it or should I? 2011-01-10T15:29:51 Dyrcona: Sounds like you have the most info. I'll defer to you. 2011-01-10T15:29:57 Ok. 2011-01-10T15:41:37 * Dyrcona just submitted https://bugs.launchpad.net/evergreen/+bug/701208 2011-01-10T15:43:19 bshum: feel free to add anything else you know about it. 2011-01-10T15:44:58 dbs: should I target the 2.0 series with that bug or 2.0RC2? 2011-01-10T15:46:55 Dyrcona: 2.0 series I think 2011-01-10T15:47:16 distinction being that there's no actual fix targeted to land in 2.0RC2 2011-01-10T15:47:18 (cause 2.0RC2 happened already!) 2011-01-10T15:51:37 *** granitize has quit IRC 2011-01-10T15:53:49 Dyrcona: I'll check the bug ticket and see if there's anything else I can add. I can tell you that I've got bugs happening with less than 100 copies on the hold shelf, so failures may occur 25+ 2011-01-10T15:54:55 *** jenny1 has quit IRC 2011-01-10T15:56:17 *** finnapz has quit IRC 2011-01-10T15:57:17 *** finnapz has joined #evergreen 2011-01-10T15:58:18 bshum: I've only really tested with 3 or four libraries. The three that trigger it all have 100+ on the hold shelf. I should test with more libraries. 2011-01-10T15:58:48 * dbs muses - max_requests defaults to 200 in opensrf.xml; if a single holds shelf call results in > max_requests, then that child would get killed off 2011-01-10T15:58:59 Dyrcona: Our 2.0 test server is using data migrated from our live 1.6.1 database, so we have... I think 7 libraries 2011-01-10T15:59:04 Of varying size and usage 2011-01-10T15:59:37 Dyrcona: just out of interest, could you try changing max_requests from 200 to something like 2000 and restarting the opensrf processes to see if that makes any difference? 2011-01-10T15:59:39 just for fun 2011-01-10T16:00:08 opensrf or ejabberd or postgres max requests? 2011-01-10T16:00:12 (it's entirely possible and probable that KCLS has a tweaked configuration) 2011-01-10T16:00:17 opensrf.xml 2011-01-10T16:00:23 dbs: will do. 2011-01-10T16:00:25 in open-ils.cstore 2011-01-10T16:00:55 it will be the max_requests element that is the direct child of 2011-01-10T16:01:57 *** Meliss has quit IRC 2011-01-10T16:04:44 dbs: it doesn't appear to help. still crashes. 2011-01-10T16:05:06 Dyrcona: okay, good to know 2011-01-10T16:05:18 General question, how long does it take before an item that's been newly added to Evergreen begins to fulfill hold requests? I think that it has something to do with how often we run the hold_targeter, but now I'm unsure. 2011-01-10T16:05:40 bshum: depends on age protection, if you use that 2011-01-10T16:05:51 mrpeters-isl: We don't use that. 2011-01-10T16:05:55 but i'd think as soon as it goes from "reshelving" to "available" 2011-01-10T16:06:15 Dyrcona: the max_children stanza defaults to 15, do you see a spike in cstore processes when the holds shelf call runs and is it hitting that max_children cap? 2011-01-10T16:06:18 newly placed holds can see it right away, but already-placed holds won't see it until they retarget 2011-01-10T16:06:24 *** b_bonner has quit IRC 2011-01-10T16:07:13 dbs: I would say, yes: ps ax | grep -c cstore gives 17. 2011-01-10T16:07:58 should I increase that by a factor of 10 also? 2011-01-10T16:08:19 phasefx: For an already-placed hold then, how does it go about retargeting? 2011-01-10T16:08:41 bshum i think you'd need to reset the last check time in the db 2011-01-10T16:08:43 to like, 1 day ago 2011-01-10T16:08:43 bshum: there's a cron job that will retarget holds periodically based on some value and when they were last "checked" 2011-01-10T16:09:08 Dyrcona: you might end up hitting a postgresql max_connections issue, and need to bump that too, but why not start with that? 2011-01-10T16:09:25 you can also do View Holds and manually retarget holds after adding items. Evergreen doesn't do that automatically for fear of performance impact at large sites 2011-01-10T16:09:27 sounds like it might be a candidate 2011-01-10T16:09:27 bshum: you can run hold_targeter.pl with a different interval, I believe the default is 24 hours. 2011-01-10T16:09:32 phasefx: Ah, alright, so the hold_targeter.pl script in /openils/bin that we set as a cron checks once every '24h' according to the script 2011-01-10T16:09:43 but remember - last check time 2011-01-10T16:09:44 So that checks against the db value for the last time a hold was checked. 2011-01-10T16:09:56 ^^ 2011-01-10T16:10:07 So if we force the last checked times to be in the past, then that would force the next running of the script to do its job. 2011-01-10T16:10:40 bshum: you could run it with 1h as the interval. that should check anything that hasn't been checked in the last hour, no? 2011-01-10T16:11:06 Dyrcona: That's something I was wondering, if modifying that value in the script would break anything bad down the road. 2011-01-10T16:11:18 it's also not enough to run the hold targeter more frequently; there's still a configured time span somewhere 2011-01-10T16:11:38 that gets added to last check time for the test on whether to retarget or not 2011-01-10T16:11:56 I know we used to run the targeter only once a night. We raised it to once every 15 minutes. But we never adjusted the value in the script itself. 2011-01-10T16:12:03 Sounds like we should do that. 2011-01-10T16:12:21 dbs: i increased max_children to 45. we'll see what happens, now. 2011-01-10T16:12:43 if libraries are printing out pull lists, they may not want things juggling too frequently, as well 2011-01-10T16:12:49 * dbs crosses fingers 2011-01-10T16:13:32 phasefx: Ah, that would be problematic. I think they do print off those requests. 2011-01-10T16:13:39 might be a sign that something is issuing transactions (to hit the master database for reads) without rolling back or committing them after 2011-01-10T16:13:40 we also run every 15 minutes 2011-01-10T16:13:50 i think the interval is 'open-ils.storage.action.hold_request.copy_targeter' => '24h' from within the script itself 2011-01-10T16:14:16 so, things with a last check time of now()-24 hours or more would be eligible for retarget 2011-01-10T16:14:37 dbs: it worked. it didn't go over 30 or so children. 2011-01-10T16:15:29 Dyrcona: no errors? 2011-01-10T16:15:39 dbs: no errors. 2011-01-10T16:15:42 sweet 2011-01-10T16:16:08 i will now try our largest library to see what happens. 2011-01-10T16:16:10 that's worth some serious documentation in the "Configuring your Evergreen system for production usage" 2011-01-10T16:16:25 yes, i agree. 2011-01-10T16:16:43 * dbs will send his consulting bill to MVLC shortly :) 2011-01-10T16:16:48 *** tjspindler has quit IRC 2011-01-10T16:17:36 mrpeters-isl: That might explain it then. The previous checktime for the holds in question was 2 am this morning. So they wouldn't be eligible again till tomorrow morning at 2 am. 2011-01-10T16:18:03 So even if we added an item right now, it wouldn't matter till the holds were retargeted tomorrow morning. 2011-01-10T16:18:10 yep thats the way i understand things 2011-01-10T16:18:30 you could always manually run the targeter again i guess 2011-01-10T16:18:39 and see what it does 2011-01-10T16:18:48 It's run a few times since. 2011-01-10T16:19:02 dbs: :) 2011-01-10T16:19:06 It's set to run every 15 minute mark, but it won't do anything if the time is set to 24 hours. 2011-01-10T16:19:18 *** rsoulliere has quit IRC 2011-01-10T16:19:24 I guess I'd have to bring it to the attention of folks here for policy thought. 2011-01-10T16:19:48 Thanks mrpeters-isl, phasefx, and Dyrcona for helping me think through some of that. 2011-01-10T16:25:11 *** ebyr is now known as eby 2011-01-10T16:27:48 so, i'll close my own bug with an entry about settings needing adjustment. 2011-01-10T16:28:43 i am curious if someone from kcls could share these types of settings with the rest of the community, though. their settings might provide a good meter stick for the rest of us. 2011-01-10T16:30:29 for some hold interfaces, I think KCLS spawned the rapid development of some alternatives. I don't remember the details, though 2011-01-10T16:31:03 *** jenny has joined #evergreen 2011-01-10T16:32:48 Dyrcona: agreed - it's one reason I've tried in the past to ask sites to post more bugs on the project bug tracker rather than filing support tickets with support companies 2011-01-10T16:33:33 sometimes the solution is "obvious" but if there's no publicly available report w/ resolution you can google, it takes some time to rediscover the obvious solution 2011-01-10T16:34:09 looks like that was for printing. Print, Print Full Pull List, Print Full Pull List (Alternate strategy) 2011-01-10T16:36:21 *** bshum has quit IRC 2011-01-10T16:41:43 *** kmlussier has quit IRC 2011-01-10T16:44:26 *** eby has quit IRC 2011-01-10T17:01:27 Dyrcona: maybe instead of marking the bug as invalid, assign it to the documentation team? (Not sure how, exactly - maybe with a "documentation" tag or something) 2011-01-10T17:01:59 *** dbs has quit IRC 2011-01-10T17:26:26 *** b_bonner has joined #evergreen 2011-01-10T17:29:42 *** Dyrcona has quit IRC 2011-01-10T17:32:28 *** bshum has joined #evergreen 2011-01-10T17:34:35 *** bshum has joined #evergreen 2011-01-10T17:46:25 *** mjg_ has joined #evergreen 2011-01-10T17:46:34 *** mjg_ has left #evergreen 2011-01-10T17:48:54 *** jamesrf has quit IRC 2011-01-10T17:55:51 *** jenny has quit IRC 2011-01-10T18:01:23 *** jamesrf has joined #evergreen 2011-01-10T18:36:23 *** b_bonner has quit IRC 2011-01-10T18:56:49 *** corretico has quit IRC 2011-01-10T18:59:32 *** corretico has joined #evergreen 2011-01-10T19:25:36 *** Dmagick-home has quit IRC 2011-01-10T19:59:39 *** b_bonner has joined #evergreen 2011-01-10T20:15:03 *** atz_ has quit IRC 2011-01-10T20:16:45 *** brian_f has quit IRC 2011-01-10T20:22:03 *** jamesrf_ has joined #evergreen 2011-01-10T20:22:16 *** jamesrf_ has left #evergreen 2011-01-10T20:22:17 *** jamesrf has quit IRC 2011-01-10T20:52:15 *** b_bonner has quit IRC 2011-01-10T21:43:07 *** rickd_ has quit IRC 2011-01-10T22:33:52 *** atheos_ has joined #evergreen 2011-01-10T22:34:54 *** atheos has quit IRC 2011-01-10T22:34:54 *** Ghidorah has quit IRC 2011-01-10T22:34:54 *** finnapz has quit IRC 2011-01-10T22:34:55 *** denials has quit IRC 2011-01-10T22:34:55 *** csharp has quit IRC 2011-01-10T22:35:23 *** atheos_ is now known as atheos 2011-01-10T22:35:32 *** Guest82635 is now known as jennam 2011-01-10T22:35:32 *** jennam has joined #evergreen 2011-01-10T22:36:01 *** Ghidorah has joined #evergreen 2011-01-10T22:37:27 *** finnapz has joined #evergreen 2011-01-10T22:37:27 *** denials has joined #evergreen 2011-01-10T22:37:27 *** csharp has joined #evergreen 2011-01-10T23:00:16 *** corretico has quit IRC 2011-01-10T23:23:20 *** phasefx has quit IRC 2011-01-10T23:25:25 *** phasefx has joined #evergreen 2011-01-10T23:37:48 *** bshum has quit IRC