Open Source Integrated Library System

Evergreen on IRC

#evergreen Logs for Tuesday, March 8th, 2011

< Monday, March 7th, 2011Raw Log FileWednesday, March 9th, 2011 >
#TimeNickMessage
#03:12:20mariolone has joined #evergreen
#03:12:20mariolone has left #evergreen
#04:33:13mariolone has joined #evergreen
#04:33:14mariolone has left #evergreen
#05:08:46tsbere_ has joined #evergreen
#05:09:38tsbere has quit IRC
#06:03:00mtisi has quit IRC
#06:43:26LBA has joined #evergreen
#07:33:57eliasp_ has quit IRC
#07:39:22eliasp_ has joined #evergreen
#07:45:50tater-laptop has quit IRC
#07:50:15sfortin has joined #evergreen
#07:54:04rickd_ has joined #evergreen
#07:55:39rickd_ has joined #evergreen
#08:04:54collum has joined #evergreen
#08:06:51rickd_ has quit IRC
#08:08:41tater-laptop has joined #evergreen
#08:15:15dbs has quit IRC
#08:18:02tater-laptop has quit IRC
#08:19:31phasebb has quit IRC
#08:22:16kmlussier has joined #evergreen
#08:31:15tsbere_ is now known as tsbere
#08:35:14rickd_ has joined #evergreen
#08:42:42mtisi has joined #evergreen
#08:47:18mrpeters-isl@later bshum - Re: SIP logging - we just log to /var/log/SIP.log and let syslog-ng read that file
#08:47:18pinesol`mrpeters-isl: Error: The "Later" plugin is loaded, but there is no command named "bshum" in it. Try "list Later" to see the commands in the "Later" plugin.
#08:47:26mrpeters-isl@later tell bshum - Re: SIP logging - we just log to /var/log/SIP.log and let syslog-ng read that file
#08:47:26pinesol`mrpeters-isl: The operation succeeded.
#08:51:42csharpwould there be a recommended minimum RAM for a standalone DB server for Evergreen (thinking 8GB as a default)?
#08:52:15csharp(PINES production current runs 132G, but I'm thinking of a "default" cluster setting)
#08:52:29csharps/current/currently/
#08:59:33Meliss has joined #evergreen
#09:05:30mrpeters-islheh our db server has 256GB
#09:05:47mrpeters-islwe did OK on 64GB though
#09:06:04mrpeters-islcsharp: i guess it'd depend on the size, you know?
#09:06:07csharpmrpeters-isl: so did we
#09:06:45mrpeters-islit's fun running the whole thing in RAM hah
#09:07:03csharpmrpeters-isl: it definitely helps if you can ;-)
#09:09:10mrpeters-isl8GB is probably a decent minimum for a small library
#09:09:23mrpeters-islbased on my limited experience at that size
#09:09:35mrpeters-islthat's probably 3 or 4 libraries for us which we were only at for a short time
#09:10:07csharpmrpeters-isl: that was my thinking too - a single PINES system would probably just need 8
#09:10:31mrpeters-isli mean, we see some legacy systems coming off of servers with 1GB or less of memory
#09:10:42mrpeters-islsmaller libraries, that is
#09:11:17mrpeters-island of course that's not evergreen, but you'd imagine the bib/patron data footprint being comparable
#09:11:35csharpmrpeters-isl: right
#09:11:38mrpeters-isli guess another consideration would be do you run all of evergreen on that server, or is it just the database
#09:11:50csharpit would just be the database
#09:12:53mrpeters-isllet me check something....curious how much "over" the data we load up in memory we're using
#09:14:18mrpeters-isllooks like we're using about 96GB beyond the 128GB in our base psql directory that we cat > /dev/null
#09:14:34mrpeters-islalmost works out to about a gig or 2 per library, if you think of it that way
#09:15:00csharpmrpeters-isl: sounds logical to me
#09:15:24bshum has joined #evergreen
#09:15:31mrpeters-isli say give 8GB a shot :)
#09:15:45yboston has joined #evergreen
#09:16:02csharpthanks for the sanity check, mrpeters-isl
#09:16:33tsbereWe are planning on trying to run our system on 24. I think.
#09:16:37tsbereFor the DB alone
#09:16:53mrpeters-islyeah, we are by all means overkill
#09:17:13gmcharltPostgres' performance is strongly dependent on the ability to cache reads of the database files in memory; ideally, having enough RAM to cache the whole thing in memory (plus enough extra for Postgres itself and the OS, natch) is good
#09:17:27tsbereWe have two servers we plan to use. One will have the DB, the other everything else. One has 24 GB, the other has 128 GB.
#09:17:28bshummrpeters-isl: Ah, thanks. I cheated and changed the SIPServer code to point to local5 since it seemed we weren't using that one, then added local5 to point to a destination log for all SIP activities.
#09:17:29mrpeters-isl^^ there is the expert :)
#09:17:45gmcharltbut if memory is tight, at minimum you need enough RAM to cache the tables in metabib.*
#09:18:13bshummrpeters-isl: For us, local6 (the default location) was getting jammed up with other osrfsys messages and it was confusing :(
#09:18:20gmcharlta rule of thumb I use is figuring on 200K-400K of total database size per bib record you intend to have in the final database
#09:18:35csharpgmcharlt: thanks so much for the tips
#09:18:38mrpeters-islgood to know, gmcharlt
#09:19:13bradlleave enough RAM for the Doom server, too. (man, I'm old)
#09:19:18mrpeters-isllol
#09:19:27csharpbradl: that's a given
#09:19:35atheosso, about 64k?
#09:19:46mrpeters-islhaha i knew atheos would chime in when the gaming talk started :)
#09:20:02bradlatheos++
#09:20:19atheoslove doom, but quake2 was the best
#09:20:22csharp plays the PINES Library Action (TM) MMORPG
#09:20:25berick gives bradl a 64k RAM PACK
#09:20:34mrpeters-islatheos - i think we still have a partition somewhere for your planned quake server haha
#09:21:17atheosmrpeters-isl - yea, too bad quake live has obsoleted it's purpose. well, not bad really, quake live is pretty cool.
#09:21:18bradlberick: 30 pin memory, right?
#09:21:31gmcharltmrpeters-isl: and with your setup, we can run THOUSANDS of Doom servers
#09:21:34gmcharltthousands, I tell you
#09:22:15mrpeters-isllets have some fun with the old vmware boxes :)
#09:23:03berickbradl: now your talking crazy
#09:23:18mrpeters-islgmcharlt: if you're in the office, we're having some trouble with slony (ticket 15327) - maybe you can peek in? John said no one was in yet that could look at it...
#09:23:51mrpeters-isli'm going to turn into a human doom server if we don't get reports back online soon :)
#09:24:27jenny has joined #evergreen
#09:25:35gmcharltmrpeters-isl: remind me, what's your replication server?
#09:25:50mrpeters-isl192.168.80.4
#09:41:08gmcharltmrpeters-isl: please try bouncing slon on .80.4
#09:46:53mrpeters-islstop start as documneted?
#09:47:00gmcharltye
#09:47:01gmcharlts
#09:47:04mrpeters-isl10-4
#09:47:20mrpeters-islpostgres@reporter:~$ slonik_unsubscribe_set set1 node2 | slonik
#09:47:20mrpeters-isl<stdin>:11: unsubscribed node 2 from set 1
#09:47:31mrpeters-islsame error on start
#09:47:38mrpeters-islpostgres@reporter:~$ slonik_subscribe_set set1 node2 | slonik
#09:47:38mrpeters-isl<stdin>:4: NOTICE: subscribe set: omit_copy=f
#09:47:38mrpeters-isl<stdin>:4: NOTICE: subscribe set: omit_copy=f
#09:47:38mrpeters-islCONTEXT: SQL statement "SELECT "_replication".subscribeSet_int( $1 , $2 , $3 , $4 , $5 )"
#09:47:38mrpeters-islPL/pgSQL function "subscribeset" line 77 at PERFORM
#09:47:39mrpeters-isl<stdin>:10: Subscribed nodes to set 1
#10:12:47dbs has joined #evergreen
#10:12:47dbs has joined #evergreen
#10:29:40tsbereI have a potential concern about two of the summer of coding project ideas.
#10:32:24tsbereI fear that "Modernize Evergreen's Web interface" may break work done for "Bring sanity to the Evergreen configuration interface", or that the latter may almost have to happen for the former.
#10:34:12dbstsbere: That's a fair concern. The idea is to provide a number of ideas for projects so that Google can decide whether the project warrants being included in the GSoC in the first place.
#10:34:52dbsIf accepted, then students can submit applications with a proposal for an idea they would like to work on (which may or may not be on the list of potential project ideas that we've provided)
#10:35:38dbsSo if we get through step 1 and are accepted into the program, then if we get students applying for the GSoC, then we can worry about whether their proposals require coordination. Make sense?
#10:36:29tsbereYea. Although perhaps we should make note on those two ideas now that they may need coordination if both are being worked on. Reasoning being the configuration interface uses a dojo interface that is no longer available for use in newer versions.
#10:38:09dbsI would leave that up to the mentor and the student to work out, personally. The student might want to explore something entirely different, such as XUL or a TT2 interface or jQuery or Pyjamas or something else as a potential solution to the problem
#10:38:47tsbereOn a different note, is "customizable toolbar(s) for the staff client" a potentially large enough project idea?
#10:39:04dbsSure!
#10:40:00tsbere needs to dig up his note about what his password is for the dokuwiki as he is only logged in on one of two domains it "lives" on
#10:40:23dbsOh, MARC::Record, what do you mean '"undefined" is not a valid tag'? DWIM! :)
#10:40:58gmcharltheh
#10:41:15tsbere corrects himself, he needs to request a password reset as his note appears to have had something spilled on it and is unreadable
#10:41:29gmcharltthe newewst application profile for ISO2709, where the tag length is set to 9
#10:41:51phasefxnot 8.3?
#10:42:08phasefxtsbere: I'll reset your pw
#10:42:21tsberephasefx: I filled out the form already.
#10:42:27phasefxah, cool
#10:42:28tsbereNot that big a deal ;)
#10:42:39phasefxforget that these things are user friendly
#10:43:03tsbereWell, the greylister here will be a PITA for a few minutes, but I really have to get back to work on setting up this helpdesk.
#10:52:31tsbereI added to the summer of code project ideas page. May need some editing tweaking.
#10:53:06dbstsbere: it all does :) thanks!
#10:53:28tsbereI also didn't list a mentor. Wasn't sure who to put.
#10:54:13dbsYeah, I left many of those blank. When gmcharlt and phasefx and I put our heads together I guess we can pencil in names
#11:18:06LBA has left #evergreen
#11:37:39dbsIt occurs to me that it would probably make sense to add grace periods as a column in either config.rule_circ_duration or config.rule_recurring_fine (probably the latter), rather than the all or nothing grace periods that we have today
#11:39:06gmcharltdbs: yes, I think that ideas been mooted before; I lean towards crrf
#11:39:53dbscould add that to the student projects list I guess
#11:40:08gmcharltyep
#11:40:57tsberedbs/gmcharlt: I wouldn't add that
#11:41:11tsbere points at his 90% completed git branch named graceperiods
#11:41:26dbstsbere: well, if we don't know about these things... :0
#11:41:37tsbereWould have had it done by now but this helpdesk project not turned out to be so time consuming
#11:41:40tsbere>_>
#11:42:44dbshah
#11:43:57tsbere grumbles about stupid db designs
#11:47:38tsbereThis uses mysql. It wants an autoincrement id value. WHY DOES IT NOT DO SO?
#11:58:39Dyrcona has joined #evergreen
#12:05:09brian_f has joined #evergreen
#12:14:30bshumorg unit shortnames are case sensitive right?
#12:14:49bshumIn that one could theoretically use the same name for shortnames for a library system and branch
#12:14:56bshumJust different case
#12:15:07bshumOr would that end up being very bad for searching in the future?
#12:15:24dbsbshum: it would work, but it's probably not a great idea
#12:16:37dbsI bet the intent was for shortnames to be case-insensitive, but we never added a LOWER() to the unique index definition
#12:17:14bshumdbs: I'm helping to write up the org units for the rest of our consortium libraries and we still haven't come up with a very strong shortname naming scheme
#12:17:46bshumI suddenly wasn't sure about how best to handle the library system org units
#12:18:43tsbereOne option is to say the system is ABC and the branch is ABC-DEF and sublibs/bookmobiles are ABC-DEF-GHI
#12:19:03bshumMost all our libraries are single branches
#12:19:12dbsbshum: understood; for individual branches we've used national library codes where available, but at the system level we have "LAURSYS" and "WINDSYS" and "ALGOMASYS"
#12:19:56Dyrconabshum: You can set an option in PostgreSQL to make select statement matches ignore case, so relying on case in a unique key is a bad idea in my opinion.
#12:21:05Callender has quit IRC
#12:24:02bshumDyrcona: That's what I was suddenly afraid of.
#12:24:16bshumdbs: Thanks Dan, that sounds like a sound approach
#12:24:46bshumThe shortnames only really appear to end users in the staff client right? In the OPAC, everything comes with the real full names
#12:25:04Dyrconabshum: We used the location codes from our legacy ILS for the shortnames, if that helps you.
#12:25:55gmcharltDyrcona: true, but that would tempt me to add some code to open-ils.search to make it refuse to run if that Pg option was enabled
#12:25:57gmcharlt;)
#12:26:03mrpeters-islwe just use a few letters that "make sense" along with the first letter of the city - Indiana State Library - Indianapolis = ISLI for example
#12:26:18mrpeters-islMooresville Public Library - Mooresville = MVPLM
#12:27:31csharpbshum: the shortnames in PINES pre-dated evergreen, and were system-branch, as in Athens Regional Library System, Athens-Clarke County Library became ARL-ATH
#12:28:09csharp figures that's obvious, but just in case it helps ;-)(
#12:28:28dbsgmcharlt: word. There are all kinds of ways to shoot yourself in the foot if you mess with config options.
#12:30:05csharpone of our library systems uses letters as in CPRL-A, CPRL-B, etc.
#12:30:26bshumDyrcona: That's what we're doing as well with our Horizon libraries. I'm just trying to resolve between branch / system level shortnames for single sites.
#12:30:38dbsI'm surprised sites don't use OCLC holdings codes
#12:30:41gmcharltbshum: in my experience, it's almost a universal that EG consortia use the first part of the shortname as a group mechanism for branches of a library
#12:31:16kmlussier_ has joined #evergreen
#12:31:26kmlussier has quit IRC
#12:31:28Dyrconabshum: I use some variation of the town name for the Systems, since we don't really have many real systems.
#12:31:28kmlussier_ is now known as kmlussier
#12:31:54phasefxoff-topic, I started typing this and found it amusing: sudo cp Cat
#12:33:06Callender has joined #evergreen
#12:34:27csharphttp://www-users.cs.york.ac.uk/susan/joke/unix.htm
#12:39:55bshumAha, thanks everyone. Org units can be such a hassle :(
#12:41:37mrpeters-islhahahahaha thanks csharp
#12:46:08kmlussier has quit IRC
#12:48:17bshumHmm, just noticed that in the 2.0 patron registration, one can select for the home library to be the system group instead of the specific branch (not that you can save it this way). This differs from how selecting permission groups works in having nonselectable choices grayed out.
#12:48:46bshumAlso in that it uses the shortname code instead of the full name.
#12:51:26bshumCan't remember if it worked that way in the original patron editor.
#12:51:36bshumAnyone have the old 1.6 around? (old? heh....)
#12:53:11gmcharltbshum: does the system OU's type have can_have_users set to true?
#12:53:53bshumgmcharlt: Nope, those are set to false for that OU type
#12:57:56mrpeters-islhmm i can confirm the same behavior
#12:58:07mrpeters-islwith 2.0.1 at least
#12:58:17jennybshum: just compared, 1.6 most definitely has the full name of the home library and 2.0 just has the short name code. We allow (some) users to be registered at the system level, so we aren't a good test of that
#12:58:26kmlussier has joined #evergreen
#12:58:29jennyalso 2.0.1
#12:58:40mrpeters-islnow, i dont know if it will let you "save" someone who has a system (with a false can_have_users value) though
#12:59:27mrpeters-isli'm betting this isn't too hard to fix
#12:59:42mrpeters-isljust need to think about what i've learned the past couple weeks about this interface for a second...
#13:00:18mrpeters-islyikes - yeah, it let me save a user after making their home_ou a system level org
#13:00:58mrpeters-isldouble checked that system does not have "can have users" enabled
#13:02:22jennyditto...made the changes and tested w/ same results as mrpeters-isl
#13:05:33mrpeters-isli'm thinking maybe register.js is linking actor.usr.home_ou = actor.org_unit.id and grabbing the "shortname" instead of the "name" field like it used to
#13:05:37dbsSounds like it's time for someone to open a bug.
#13:05:42mrpeters-islagreed
#13:05:45bshumYay more bug tickets :)
#13:08:04jamesrf has quit IRC
#13:08:46eeevilspeaking of bugs (and, better, fixes) ... it's about time for a 2.0.3, and a technology preview (alpha, but I've gotten complaints about "finishedness" using that term) for 2.1 ... and a 1.6.1 ... thoughts?
#13:09:22tsbereNo opinion on 2.0.x, but any chance of my patron registration patch being looked at again before a 2.1 "preview"?
#13:09:42mrpeters-isl wonders if the patron registration uses FETCH_ORG_BY_SHORTNAME : ['open-ils.actor','open-ils.actor.org_unit.retrieve_by_shortname'],
#13:10:15mrpeters-islreally just feeling around blindly though...
#13:11:07jamesrf has joined #evergreen
#13:14:57eeeviltsbere: unsure ...
#13:17:05eeeviltsbere: ... berick will look it over soon, so "yes"
#13:17:15dbswould be nice to have tsbere's patch in 2.1, agreed
#13:17:30mrpeters-isl+1 to that
#13:17:32eeeviloh, I had no worries about it getting in in some form
#13:17:40eeeviljust unsure on immediate berick tuits
#13:17:48dbswhat about https://bugs.launchpad.net/evergreen/+bug/724583 - patron credit patch via dbwells?
#13:18:36dbs(for 2.0.x)
#13:19:06eeevilphasefx: did you look at the patch? (I'm looking now, fwiw)
#13:19:55dbwellsdbs: eeevil I can apply that right away, if no objections. It does lead to another bug being possible, but one thing at a time isn't bad.
#13:20:23phasefxeeevil: I have not
#13:20:45eeevildbwells: another identified bug, or just "changes can introduce bugs"?
#13:21:07eeevildbwells: if the latter, +1 on committing, from me
#13:23:09dbsphasefx: if you could identify "the bit of code" that's not being reached for display of pre-cats in https://bugs.launchpad.net/evergreen/+bug/723895, it would save some rediscovery time :)
#13:23:52dbwellseeevil: actually, there is a related bug where you can generate credit by deliberately overpaying with a patron credit, but it isn't directly caused by the fix, so I'll commit and create a separate bug for that. It is certainly less serious than 'not working at all'
#13:24:27eeevildbwells++
#13:27:47phasefxdbs: done
#13:28:02dbsdbwells: i think you opened https://bugs.launchpad.net/evergreen/+bug/727432 for that already
#13:28:21dbsphasefx++
#13:28:27dbwellsdbs: heh, so I did :)
#13:28:43dbsyou're good !
#13:29:11dbwellsbeen in a buggy mood lately, I guess
#13:30:51bericktsbere: my only suggestion, and this can come post-commit, would be to cache the datePattern org setting value inside AFW (when it has to be fetched), so it's only fetched once per page.
#13:31:00berickotherwise, tsbere++
#13:45:59tsbereberick: That is a good idea, but I wasn't sure how to do that yet. Haven't dug into the widgets enough yet.
#14:21:15mtate has quit IRC
#14:28:55brian_f has quit IRC
#14:29:04tater has joined #evergreen
#14:29:35tater is now known as mtate
#14:41:31brian_f has joined #evergreen
#14:58:27jenny has quit IRC
#15:00:49jenny has joined #evergreen
#15:23:30brian_f has quit IRC
#15:25:34brian_f has joined #evergreen
#15:28:08dbsdbwells: are you going to merge a fix for 731564?
#15:29:13dbsalso, boy howdy could we use a standard set of tests: import records; run searches; confirm expected results are returned - especially with the big old SVF in trunk/2.1 to shake everything up again
#15:31:07dbwellsdbs: certainly when I find the time, which may not be today, so I thought I would leave it an open invitation :)
#15:31:37dbsdbwells: okey doke, assuming eeevil hasn't begun the 2.0.3 rolling, I'll try and get that fix into place now
#15:31:46dbsdbwells++
#15:35:41dbwellscommitting db changes generally takes me way more time than it probably should, as my level paranoia about making a mistake on them is pretty high, so...
#15:36:00dbwellsdbs++ # lightening my load :)
#15:37:12eeevildbs: I have not
#15:38:16gmcharlt0666.schema.royt-is-proven-right.sql # alter table biblio.record_entry rename column marc to ...
#15:40:33eeevilhehe
#15:40:34dbsgmcharlt++
#15:45:56jeffgmcharlt++ # ha!
#15:48:26eeevildbs / dbwells: thanks for that fix!
#16:00:05Meliss has quit IRC
#16:00:31kmlussier has quit IRC
#16:00:51jennam has quit IRC
#16:16:38sfortin has quit IRC
#16:20:08phasefx has quit IRC
#16:20:39jenny has quit IRC
#16:20:46phasefx has joined #evergreen
#16:25:46_bott_Anyone ever have trouble with find_orphaned_reports.pl, looking for part of the output_base, instead of the report ID? e.g. LINE 1: SELECT id FROM reporter.schedule WHERE id = web
#16:26:11_bott_Where the output_base is: /openils/var/web/reporter
#16:30:28phasefx has quit IRC
#16:31:09phasefx has joined #evergreen
#16:33:42jenny has joined #evergreen
#16:34:29gmcharlt is now known as gmcharlt_md
#16:34:35gmcharlt_md is now known as gmcharlt
#16:41:35collum has quit IRC
#16:42:19gdunbar has quit IRC
#16:54:46brian_fIs berick around, or somebody who knows how to install the Template Toolkit version of evergreen?
#16:55:44berickbrian_f: howdy
#16:56:14brian_fhi berick!
#16:57:03brian_fberick: would it be easy for you to post install instructions to the wiki? I'd like to see what is involved in it
#16:57:33berickbrian_f: yes, i can do that. i can probably get a draft up there today
#16:57:34brian_for if they are very rough, i'd be happy to get an e-mail version and help clean them up for you
#16:58:01brian_fberick: that would be great!
#16:58:26berickbrian_f: thanks. i'll let you be the judge of roughness...
#16:58:59brian_fsounds more than fair to me!
#17:00:00dbsas I recall there was very little outside of the norm to do to get opac-tt-poc rolling - small addition to apache conf I think?
#17:00:12dbsberick++ in any case
#17:01:16berickyeah, apache conf, 2 or 3 perl packages
#17:03:31berickdbs++ # for introducing me to KeyPassX many months ago via coffeecode
#17:03:59bericki was reminded as I went to log into the wiki...
#17:06:02dbsheh
#17:06:13dbspleased to be of assistance in whatever meagre way I can
#17:27:09Dyrcona has quit IRC
#17:29:31berickbrian_f: don't have time to test the steps ATM, so I probably left something out. if you try it and have problems, let me know. http://evergreen-ils.org/dokuwiki/doku.php?id=dev:opac:template-toolkit
#17:29:59berick will post to -dev as well
#17:30:27brian_fberick++ thanks - will let you know
#17:32:09jenny has left #evergreen
#17:40:56yboston has quit IRC
#17:45:30dave-esi has quit IRC
#17:46:58dbs has quit IRC
#19:07:53rickd_ has quit IRC
#20:32:40brian_f has quit IRC
#20:56:54finnapz has joined #evergreen
#21:10:55bshumNeato, tried the new opac with a copy of our 2.0.1 system, pretty fun watching it go :)
#21:10:58bshumberick++
#21:12:21bshumWhen copying the perlmod contents for OpenILS/WWW, I ended up using a different path for the actual installed contents in /openils/lib instead of the path specified in the wiki.
#21:24:32eliasp_ has quit IRC
#21:31:21eliasp has joined #evergreen
#21:38:31tater-laptop has joined #evergreen
#21:54:13mrpeters-isl_ has joined #evergreen
#22:00:07jamesrf has left #evergreen
#22:23:05mrpeters-isl_ has quit IRC
< Monday, March 7th, 2011Raw Log FileWednesday, March 9th, 2011 >