Time  Nick      Message
13:11 kyle      FYI: in the rel2_2 cvs there is no default set for the opac color stylesheet or the opac layout stylesheet for npl. It might be a good idea to have a default there or it just looks awful ; )
13:12 kados     yea, I noticed that
13:12 kados     next time you see owen in here holler at him
13:14 kados     10:59 < kyle> FYI: in the rel2_2 cvs there is no default set for the opac color stylesheet or the opac layout stylesheet for npl. It might be a
13:14 kados                   good idea to have a default there or it just looks awful ; )
13:14 kados     owen: ^^^
13:14 kados     11:00 <@kados> yea, I noticed that
13:14 kados     11:01 <@kados> next time you see owen in here holler at him
13:15 owen      Yeah, I wasn't sure about how to handle that.
13:15 owen      I guess at this point it would be fine to change updatedatabase, since those values don't have any bearing on the default template
13:17 owen      It points to the greater problem of needing a better way to manage all those settings
13:18 kados     yea, definitely
13:18 owen      That is, how do we link the choice of using the NPL template with the options for stylesheets
13:18 kados     hmmm
13:18 kados     owen: btw: you up for a phone call?
13:19 owen      Sure, do you have my cell number?
13:19 kados     owen: at NPL still?
23:39 kados     chris: got a sec?
23:39 chris     yep, whats up?
23:39 kados     chris: I'm trying to work through the steps for getting tumer's zebra going
23:39 kados     I know we converted to utf8
23:39 kados     wasn't there some other db changes?
23:39 chris     yeah i scripted them all
23:39 kados     cool ... where at?
23:40 chris     in the dev-week branch
23:40 chris     in mis
23:40 chris     c
23:41 kados     do you remember which ones they are?
23:41 chris     there was move_marc_to_biblioitems
23:42 kados     cool, thx
23:42 chris     biblio_framework.sql
23:43 chris     phrase_log.sql
23:43 kados     cool, thanks
23:43 chris     i think thats all of them
23:43 kados     I'll try to get it going and report back tomorrow
23:43 chris     cool
23:44 kados     should be able to wack out the plugin in a few hours once I remember the steps
23:45 chris     sweet
23:48 kados     weird
23:48 kados     grabbed npl's db with mysqldump
23:48 kados     imported it on the new server
23:49 kados     ran 2.2's updatedatabase
23:49 kados     ran the fixmissin 090 script
23:49 kados     and the change to utf8 script
23:49 kados     as well as the ones above
23:49 chris     right
23:49 kados     slapped dev-week in place
23:49 kados     now, when I do a search I get a mysql query that hangs forever:
23:49 kados     | 135 | kohaadmin | localhost | oldkoha | Query   | 11   | Copying to tmp table | SELECT DISTINCT B.biblionumber AS biblionumber , ( MATCH (title,seriestitle,unititle,B.author,subjec |
23:50 kados     zebraidx on the data
23:50 kados     started the server
23:50 kados     can search it with yaz
23:51 chris     you doing the search from the special page?
23:51 chris     afaik it only works from there
23:51 kados     yea
23:51 kados     catalogue-home.pl
23:51 chris     is it searching zebra?
23:52 chris     cos that search looks like its doing his fallback search
23:52 kados     yea
23:52 kados     it's not searching zebra
23:52 chris     show full processlist will give you the whole query btw
23:53 chris     so for some reason it think zebra isnt there
23:53 kados     ahh, cool
23:53 chris     so is falling back
23:53 kados     yep, the conn must be working
23:53 kados     bugger all ... I'm dog tired
23:53 chris     :)
23:53 kados     I'll work on this tomorrow
23:53 kados     damn zebra
23:53 chris     sounds like a plan
23:53 kados     night
02:05 paul_away tiens, il y a quelques courageux qui ne font pas le pont en France ;-)
02:06 osmoze    hello
02:06 dewey     hello, osmoze
02:07 osmoze    le seul service de la communauté.... C est d un fatigant la culture :)
02:07 paul      :)
05:02 paul      le gadget de la mort qui tue :
05:02 paul      http://www.silicon.fr/articles/15310/Le-gadget-qui-tue-la-chaussure-Nike-qui-communique-avec-l-iPod.html
07:37 qiqo      hello everybody
07:44 qiqo      tuut
08:16 paul      hello tnb/tina
08:16 ToinS     hello tnb
08:16 tnb       hey Antoine :)
08:16 tnb       ToinS: how's it going in Marseille?
08:16 ToinS     how are you ?
08:17 tnb       good.  just got up :>)
08:17 paul      summer is arrived ... very hot today !
08:17 tnb       paul: hi!
08:17 ToinS     32°C
08:17 tnb       It is still cool, but warm here.  Perfect spring
08:18 tnb       can I ask dewey to translate to farenheit?
08:18 tnb       ;)
08:18 ToinS     hehe
08:18 ToinS     try !
08:20 tnb       eek :/  I remember I have a meeting this morning.  I'll be back later. cya :)
08:20 ToinS     i think farenheit = (celsius * 1.8) + 32
08:20 ToinS     so it 's 89.6 today
08:21 tnb       pretty nice weather!
08:21 tnb       *tina gets to meeting fast
09:05 tumer     paul are you around?
09:05 paul      yep
09:05 paul      hello tumer
09:05 tumer     date:manip whats the problem you are saying about it
09:06 tumer     how slow is circulation in your installations?
09:07 paul      it's not a problem for my clients.
09:07 tumer     I have removed date:manip from circ2.pm but cannot detect much differnce
09:07 paul      because they are either small libraries or libraries with a few circ
09:07 tumer     So shall we bother?
09:07 paul      but kados reported that 60% of their circ slowness was due to compiling Date::Manip iirc
09:07 owen      tumer: have you asked kados about it?
09:07 paul      hello owen
09:08 owen      Hi paul
09:08 tumer     I dont know what time he is awake
09:08 tumer     hi :owen
09:08 tumer     opps hi:owen
09:08 paul      tumer: kados is ALWAYS awake. He don't need to sleep. never.
09:09 owen      :D
09:09 paul      he is sometimes away, because he has meetings.
09:09 tumer     :-)
09:09 paul      or has some time with tina. but never sleeps.
09:09 tumer     ;-)
09:10 paul      dewey : kados is never sleeping, can you learn that !
09:10 dewey     ...but kados is becoming a true Perl Monger......
09:10 paul      (that's from me too ;-) )
09:12 tumer     paul:have you resolved the issue of what to do with biblio.pm with kados?
09:12 paul      we spoke of it, and he wrote something on the wiki
09:13 tumer     you mean on the devel-list?
09:13 paul      no, on wiki.koha.org
09:13 tumer     I'll check that thanks
09:13 paul      API, adding database considerations
09:14 paul      last entry in the RSS feed
09:48 kados     hey tumer
09:49 kados     hi all
09:49 paul      hey kados
09:49 tumer     hi kados
09:49 kados     tumer: the calculations of circ time were done by chris and I
09:49 kados     tumer: we used dprof
09:49 kados     tumer: it's more than 60% ... something like 70%
09:49 kados     tumer: and if you're going to have a collection with over 100K records it's critical to speed up circ
09:50 kados     tumer: because it's sometimes 4-5 seconds per transaction
09:50 tumer     well I never get that long transaction periods even with date:manip
09:51 kados     yea? how many records in the collection?
09:51 tumer     150K
09:51 kados     might not be large enough yet
09:51 kados     also, with zebra, it might alleviate the mysql strain
09:51 kados     since I have both of you ...
09:51 kados     paul:
09:51 kados     tumer:
09:51 kados     I have a proposition
09:52 kados     I've been reading through HEAD and rel_2_2 and dev_week code these past few adys
09:52 kados     days even
09:52 tumer     what's it got to do with the size of collection? its a single search of a barcode whether zebra or mysql its very fast
09:52 kados     and more and more I'm thinking we are going to need to do major cleanup for 3.0
09:52 paul      tumer/kados : maybe the circ stuff goes slower when issues tables is filled with X00 000 issues
09:52 kados     paul: nope, I've tried emptying it out
09:53 kados     tumer: I'm talking about circulation, issues, returns, borrower records, etc.
09:53 tumer     and I also populate marc record and zebra with this transaction as well
09:53 kados     hmmm
09:53 kados     I'm surprised you're not seeing any speed issues
09:53 kados     with circ transactions
09:54 kados     especially since I thought zebra was actually slower on inserting than mysql
09:54 tumer     what is slow is try and get a books on loan or overdue books report. That is slow but circulation is not
09:54 kados     interesting
09:55 kados     tumer: how many librarians at the same time issuing books?
09:55 paul      probably an index would help
09:55 kados     tumer: at NPL there are seven branches
09:56 tumer     2 librarians almost hitting <enter> at the same time will get an issue almost 1 1.5 secons
09:56 kados     tumer: at around 3-4pm the system starts crawling
09:56 kados     tumer: so maybe that's it ...
09:56 kados     tumer: because NPL has times where there are 14 librarians at the same time doing issues
09:56 kados     tumer: plus other librarians doing searches and editing records, etc.
09:57 paul      didn't mod_perl help ?
09:57 kados     paul: koha's not clean enough to use it ... it causes major data errors
09:57 paul      right.
09:57 tumer     I did not use mod_perl I tryed using mysql functions for it?
09:58 kados     yea, mod_perl's disabled on NPL's server too
09:58 kados     anyway ... got a question for the both of you
09:58 kados     paul: mainly for you
09:59 kados     I've been reading through lots of code these past few days
09:59 kados     and I'm prepared to invest a lot of time cleaning up for 3.0
09:59 kados     in fact, I've already begun
09:59 kados     however ...
09:59 tumer     I personally believe that we are re-reading the borrower information and issue and item information every time we issue which slows things
10:00 kados     I strongly suggest that we include a plugin for 2.4 for Zebra as per our original plan
10:00 kados     I have to get NPL running Zebra asap (next few weeks)
10:01 kados     and to do that, I'm going to be merging dev-week code, and head code into rel_2_2
10:01 kados     (and asking tumer for some help as I'm having some trouble ;-))
10:02 tumer     :-)
10:02 paul      mmm... i'm really afraid with this idea.
10:02 paul      maybe another option would be to have :
10:02 paul      - Koha 2.2.6
10:02 kados     paul: you might not be if you hear how I plan to implement it
10:02 tumer     so am I
10:02 paul      - Koha 2.4.0
10:02 paul      - Koha 3.0
10:02 paul      ok, i wait for your proposal
10:03 kados     what I propose, is that it be implemented as a separate tar file
10:03 kados     with utilities to automatically perform the upgrade
10:03 kados     (a true plugin)
10:03 kados     so we completely leave 2.4 code alone in the release
10:03 kados     but we announce a plugin is available for zebra
10:04 kados     paul: what do you think?
10:04 tumer     what is 2.4 code (modified 2.2.5?
10:05 kados     tumer: paul was planning to release rel_2_2 as 2.4 once we declare it stable
10:05 paul      I think 2 things :
10:06 paul      1- we agreed that we didn't want to add any new feature to Koha 2.x branch, just fix bugs. IF your plugin, when uninstalled, strictly changes nothing to Koha 2.x, then I could be OK.
10:07 paul      2- "major code cleaning on head" => what does it mean on release time ? because there are many many improvements in head that libraries are waiting impatiently. And I think we could miss something (including customers ;-) ) delaying new features too long.
10:08 paul      however, I agree we NEED a good coding rules & modify head accordingly.
10:08 kados     maybe the solution is:
10:08 kados     rel_2_2 => 2.2.6
10:08 kados     head => 2.4
10:08 kados     newhead => 3.0
10:08 kados     (with code cleaning)
10:09 paul      mmm... not sure, as head is already very unstable
10:09 kados     yep
10:09 kados     the problem I've got
10:09 kados     is I've invested over $50K in Koha this year
10:09 kados     and I'm getting tapped out
10:09 kados     plus, I've got clients who need Zebra
10:10 kados     as much as I hate to say this
10:10 kados     I'm starting to think I'm going to have to maintain a liblime version of Koha
10:11 kados     I think with tumer's help I can get a rel_2_2-based Koha up and running on Zebra
10:11 kados     (at least I hope so, since tumer has one)
10:12 kados     but unless that code is committed to CVS it does noone any good
10:12 tumer     kados: be careful
10:12 kados     careful?
10:12 tumer     one sec phone
10:13 kados     k
10:13 paul      I back to point 2 : what is your timeline for 3.0 ?
10:13 kados     paul: 3.0 isn't timeline based IMO
10:13 paul      it seems we have 3 differents problems on HEAD :
10:13 kados     paul: it's feature based
10:13 paul      - tools & technical backend (like zebra things)
10:13 paul      - code cleaning
10:13 paul      - new features
10:13 kados     yep
10:13 paul      maybe we were wrong starting all of them at the same time.
10:13 kados     to me, new features aren't important
10:13 kados     code cleaning is important for 3.0
10:14 kados     and zebra is #1 for me
10:14 paul      so, maybe we could :
10:14 paul      maybe we could rank all objectives and say :
10:14 paul      - OK, zebra things is working correctly
10:15 paul      - now we do code cleaning & only code cleaning for, say 3 months.
10:15 tumer     kados: my code leaves out marc_tables thus no more searches like fieldxxxsubfieldx for advanced marcseraches
10:15 kados     tumer: my clients are ok with that
10:15 paul      - in 3 month, we start new feature commiting for 3 months
10:15 kados     tumer: and if they need it we can use xpath
10:15 tumer     yep
10:15 kados     paul: it sounds good
10:16 paul      I planned to spend some time on head in the next 2-3 months. It can be for code cleaning.
10:16 kados     right now, my #1 priority, above everything else, is getting a stable Koha going for Zebra
10:16 kados     s/for/with/
10:16 kados     I don't care if the code is super clean as long as it works well
10:16 kados     #2 priority is working on code cleaning for 3.0
10:17 paul      so you're saying code cleaning is not that important ? ;-)
10:17 kados     #3 priority is working on the new features for 3.0
10:17 kados     code cleaning is not important until Zebra is production-ready (IMO)
10:17 kados     but ...
10:17 paul      be careful, because I have some libraries that are waiting for Koha 3.0 impatiently. so we can't delay it too much.
10:18 paul      (most of those features being already in CVS, even if not 100% done)
10:18 kados     something else to consider
10:18 kados     when i was elected as RM for next major version
10:18 kados     it was agreed that 'joshua is not a programmer' ...
10:19 paul      and it seems joshua has a new diploma now ;-)
10:19 kados     hehe
10:19 kados     well, I've learned quite a bit these past few months
10:19 kados     but to be honest, I've been a bit dissapointed at the speed of zebra developemnt for Koha
10:19 paul      right, kados. but we are back to the problem discussed at devweek : having someone working at full time for the project itself !
10:20 kados     even though I've invested over $50K in programming and support costs
10:20 kados     (sub-contracted out)
10:20 paul      same for me if you want to know the truth...
10:20 kados     some of that was sponsored, but large portions of it have come from LibLime's budget
10:20 paul      (disapointing, not investing 50K$)
10:21 kados     (right)
10:21 kados     so we're in a bit of a pickle now
10:21 paul      bit of a pickle ?
10:21 kados     it's an expression ... meaning ...
10:21 kados     bit of trouble ...
10:22 tumer     kados:zebra in itself is a major improvement but everyone else but me is working on other improvements. Priorities?
10:22 kados     tumer: right
10:22 paul      tumer : I worked for some time on zebra as well.
10:22 kados     yep ...
10:22 tumer     paul: I know but you stopped
10:23 paul      but we face a problem here, already encountered 10000 times : a library want to fund something that is seen, and zebra is not really seen...
10:23 kados     the bottom line is, I don't have the expertise to put everything together and come up with a stable Koha with Zebra with what we've got now
10:23 kados     As the RM, I'm relying on programmer expertise from the dev team
10:23 paul      tumer: right. I won't tell you too long why, but I got a contract that was really a pain. it's 99% done now.
10:24 kados     fortunately, tumer came along
10:24 paul      that's why i'm expecting getting some time to work on head
10:24 kados     otherwise, I've no idea what we would have done ... zebra would be delayed for a long time I expect
10:24 paul      (my position is opposite to kados : I invest time when i have money. I don't invest money to have time ;-) )
10:24 paul      kados: right, but at the beginning, the zebra stuff was supposed to be easy. or at least easier.
10:25 tumer     I hane neither :-(
10:25 kados     yep
10:25 tumer     s/hane/have
10:26 tumer     kados:zebra learning curve is steep but easy once climbed
10:26 paul      tumer: not exactly : you're sure to have money at end of month, from your employer. If I don't earn money, strictly no money...
10:26 paul      so you have money, but no time at all.
10:26 paul      fortnuatly, you odn't need to sleep, like joshua.
10:26 kados     and while tumer's commits to dev-week are very helpful, the cvs of dev-week doesn't just work out of thebox
10:27 kados     there are tons of things that need to be tweaked, config files, directories, indexing, scripts to run on the db, etc.
10:27 kados     there is no OPAC ...
10:28 tumer     kados:without changes to database my code will not work its a hybrid 2.20=>3.0 database I'm handling
10:28 kados     tumer: yep, I completely understand that
10:29 kados     tumer: while you're here, I suppose I should mention a problem I'm having with dev-week
10:30 tumer     kados:by the way zebra is behaving very well lately.No dropouts except when I have non-utf8 chars in records
10:30 kados     excellent
10:30 tumer     kados:shoot
10:31 kados     for some reason, it's not finding the zebra server
10:31 kados     and the fallback mechanism isn't working
10:31 kados     it just creates a msyql process that lasts forever
10:31 tumer     very strange?
10:32 tumer     can you reach the zebraserver from yaz client?
10:32 kados     tumer: http://zoomintranet.liblime.com/cgi-bin/koha/opac-catalogue-home.pl
10:32 kados     tumer: yep ... but I do get one strange error
10:32 kados     lemme pull it up
10:33 kados     00:33:16-26/05 zebrasrv(1) [log] term="new" nn=662 type=void count=626
10:33 kados     00:33:16-26/05 zebrasrv(1) [log] user/system: 0/0
10:33 kados     00:33:16-26/05 zebrasrv(1) [request] Search OK 626 2 1+0 RPN: @attrset Bib-1 new
10:33 kados     00:33:23-26/05 zebrasrv(1) [warn] Couldn't open record.abs [No such file or directory]
10:33 kados     00:33:23-26/05 zebrasrv(1) [log] user/system: 0/0
10:33 kados     00:33:23-26/05 zebrasrv(1) [request] Present OK - 2 1+1
10:33 kados     00:34:18-26/05 zebrasrv(1) [session] Connection closed by client
10:33 kados     00:34:18-26/05 zebrasrv(1) [log][app2] zebra_register_close p=0x80b8898
10:33 kados     I'm also confused about the database naming with the new koha.xml
10:33 kados     should the biblioserver and authorityserver both be 'default'?
10:34 tumer     the problem seems to be that your /tabs folder is missing or zebra.cfg does not point to it properly.
10:35 kados     hmmm
10:35 kados     I have:
10:35 kados     profilePath:${srcdir:-.}:/usr/share/idzebra/tab/:${srcdir:-.}/tab/
10:35 kados     and in the dir where zebra.cfg is
10:35 kados     I have tab/record.abs
10:36 tumer     hmm one sec
10:36 kados     (also, I'm confused about the user/password, is it supposed to be in zebra.cfg/kohails or in the koha.xml file? or both?)
10:38 kados     I tried copying it to the dir where zebra.cfg is but it still doesn't work
10:38 tumer      remove the trailing // after tab.Mine does not have it .I'll walk you in koha.xml in aminute
10:39 kados     tumer: do I need to reindex?
10:39 kados     tumer: or just restart zebra?
10:40 kados     tried removing the / and restarting zebra same error
10:40 tumer     if it did not read tab it should not have indexed. So try restarting zebra
10:40 tumer     I mean reindex
10:40 kados     ok ...
10:40 kados     question about file locations
10:41 kados     I have /koha/etc
10:41 kados     with:
10:41 kados     koha.xml
10:41 kados     pqf.properties
10:41 kados     zebra.cfg
10:41 kados     /koha/etc/tab
10:41 kados     default.idx  kohalis  record.abs  sort-string-utf.chr
10:43 tumer     and where does the zebra db reside?
10:43 kados     then I've got (for the zebra database):
10:43 kados     /koha/zebradb
10:43 kados     that has:
10:43 kados     authorities  biblios  lock  records  register  shadow  tmp
10:43 kados     (all directories)
10:43 kados     records/koha.iso2709
10:44 kados     can we name the databases:
10:44 kados     authorities
10:44 dewey     i heard authorities was authority records
10:44 kados     biblios
10:44 kados     items
10:44 kados     (and eventually reservoir)
10:44 kados     ?
10:45 kados     I'm confused about running multiple databases at the same time
10:45 tumer     you miss 2 folders for zebraopfiles
10:45 kados     wondering if the indexes, registers, shadow, lock and tmp need to all be in different locations
10:45 tumer     specialUpdate and recordDelete (both folders)
10:46 kados     where do those live?
10:46 kados     (is there a config file setting for them?)
10:47 tumer     running 2 different databases from the same location I found to be tricky and confusing
10:47 kados     (I'd like to get a standard filesystem layout for koha:zebra
10:47 kados     something like:
10:47 kados     zebradb/
10:47 kados     authorities biblios items reservoir
10:47 kados     then, each of them has:
10:47 kados     register
10:47 kados     shadow
10:47 kados     lock
10:47 kados     tmp
10:48 kados     errors/ (with specialUpdate and recordDelete)
10:48 kados     is that possible?
10:48 tumer     so I have zebrabiblios folder containing folders: tmp lock record register shadow specialUpdate  recordDelete
10:48 kados     then, when installing Koha
10:48 kados     all the user has to specify is, where is 'zebradb'
10:48 tumer     and zebraauthorities folder with similar folders
10:48 kados     ahh, exactly what I thought
10:48 kados     great
10:49 tumer     sorry both of them also have a tab folder
10:49 kados     really!
10:49 kados     hmmm ... maybe that's my problem
10:49 tumer     because record.abs is differnt for both
10:49 kados     ahh, I didn't realize that
10:50 tumer     I have commited record.abs to head under a new folder
10:50 tumer     its called zebraauthority
10:50 kados     ok
10:51 kados     so is it required to have a tab dir in the same dir as zebrabiblios?
10:52 tumer     Not necessarily but makes things simpler with srcdir/tab declaration
10:52 kados     so to index, I go:
10:52 kados     cd /koha/zebradb
10:52 kados     (where there is a records folder containing npl.iso2709
10:52 kados     )
10:52 kados     then:
10:53 tumer     I also put the zebra.cfg of each into their directories
10:53 kados     zebraidx -g iso2709 -c /koha/etc/zebra.cfg -d biblios update records
10:53 tumer     so that your code above can find it more easily
10:53 kados     that will create a database called 'biblios' with the records in 'records', right?
10:54 tumer     yes thats right unless your zebra.cfg actually defines where to look for /tab folder
10:54 tumer     s/unless/if
10:54 kados     hmmm
10:54 kados     but ...
10:55 kados     zebra.cfg defines the location of register, tab, etc.
10:55 kados     isn't that supposed to be different for each database?
10:55 kados     is there a way to specify: database:register:/path:sizeG ?
10:55 tumer     yes differnt thats why together with all db
10:55 kados     ?
10:55 tumer     ?
10:56 kados     my zebra.cfg has:
10:56 kados     register: /koha/zebradb/register:4G
10:56 kados     setTmpDir: /koha/zebradb/tmp
10:56 tumer     yes
10:56 kados     keyTmpDir: /koha/zebradb/tmp
10:56 kados     but ...
10:56 kados     register is different if it's biblios or authorities
10:56 kados     register should be:
10:57 kados     /koha/zebradb/biblios/register (for biblios)
10:57 kados     /koha/zebradb/authorities/register (for authorities)
10:57 kados     so do I need two zebra.cfg files?
10:57 kados     or can I specify that in just the one?
10:57 tumer     yes one for each
10:57 kados     ok ... in that case
10:57 kados     I propose:
10:57 kados     zebra-biblios.cfg
10:57 tumer     so 2 zebra.cfg's
10:57 kados     zebra-authorities.cfg
10:58 kados     both live in /koha/etc/
10:58 kados     so to index biblios I go:
10:58 kados     zebraidx -g iso2709 -c /koha/etc/zebra-biblios.cfg -d biblios update records
10:58 kados     for authorities I go:
10:58 kados     zebraidx -g iso2709 -c /koha/etc/zebra-biblios.cfg -d authorities update records
10:58 kados     make sense?
10:58 tumer     exactly
10:58 kados     ok, I'll try that
10:59 tumer     first get an indexed zebradb then we'll get zebrasrv started with koha.xml properly set up
11:02 kados     tumer: when I start zebra, which zebra.cfg do I use?
11:03 tumer     you mean zebraserver?
11:03 kados     tumer: does it matter? (do I need to start zebra with both)?
11:03 kados     yea
11:03 tumer     yes
11:03 kados     in koha.xml
11:03 kados     there is a line for zebra.cfg
11:03 kados     but now I have two different zebra.cfg
11:03 kados     one for biblios and one for authorties
11:03 kados     or do I need also a third?
11:04 kados     for starting the zebra server?
11:04 tumer     your koha.xml should define them both and start zebrasr -f/koha/etc/koha.xml
11:04 tumer     slow down I can not write that fast
11:05 tumer     did you index anything?
11:05 tumer     or you will try later? shall I continue?
11:05 kados     tumer: ok, it's working now
11:05 kados     tumer: yaz-client is returning results
11:06 kados     now I'll test the search in koha again
11:06 tumer     your page gives me 404 on search
11:08 kados     ok, that's fixed, now it's just hanging again
11:08 kados     lemme check mysql
11:09 kados     yea, same as before:
11:09 kados     | 155 | kohaadmin | localhost | oldkoha | Query   | 65   | Copying to tmp table | SELECT DISTINCT B.biblionumber AS biblionumber , ( MATCH (title,seriestitle,unititle,B.author,subjec |
11:09 kados     need to kill that process manually
11:09 kados     zebra's not getting called
11:10 tumer     is everything updated as head?
11:11 kados     I'm running dev-week
11:11 kados     with Context.pm from head
11:11 tumer     and koha.xml from head?
11:12 kados     yep
11:12 kados     here's in the apache logs:
11:12 kados     Error 10000: Connect failed
11:12 kados     DBD::mysql::st execute failed: Server shutdown in progress at /koha/intranet/modules/C4/Search.pm line 607.
11:12 kados     DBD::mysql::st execute failed: Lost connection to MySQL server during query at /koha/intranet/modules/C4/Search.pm line 615.
11:12 kados     DBD::mysql::st fetchrow_hashref failed: fetch() without execute() at /koha/intranet/modules/C4/Search.pm line 616.
11:12 kados     DBD::mysql::st fetchrow failed: fetch() without execute() at /koha/intranet/modules/C4/Search.pm line 642.
11:12 tumer     and yor apache logs should give you the zebra errors as well
11:13 kados     those are the only errors I get
11:13 kados     zebra never gets the request
11:13 kados     probably because 'Error 10000: Connect failed'
11:14 tumer     error 1000 is usually when there is no zebrasrv running -*veery strange*
11:15 tumer     can you read me the listen lines of your koha.xml better still send me a copy
11:16 kados     k ... which email ... yahoo?
11:16 kados     I don't remember your email at yahoo
11:16 tumer     tgarip@neu.edu.tr
11:16 kados     last time that one took several weeks to get a message from me ;-)
11:16 tumer     try again its still the fastest
11:17 kados     tumer: sent
11:17 tumer     dont worry I'll get you running before going to milonga tonight
11:17 kados     tumer: thanks!
11:17 kados     actually I could just post it
11:18 kados     tumer: http://zoomintranet.liblime.com/koha.xml
11:23 kados     don't know if it helps, but in yaz-client I must go:
11:23 kados     open localhost:9900/biblios
11:23 tumer     sorry this configuration will not work with your biblio.pm I have to send you a new one give me an email
11:23 kados     jmf@liblime.com
11:23 kados     but search.pm is what we're having trouble with now right?
11:24 kados     well actually, connect is probably in context.pm
11:24 tumer     I have changed biblio.pm to accept this new koha.xml but did not commit it waiting your dev-week commits. ý mean search.pm & biblio.pm
11:24 kados     ok
11:24 kados     if you commit to dev-week that would work
11:25 kados     or just email
11:25 tumer     wait till I get things from the server. 2 min
11:25 kados     k
11:33 tumer     I'sent the search.pm
11:34 tumer     I'll commit a newer context.pm biblio.pm once we get this working
11:35 tumer     kados: I newer got your mail by the way
11:35 kados     sounds good, thanks
11:37 tumer     any success?
11:37 kados     tumer: new dependency, Date::Calc
11:37 kados     installing now
11:40 tumer     why did I need that I wonder?
11:40 kados     hmmm
11:40 kados     it was in Search.pm
11:40 kados     so ...
11:40 kados     still getting:
11:40 kados     Error 10000: Connect failed
11:40 kados     and that hanging mysql process
11:41 kados     why isn't it falling back on the old koha tables I wunder
11:41 tumer     very strange indeed.
11:42 tumer     I'm getting dev week stuff now to test and see whats happening
11:42 kados     was there a name change in the connection object?
11:42 tumer     yes
11:43 kados     that could be it
11:43 kados     Zconn in HEAD
11:43 kados     what was it in dev-week?
11:43 tumer     the new change allows Zconn to connect to either biblioserver or authority server
11:44 kados     could that be the root of the problem?
11:44 tumer     before it was C4::Context->Zconn now C4::Context->Zconn("biblioserver")
11:45 kados     wasn' it ->new_Zconn before?
11:46 kados     nevermind, I see it now
11:46 kados     hmmm
11:46 paul      bye bye guys. see on on tuesday.
11:46 kados     bye paul
11:46 tumer     bye
11:46 paul      reminder : pierrick & me won't probably be here for monday meeting
11:46 paul      as we have a meeting in Lyon
11:46 kados     yep, got the email, see you on tueesday
11:46 paul      (with the university)
11:46 kados     have a good weekend
11:47 paul      thx
11:47 paul      you too. maybe 2 days off from Koha will be a good thing
11:47 kados     :-)
11:49 kados     tumer: is the name of the zebra biblio server always supposed to be 'biblioserver' cause I've got it as 'biblios'
11:49 kados     ahh, I see in the config.pm now
11:49 kados     it's handled properly based on the koha.xml file
11:49 kados     nice work there!
11:51 tumer     kados:biblios is tha name of yor database biblioserver is not and yes it should stay as that
11:52 kados     I'm throwing some warns into context.pm to see if I can track down where the connections' failing from
11:52 tumer     yes try and get the $port as well and the databaseName
11:53 kados     In Zconn
11:53 kados     Use of uninitialized value in list assignment at /usr/local/lib/perl/5.8.4/ZOOM.pm line 289.
11:53 kados     Use of uninitialized value in subroutine entry at /usr/local/lib/perl/5.8.4/ZOOM.pm line 295.
11:53 kados     Error 10000: Connect failed
11:53 kados     warn in new_Zconn
11:53 kados     sec, I'll get the port, etc.
11:57 kados     interesting
11:57 kados     In new_Zconn localhost
11:57 kados     Use of uninitialized value in list assignment at /usr/local/lib/perl/5.8.4/ZOOM.pm line 289.
11:57 kados     Use of uninitialized value in subroutine entry at /usr/local/lib/perl/5.8.4/ZOOM.pm line 295.
11:57 kados     Error 10000: Connect failed
11:57 kados     is all I get when I put in:
11:57 kados     my ($tcp,$host,$port)=split /:/,$context->{"listen"}->{$server}->{"content"};
11:57 kados     warn "In new_Zconn ".$context->config('hostname')." ".$port.$context->{'config'}->{$server}."\n";
11:58 kados     so it must not be getting the port properly
11:58 kados     or the context
11:58 tumer     yes not reading your xml file properly
11:58 kados     hmmm
11:59 tumer     try using warn ($port,$context->{'config'}->{$server}) with no quotes so we get actual values