← Previous day | Today | Next day → | Search | Index
All times shown according to UTC.
| Time | Nick | Message |
|---|---|---|
| 00:46 | Waylon joined #koha | |
| 00:46 | Waylon | hey all. |
| 00:48 | what determines the display format of search results? | |
| 00:48 | is there a xslt file involved? | |
| 00:48 | or something? | |
| 00:49 | * jcamins | stops by while waiting for his bread to bake. |
| 00:49 | jcamins | Waylon: it depends. If you have XSLT mode turned on then yes, XSLT is involved. |
| 00:49 | Otherwise, no. | |
| 00:49 | Waylon | okay, what is XSLT? |
| 00:49 | wahanui | XSLT is a maze of twisty turny passages, all alike. |
| 00:50 | jcamins | lol |
| 00:50 | Waylon | no.. that sounds like zork. |
| 00:50 | jcamins | XSLT is not that hard, wahanui. |
| 00:50 | wahanui | ...but xslt is a maze of twisty turny passages, all alike.... |
| 00:50 | Waylon | grue? |
| 00:50 | xyzzy? | |
| 00:50 | jcamins | eXtensible Stylesheet Language for Transformations. |
| 00:51 | Waylon | xyzzy is cheating! Cheater! |
| 00:51 | xyzzy? | |
| 00:51 | wahanui | xyzzy is cheating! Cheater! |
| 00:51 | Waylon | ah.. exactly what i want to use then. |
| 00:53 | jcamins | wahanui: xslt is also http://en.wikipedia.org/wiki/XSLT and the pages at http://cs.au.dk/~amoeller/XML/transformation/ may be useful for learning the basics. |
| 00:53 | wahanui | okay, jcamins. |
| 00:53 | jcamins | xslt? |
| 00:53 | wahanui | xslt is a maze of twisty turny passages, all alike. or http://en.wikipedia.org/wiki/XSLT and the pages at http://cs.au.dk/~amoeller/XML/transformation/ may be useful for learning the basics. |
| 00:56 | jcamins | Waylon: XSLT has, generally-speaking, replaced "normal" mode. |
| 00:57 | One of the neat features for XSLT is that if the display you want is not at all suitable for inclusion upstream, you can specify your own custom XSLT. | |
| 00:58 | However, if what you need is generally applicable, it's much better to get your changes upstream, so that any future improvements are available to you. | |
| 00:58 | Waylon | right. XSLT happens serverside, right? |
| 00:59 | jcamins | Yes. |
| 01:00 | I believe it is possible to load a remote XSLT stylesheet, though, by using a fully-qualified URL. | |
| 01:00 | Waylon | changes i'll be making, is making the existance of a download, 856u, visible, but the actual url not there, instead, takes one to opac-download.pl?biblio=<biblio> for download tracking purposes, using C4::Log and BorrowerAttributes |
| 01:01 | which probably won't be applicable to general use? | |
| 01:01 | jcamins | Uhhh... |
| 01:01 | Waylon | and therefore, XSLT will be the thing to use. |
| 01:01 | jcamins | frankly, that sounds like a bad idea on so many levels I don't have a response. |
| 01:02 | Waylon | ah? |
| 01:02 | wahanui | ah is probably ? |
| 01:02 | Waylon | whats wrong with it? |
| 01:03 | jcamins | You are eliminating the repeatability of 856s, to start with. |
| 01:06 | Waylon | ah, right. yes, so this is strictly a inhouse mod then, as for the company i work with, only one 856 per record. |
| 01:06 | jcamins | Second, you're trying to use the log for analytics, which I expect is going to run into issues of database size. |
| 01:07 | (C4::Log, I mean, not logs generally) | |
| 01:07 | Waylon | hmm? what issues? |
| 01:14 | jcamins | The size is going to explode, and it's going to be as slow as heck to get any information out of the log. |
| 01:14 | Didn't you say you had a userbase in the 10000s? | |
| 01:15 | Waylon | university students yeah. |
| 01:16 | jcamins | Even just one hit per week from each of those is 500k entries in the log. |
| 01:16 | Waylon | one account per university (for some reason i don't know, but thats the way it is) |
| 01:16 | so i think we actually only have 30 accounts.. i think.. | |
| 01:16 | * Waylon | shrugs. |
| 01:17 | jcamins | Don't forget you can't free up space. |
| 01:17 | Waylon | only by dumping and reinserting yeah. |
| 01:17 | jcamins | And you get to upgrade things how often? |
| 01:17 | upgrade/maintain | |
| 01:19 | * jcamins | is a strong believer in trying to anticipate problems. |
| 01:20 | jcamins | I think you're better off using an actual analytics package, rather than trying to use Koha's auditing functionality for crude analytics. |
| 01:21 | Waylon | we are starting to use piwik, i could make the suggestion. |
| 01:22 | okay, back in 45 minutes. lunch | |
| 01:22 | * jcamins | leaves to eat dinner, probably won't be back. |
| 01:48 | tater joined #koha | |
| 02:35 | Waylon | hmmm. |
| 02:36 | took a old 3.0 biblio db.. upgraded it to master.. used the same zebra config as the old setup... reindexed. | |
| 02:36 | no results. | |
| 02:38 | oooo | |
| 02:38 | the xsl thing. | |
| 02:38 | forgot to create it. | |
| 02:45 | k.. wasn't taht. | |
| 03:18 | after copying over the old xslt files to new koha, reindexing. | |
| 04:31 | cait joined #koha | |
| 04:31 | cait | good morning #koha |
| 05:01 | bye #koha | |
| 06:04 | wajasu | Waylon: i |
| 06:04 | i've got to migrate from 3.0 as well. | |
| 06:05 | Waylon | ... after another reindex.. still no results |
| 06:06 | wajasu | when you run a rebuild_zebra -b -a -v do you see reocrds being dumped. |
| 06:07 | and are you sure the zerbasrv is started with the correct KOHA_CONF and PERL5LIB | |
| 06:14 | Waylon | yes and yes. |
| 06:14 | gotta go, back laters | |
| 06:15 | wajasu | later |
| 06:40 | cait joined #koha | |
| 06:40 | cait | hi #koha |
| 06:44 | mbalmer joined #koha | |
| 06:48 | cait1 joined #koha | |
| 06:57 | cait joined #koha | |
| 08:08 | cait joined #koha | |
| 08:09 | jenkins_koha | Starting build #173 for job Koha_3.8.x (previous build: SUCCESS) |
| 08:43 | drojf | good day #koha |
| 08:44 | drojf left #koha | |
| 08:47 | cait joined #koha | |
| 08:49 | * cait | waves again |
| 08:52 | drojf joined #koha | |
| 09:03 | cait joined #koha | |
| 09:11 | jenkins_koha | Project Koha_3.8.x build #173: SUCCESS in 1 hr 2 min: http://jenkins.koha-community.[…]b/Koha_3.8.x/173/ |
| 09:11 | * jcamins: Bug 8447: Make sure we have enough subfields for broader_headings | |
| 09:11 | * Chris Cormack: Bug 7409 : Missing dependency on libtest-mockmodule-perl | |
| 09:11 | huginn | 04Bug http://bugs.koha-community.org[…]w_bug.cgi?id=8447 critical, P1 - high, ---, gmcharlt, Pushed to Stable , can't save new records / indicator error |
| 09:11 | 04Bug http://bugs.koha-community.org[…]w_bug.cgi?id=7409 normal, P5 - low, ---, robin, Needs Signoff , Missing dependencies for Debian package | |
| 09:11 | jenkins_koha | Starting build #174 for job Koha_3.8.x (previous build: SUCCESS) |
| 09:33 | drojf1 joined #koha | |
| 09:46 | cait joined #koha | |
| 09:59 | cait left #koha | |
| 10:13 | jenkins_koha | Project Koha_3.8.x build #174: SUCCESS in 1 hr 2 min: http://jenkins.koha-community.[…]b/Koha_3.8.x/174/ |
| 10:13 | * colin.campbell: Bug 6976 Close loophole allowing borrowers extra holds via opac | |
| 10:13 | * jonathan.druart: Bug 8002: FIX Change datatype from varchar 1 to 10 for borrower_attribute_types.category_code | |
| 10:13 | * Chris Cormack: Bug 8002 : Database bump | |
| 10:13 | * jcamins: Bug 8743: ZOOM error when merging authority change to biblios | |
| 10:13 | * Paul Poulain: Bug 8743 follow-up, throw a more explanatory error message | |
| 10:13 | * colin.campbell: Bug 8653 remove erroneous whitespace blocking indexing | |
| 10:13 | huginn | 04Bug http://bugs.koha-community.org[…]w_bug.cgi?id=6976 major, P1 - high, ---, colin.campbell, Pushed to Stable , User hold limit can be avoided by users in OPAC |
| 10:13 | jenkins_koha | * tomascohen: Bug 8520 - Authorities summary incorrect on posible duplicate notice |
| 10:13 | huginn | 04Bug http://bugs.koha-community.org[…]w_bug.cgi?id=8002 critical, P2, ---, oleonard, Pushed to Stable , Can't add patron attribute type in newer installation |
| 10:13 | 04Bug http://bugs.koha-community.org[…]w_bug.cgi?id=8743 critical, P3, ---, jcamins, Pushed to Stable , ZOOM error when merging authority change to biblios | |
| 10:14 | 04Bug http://bugs.koha-community.org[…]w_bug.cgi?id=8653 normal, P4, ---, colin.campbell, Pushed to Stable , With DOM indexing turned on subject authority links are not indexed | |
| 10:14 | 04Bug http://bugs.koha-community.org[…]w_bug.cgi?id=8520 critical, P5 - low, ---, jcamins, Pushed to Stable , Authorities display incorrectly in staff results | |
| 10:14 | jenkins_koha | Starting build #175 for job Koha_3.8.x (previous build: SUCCESS) |
| 10:19 | Waylon joined #koha | |
| 10:19 | Waylon | hiya all! |
| 10:41 | took a old 3.0 biblio db.. upgraded it to master.. used the same zebra config as the old setup... reindexed, got no results. help? | |
| 10:45 | mtj | did you follow the upgrade notes in INSTALL.debian ? |
| 10:46 | and run this script? -> misc/maintenance/remove_items_from_biblioitems.pl --run | |
| 10:47 | Waylon | yes, i did. |
| 10:48 | and also ran updatedatabase too, since it found it needed doing. | |
| 10:48 | mtj | have you checked all the obvious stuff? |
| 10:48 | Waylon | whats obvious? |
| 10:49 | mtj | run zebra_rebuild -r -b -x - v , and got no errors |
| 10:49 | and got a sane count of bibs processes, on completion | |
| 10:49 | Waylon | did zebra_rebuild -r -b -a -v -v |
| 10:50 | why -x? | |
| 10:50 | mtj | did it complete ok? or barf? |
| 10:50 | Waylon | said it completed okay. |
| 10:51 | mtj | do yu have a stable koha to compare against? |
| 10:51 | Waylon | [log] Records: 196893 i/u/d 196893/0/0 |
| 10:51 | mtj | that looks good |
| 10:51 | Waylon | hmm? only the koha currently in production. |
| 10:51 | eythian joined #koha | |
| 10:51 | Waylon | thats 3.0 |
| 10:52 | [log] Iterations: isam/dict 291501887/3024638 | |
| 10:52 | mtj | my advice…. sort a stable koha to compare against |
| 10:52 | Waylon | [log] Dict: inserts/updates/deletions: 3024638/0/0 |
| 10:53 | mtj | if you get different outputs to your stable koha, you've got a debuggable problem |
| 10:55 | so, is your zeb-rebuild behaving similar to your stable koha | |
| 10:55 | is your zebrasrv behaving similar to your stable koha | |
| 10:56 | if no, you've got a debuggable problem - if yes, you've got a trickier problem :/ | |
| 10:56 | Waylon | turned on debug level logging in koha-zebra.. |
| 10:57 | huh.. new Zebra doesn't give the query used? | |
| 10:58 | 05:56:56-16/09 zebrasrv(1) [debug] Got PDU, 119 bytes: lead=B4 75 83 | |
| 10:58 | 05:56:56-16/09 zebrasrv(1) [debug] server select: config=/usr/local/apps/koha/kohadev-etc/etc/zebradb/zebra-biblios.cfg | |
| 10:58 | 05:56:56-16/09 zebrasrv(1) [debug] bend_init | |
| 10:58 | 05:56:56-16/09 zebrasrv(1) [debug] zebra_open zs=0x1250570 returns 0x1250e20 | |
| 10:58 | 05:56:56-16/09 zebrasrv(1) [debug] Sort handler installed | |
| 10:58 | 05:56:56-16/09 zebrasrv(1) [debug] Search handler installed | |
| 10:58 | 05:56:56-16/09 zebrasrv(1) [debug] ESRequest handler installed | |
| 10:58 | 05:56:56-16/09 zebrasrv(1) [debug] Delete handler installed | |
| 10:59 | 05:56:56-16/09 zebrasrv(1) [debug] Scan handler installed | |
| 10:59 | 05:56:56-16/09 zebrasrv(1) [debug] Segment handler installed | |
| 10:59 | 05:56:56-16/09 zebrasrv(1) [debug] result immediately available | |
| 10:59 | 05:56:56-16/09 zebrasrv(1) [debug] ir_session (output) | |
| 10:59 | 05:56:56-16/09 zebrasrv(1) [debug] Wrote PDU, 183 bytes | |
| 10:59 | 05:56:56-16/09 zebrasrv(1) [debug] zebra_close zh=0x1250e20 | |
| 10:59 | cait joined #koha | |
| 10:59 | cait left #koha | |
| 10:59 | Waylon | thats all there is. |
| 10:59 | mtj | http://paste.koha-community.org/ |
| 11:00 | Waylon | hm. right.. ah.. too late |
| 11:00 | mtj | so, what happens if you do a search? |
| 11:01 | Waylon | that is from a search |
| 11:01 | from a intranet search | |
| 11:02 | mandumah.com, working server, is zebra 2.0.43, the new server in devel, is zebra 2.0.52 | |
| 11:02 | hmmmmmmm.. | |
| 11:02 | mtj | hmm, perhaps use the new zebra conf files, much has changed since 3.0 |
| 11:03 | Waylon | right, but that would take out all the customization that mandumah did. |
| 11:03 | mtj | do a diff, lots of changes yeah? |
| 11:04 | yep, you would lose those mods :/ but at least you might have a working search… to start with | |
| 11:04 | Waylon | maybe.. okay.. ill swap the configs. |
| 11:04 | mtj | make sense, yeah |
| 11:05 | use git on your zebra files | |
| 11:05 | Waylon | on its way. |
| 11:05 | mtj | start fresh and functioning, then add your mod... |
| 11:06 | when it breaks, your last added mod was the problem | |
| 11:06 | Waylon | if this works, guess i will have to manually go through the marc setup. |
| 11:15 | jenkins_koha | Project Koha_3.8.x build #175: SUCCESS in 1 hr 1 min: http://jenkins.koha-community.[…]b/Koha_3.8.x/175/ |
| 11:15 | Chris Cormack: Bug 7409 : Fixing typo | |
| 11:15 | huginn | 04Bug http://bugs.koha-community.org[…]w_bug.cgi?id=7409 normal, P5 - low, ---, robin, Needs Signoff , Missing dependencies for Debian package |
| 11:15 | jenkins_koha | Starting build #176 for job Koha_3.8.x (previous build: SUCCESS) |
| 11:16 | mtj | hmm, usually no - but if youve heavily tweaked your zebra files and marc framework, …then perhaps? |
| 11:57 | Waylon | mtj, koha master config, fails. |
| 11:58 | still no results | |
| 12:04 | jcamins_away | Sounds like you adjusted your Zebra configuration in a non-compatible way. |
| 12:05 | Waylon | hrm. |
| 12:05 | jcamins_away | Actually, looking at the output you pasted into the channel, sounds like Zebra isn't running. |
| 12:05 | Waylon | i mean, koha master zebra config fails. |
| 12:06 | ah? zebra is running. | |
| 12:07 | jcamins_away | This bread is amazing. |
| 12:08 | Waylon | huh? |
| 12:08 | jcamins_away | In that case, I think I'm back to theory A: you made non-compatible changes to your Zebra config and are storing necessary data in strange places. |
| 12:11 | lodac joined #koha | |
| 12:11 | lodac | Morning |
| 12:13 | I work in the education sector (k-12). Any recommendation of some American based companies that would provide support to a koha user for a district? | |
| 12:13 | paid of-course | |
| 12:13 | jcamins_away | support? |
| 12:13 | wahanui | support is probably found at http://koha-community.org/support/ |
| 12:13 | jcamins_away | paid support? |
| 12:13 | wahanui | please see http://koha-community.org/support/paid-support/ for a list of paid Koha support vendors. |
| 12:13 | jcamins_away | ^^ there is a list, though not everyone is on it (I'm not, for example) |
| 12:14 | lodac | I notice that list for US is 247 days old... so that is why I asked here |
| 12:14 | jcamins_away | That list is kept up to date. |
| 12:15 | lodac | jcamins_away: are you looking for customers? |
| 12:15 | jcamins_away | lodac: responding off-channel |
| 12:17 | jenkins_koha | Project Koha_3.8.x build #176: SUCCESS in 1 hr 1 min: http://jenkins.koha-community.[…]b/Koha_3.8.x/176/ |
| 12:24 | Waylon | any way to examine the contents of the zebradb? |
| 12:24 | without asking a specific query? | |
| 12:27 | jcamins_away | Waylon: not easily, no. |
| 12:29 | Waylon | ooo. |
| 12:29 | got it working | |
| 12:29 | in -v all | |
| 12:29 | mode. | |
| 12:29 | hmmmmm.. | |
| 12:33 | okay, yay. hmm | |
| 12:39 | jcamins_away | Waylon: what was the problem? |
| 12:40 | Waylon | not sure. running it again with the old zebra config. |
| 12:40 | and reindexing | |
| 12:40 | wajasu | maybe you might want to look into the yaz-client? |
| 12:40 | jcamins_away | ^^ good ides |
| 12:40 | *idea | |
| 12:40 | yaz? | |
| 12:40 | wahanui | rumour has it yaz is from IndexData and the Koha XSLT filters are essentially copied from YAZ |
| 12:41 | jcamins_away | yaz client? |
| 12:41 | troubleshooting with yaz client? | |
| 12:41 | wajasu | man yaz and man yaz-client will help |
| 12:42 | jcamins_away | understanding zebra indexing? |
| 12:42 | understanding zebra indexing is http://wiki.koha-community.org[…]ng_Zebra_indexing | |
| 12:42 | yaz client is http://wiki.koha-community.org[…]ng_Zebra_indexing | |
| 13:24 | cait joined #koha | |
| 13:27 | jcamins_away | mtj: 8210 is rebased |
| 13:58 | tcohen joined #koha | |
| 14:08 | Waylon | zebra? |
| 14:08 | wahanui | it has been said that zebra is a fielded free text indexing and retrieval engine with a Z39.50 front-end. You can use any compatible, commercial, or freeware Z39.50 client to access data stored in Zebra. or the search engine koha uses |
| 14:08 | Waylon | zebra troubleshooting is see [understanding zebra indexing] and [yaz client] |
| 14:09 | zebra troubleshooting? | |
| 14:09 | wahanui | well, zebra troubleshooting is see [understanding zebra indexing] and [yaz client] |
| 14:09 | Waylon | okay, bed time. |
| 14:09 | laters all | |
| 14:24 | drojf joined #koha | |
| 14:52 | cait joined #koha | |
| 14:52 | chris_n` joined #koha | |
| 14:55 | cait left #koha | |
| 15:20 | tcohen | hi #koha |
| 15:21 | mtj, have you seen bug 8773? | |
| 15:21 | huginn | 04Bug http://bugs.koha-community.org[…]w_bug.cgi?id=8773 enhancement, P5 - low, ---, tomascohen, Needs Signoff , Add per-instance koha-index-daemon in .deb setup |
| 15:56 | tomas joined #koha | |
| 16:28 | mib_3j8e4w joined #koha | |
| 17:09 | cait joined #koha | |
| 17:25 | mbalmer joined #koha | |
| 17:48 | cait joined #koha | |
| 17:54 | * cait | waves |
| 18:23 | drojf1 joined #koha | |
| 18:31 | mbalmer left #koha | |
| 18:44 | cait | hi drojf :) |
| 19:24 | rangi | @wunder nzwn |
| 19:24 | huginn | rangi: The current temperature in Wellington, New Zealand is 12.0°C (7:20 AM NZST on September 17, 2012). Conditions: Rain Showers. Humidity: 94%. Dew Point: 11.0°C. Pressure: 29.65 in 1004 hPa (Steady). |
| 19:25 | cait | ew |
| 19:25 | @wunder Konstanz | |
| 19:25 | huginn | cait: The current temperature in Taegerwilen, Taegerwilen, Germany is 12.8°C (9:25 PM CEST on September 16, 2012). Conditions: Clear. Humidity: 88%. Dew Point: 11.0°C. Pressure: 29.98 in 1015 hPa (Steady). |
| 19:29 | wajasu | i played with DBIx last night. I used the DBIx::Class::Schema::Loader to completely generate DBIx classes for our schema. wrote some simple queries using the objects. |
| 19:30 | rangi | Yep there's a branch with that all done |
| 19:30 | The schema is all that interests me | |
| 19:30 | wajasu | now i a thinking of using them to help write tests. |
| 19:31 | rangi | Specially ::versioned |
| 19:31 | wajasu | do you want to use the db schema migration stuff? |
| 19:31 | rangi | Hmm I wouldn't ... You wouldn't be testing the actual queries if you did that |
| 19:32 | But yes instead of updatedatabase.pl and the .sql | |
| 19:32 | Use versioned schema | |
| 19:35 | wajasu | plus it has that DateTime component. it would be nice to have time/durections normalized so folks writing scheduling/calender sensitive stuff would have an easier time. |
| 19:35 | durections/durations | |
| 19:37 | i was thinking if someone wrote a OO domain model, that used the DBIx underneath, we could write tests for all the DB layer and the OO domain and services. before the implementation | |
| 19:38 | we could assign bugs to folks to code/wire up the implementations, and uncomment the tests as part of the bugs. then we could get the tests written. | |
| 19:39 | rangi | Yep but in the mean time (next 3 feature releases at least) we need to go from 56 to close to 100 |
| 19:39 | Coverage | |
| 19:39 | Before we start a major rewrite | |
| 19:39 | Or we will break tons more than we fix | |
| 19:39 | cait | good point |
| 19:40 | rangi | Rewrites are fun |
| 19:40 | But gotta do the hard work first | |
| 19:40 | wajasu | ok. so that is the plan. i'm with you. so as i get more familiar with the codebase, i should just write tests so we don't fall behind. |
| 19:41 | rangi | That would be awesome |
| 19:43 | wajasu | for consistency sake, i would like a nice new OO domain model in there. I wish I could test against that with the implementation calling our current stuff. then let people rewrite the implementations, while keeping the tests passing. |
| 19:45 | rangi | OO isn't a panacea and shouldn't be thrown at everything |
| 19:45 | Especially structures that you have to coerce into being an object | |
| 19:45 | wajasu | true. but encapsulation is the goal. composition... |
| 19:45 | rangi | But I mostly agree with u |
| 19:46 | Goal is a fast system for end users | |
| 19:46 | Everything else is a means to that | |
| 19:47 | wajasu | thats why i want to set up a stack with and without layers (with erformance/profiler, etc). |
| 19:47 | rangi | Fast and stable |
| 19:49 | wajasu | i did a reverse engineer of the db a coupld of months ago in java/hibernate, and there a number of schema issues (nullable keys, etc). when i did the same with DBIx, those issue are still there. |
| 19:50 | so as the architects discover them, i guess they can write a bug and baby step the schema to where is need to be. | |
| 19:50 | rangi | Fixing kohastructure.sql would be the first step then |
| 19:52 | My stop bbiab | |
| 20:14 | kathryn joined #koha | |
| 20:17 | wajasu | http://wiki.koha-community.org[…]Coding_Guidelines has a link under (PERL13) POD that is broken these days. |
| 20:23 | jcamins_away | rangi: I think I have to call BS on rewrites being fun. |
| 20:23 | Rewrites suck, and make me grouchy. | |
| 20:23 | rangi | people should stop doing them then |
| 20:23 | or at least stop making them the first thing they do, rather than the last | |
| 20:24 | cos every rewrite of a major feature in koha, to this point | |
| 20:24 | has been a huge pita for the next at least 6 months | |
| 20:25 | jcamins_away | This is true. Unfortunately, C4::Search is still sitting there, mocking our attempts to make any arbitrary input from the user work. |
| 20:25 | cait | jcamins_away: but you started with writing tests for it :) |
| 20:25 | * cait | is optimistic |
| 20:25 | * magnuse | too |
| 20:25 | jcamins_away | cait: I did. |
| 20:26 | And it's a damn fine test, if I say so myself. | |
| 20:26 | magnuse | hehe |
| 20:26 | cait | :) |
| 20:26 | hi magnuse :) | |
| 20:26 | wajasu | i also wish there was separate "test" dependencies. |
| 20:27 | magnuse | hiya cait |
| 20:27 | wajasu | if say i want to use Test::Spec for BDD and write 50 tests. i wouldn't want that to affect production deps. |
| 20:27 | jcamins_away | wajasu: add it as optional. |
| 20:28 | wajasu | ahh |
| 20:28 | jcamins_away | I don't see any problem with that. |
| 20:28 | rangi | and make it not in t/ |
| 20:28 | but lower | |
| 20:28 | because if it is in t/ | |
| 20:28 | jcamins | Right. |
| 20:28 | rangi | its not optional it is mandatory |
| 20:28 | wajasu | gotch. like db_dependent |
| 20:29 | rangi | yep |
| 20:30 | jcamins | Or t/bdd/ |
| 20:31 | wajasu | getting back to the DBIx test idea. i want to use th DBIx::Class::Schema::Loader to generate OO code that is independent from the DB (myslq/postgres), then call our code, but use the DBIx classes to inspect the results. so when we rewrite, only the current calls have to be coded in the given: when: then scenarios. |
| 20:32 | jcamins | wajasu: if that technique can test functionality *now* it sounds like a fine idea. |
| 20:32 | wajasu | and the tests are more db independent |
| 20:33 | i'll try it out on something simple, then something complex and you let me know if its good to go. | |
| 20:33 | jcamins | wajasu: as long as the tests give meaningful results, I will support the inclusion in Koha. |
| 20:37 | wajasu | ok. |
| 20:40 | oh. by the way tried that XML::Easy and replaced XML::Simple. Did the load_testing/benchmark_staff.pl and got the same results. | |
| 20:41 | i did notice: the zebra config sections have XIncludes and go grab those files and parse them as well. (XML::Simple does I assume), not that it makes a difference. | |
| 20:42 | jcamins | wajasu: actually, the Zebra config sections are not used, I think. |
| 20:43 | marcelr just reported that. | |
| 20:43 | wajasu | i wonder if creation of a new db connection is the issue, or if the fact that many places don't dbh->disconnect, causes mysql to "timeout" before cleanup. |
| 20:44 | jcamins | wajasu: huh? |
| 20:44 | the first question? | |
| 20:44 | wahanui | "What are you trying to do?" "What is the goal?" or "What problem are you experiencing?" |
| 20:46 | wajasu | i was in C4, and grepped for dbh->disconnect. I read that CGI supposedly cleans up? but i also read on a site that if we don't cleanup the resource at the end of scripts, the database may have to "timeout". so since i was running the apachebench benchmark, i wonder. |
| 20:46 | jcamins | wajasu: so what is the problem you've encountered? |
| 20:47 | wajasu | i'm wondering why/how that bug wants to get more performance. |
| 20:48 | the removal of XML::Simple. I understand about memcache probably taking care of it. | |
| 20:48 | jcamins | What bug? XML parsing has nothing to do with connecting to MySQL. |
| 20:49 | wajasu | bug 7170. |
| 20:49 | wahanui | bug 7170 is what I was looking at |
| 20:49 | huginn | 04Bug http://bugs.koha-community.org[…]w_bug.cgi?id=7170 normal, P3, ---, paul.poulain, REOPENED , Remove use of XML::Simple |
| 20:49 | jcamins | Right, that has nothing to do with MySQL. |
| 20:49 | That is about how long it takes to parse koha-conf.xml. | |
| 20:51 | wajasu | i left the XML::Simple in, and added the parsing with XML::Easy and the benchmark ran with the same time. |
| 20:51 | jcamins | Right, it's not an important bug. |
| 20:55 | wajasu | i was trying to understand if there was a perfomance issue setting up Context.pm "too many queries, plus perl startup, plus db connection(no connection pool). And if that was the driving force to plackify. If so, we should code a plackified context and benchmark. |
| 20:57 | jcamins | C4::Context is not very efficient, but the bigger issue is the overall load time. |
| 20:58 | The difference between 22ms and 20ms just isn't that big a deal when you're talking about a total load time of 400ms+. | |
| 20:59 | If you want to see what I'm talking about, run the following command: perl -MDevel::NYTProf -e "use C4::Search; use C4::Circulation; print 'Hello, world!';" | |
| 20:59 | Then take a look at the results. | |
| 21:00 | Obviously the faster your computer the less dramatic it will be. | |
| 21:00 | wajasu | ok. i will. |
| 21:01 | you mean the load time for perl. right? | |
| 21:01 | jcamins | Load time for Koha's modules. |
| 21:01 | wajasu | ahh. |
| 21:01 | jcamins | I think I rembered the correct modules for the most dramatic effect. |
| 21:01 | *remembered | |
| 21:02 | Jared joined #koha | |
| 21:03 | jcamins | ^^ not me |
| 21:03 | rangi | heh |
| 21:04 | Jared | Hello, I'm looking for information on Koha - Zebra permissions. I'm installing on Fedora and am getting "Search biblios ERROR 109" from zebrasrv. |
| 21:05 | jcamins | Jared: have you reindexed yet? |
| 21:06 | Jared | files in "/var/lock/koha/zebradb/" are owned by the system user "koha". Is this the same user as what is in /etc/koha/zebradb/zebra-biblios.cfg? |
| 21:07 | Have I reindexed? I think so. I have run "/usr/share/koha/bin/migration_tools/rebuild_zebra.pl -v -v -b -a" I can send the output if that is helpful. | |
| 21:07 | jcamins | Jared: no. Zebra user and system user are unrelated. |
| 21:08 | Jared | This is a new install...not an upgrade. I did put one card/item in the catalog. |
| 21:08 | jcamins | Jared: stick the output in paste.koha-community.org. |
| 21:08 | Full disclosure: I dislike non-Debian systems because they don't play nicely with Koha. | |
| 21:09 | pastebot | "Jared" at 127.0.0.1 pasted "rebuild script output" (29 lines) at http://paste.koha-community.org/175 |
| 21:09 | jcamins | Umm, add another -v. |
| 21:10 | Jared | jcamins, can you describe how the permissions work? HTTPd runs as user "apache"...that then calls out to zebrasrv. zebrasrv (somehow) connects to actual files on the filesystem. |
| 21:11 | I added another -v, output is the same. I can re-paste | |
| 21:11 | jcamins | No need. |
| 21:11 | See above about my hatred of non-Debian systems. ;) | |
| 21:11 | zebrasrv is running as your koha user. | |
| 21:11 | Jared | Is there any concern about using a version of zebra that is too new? |
| 21:11 | jcamins | The CGI script connects via socket. |
| 21:12 | No. | |
| 21:12 | Jared | yes, zebrasrv is being run after "su koha -" |
| 21:13 | jcamins | You must run rebuild_zebra as that same user. |
| 21:14 | All Zebra files must be owned by that user. | |
| 21:14 | However, 109 isn't a permissions error, I don't think... that sounds like a "Zebra isn't being run with the correct configuration" error to me. | |
| 21:14 | Jared | I did run rebuild_zebra as that user. |
| 21:15 | as far as I can tell, everything under /var/*/koha is owned by koha.koha | |
| 21:15 | jcamins | Yeah, so I think the problem you're running into is that Zebra is failing to use your configuration file. |
| 21:15 | Wait, how are you starting Zebra? | |
| 21:16 | Jared | how am I starting zebrasrv? For now just on the command-line: "/usr/local/bin/zebrasrv -v all -f /etc/koha/koha-conf.xml" |
| 21:16 | * jcamins | adds for the benefit of those just joining it that "Jared" is not me. |
| 21:17 | jcamins | Paste that into paste.koha-community.org? |
| 21:17 | *in | |
| 21:17 | Jared | I can paste the output of ls -l on /var/*/koha...but all of the permissions look ok |
| 21:17 | I'll paste the zebrasrv output first. | |
| 21:17 | jcamins | No, nothing to do with permissions. |
| 21:17 | You should be getting a ton of output from zebrasrv. | |
| 21:18 | pastebot | "Jared" at 127.0.0.1 pasted "zebrasrv output showing error 109 (Fedora)" (109 lines) at http://paste.koha-community.org/176 |
| 21:19 | Jared | Is debian the only "officially" supported OS distro for koha? |
| 21:20 | jcamins | Jared: I don't know about "official," but certainly you'll get a lot more help with Debian/Ubuntu. |
| 21:21 | Jared: try using yaz client directly. | |
| 21:21 | yaz client? | |
| 21:21 | wahanui | yaz client is http://wiki.koha-community.org[…]ng_Zebra_indexing |
| 21:21 | Jared | same error if I use Yaz client. I can do it again, but previously, I got identical stuff in the zebrasrv log |
| 21:21 | jcamins | I want to know what it says *before* you do the search. |
| 21:22 | When you connect to the server and set the database. | |
| 21:24 | Jared | from the yaz-client? |
| 21:24 | jcamins | Right. |
| 21:24 | wajasu | http://koha.1045719.n5.nabble.[…]em-td3072111.html |
| 21:24 | jcamins | It should give you some sort of diagnostics, as I recall. |
| 21:25 | Jared | there is some extra info, but it appears (to me) to connect to zebrasrv ok |
| 21:26 | jcamins | Jared: hm, nothing about unrecognized option? |
| 21:27 | Jared | I did paste the yaz-client output...but didn't see the paste link. I'll paste it again. |
| 21:28 | pastebot | "Jared" at 127.0.0.1 pasted "yaz-client output" (24 lines) at http://paste.koha-community.org/178 |
| 21:32 | jcamins | Jared: try doing the database init thing. |
| 21:32 | You won't get any results, but you might not get the error after it's done. | |
| 21:32 | faq? | |
| 21:32 | wahanui | well, faq is found at http://koha-community.org/documentation/faq/ |
| 21:32 | Jared | database init is the reindex_zebra script? |
| 21:33 | jcamins | No. |
| 21:33 | At the end of that file it has a suggestion of resetting your Zebra indexes. | |
| 21:34 | Under "I have double ups..." | |
| 21:34 | Oh, the init was removed. | |
| 21:34 | Jared | can you clarify "the database init thing" Are you referring to "koha-rebuild-zebra"? |
| 21:35 | jcamins | So, run that drop command, then do: zebraidx -c /path/to/zebra-biblios.cfg -d biblios init |
| 21:35 | I am telling you what to do. | |
| 21:35 | Then do the commit. | |
| 21:36 | Jared | ok. Do this as the koha user, I assume |
| 21:36 | jcamins | Yes. |
| 21:36 | What version of Koha are you using? | |
| 21:36 | I should have asked this first. | |
| 21:40 | pastebot | "Jared" at 127.0.0.1 pasted "biblios rebuild" (48 lines) at http://paste.koha-community.org/179 |
| 21:40 | jcamins | Ah-ha. |
| 21:40 | This is helpful. | |
| 21:40 | Let me figure out what it helped us find out. | |
| 21:41 | Jared | interestingly, it says there was no biblios db when dropping it. |
| 21:41 | I have installed Koha koha-3.08.04 | |
| 21:42 | rangi | id do the create before rebuild_zebra see if that fixes it |
| 21:42 | wajasu | database unavailable: the user/pass OR the perms usually. following the KOHA_CONF to the koha-conf.xml /var/lib/koha/... make sure the user you run zebrasrv as has perms to read/write all the way down that heirarchy. |
| 21:45 | jcamins | wajasu: this isn't the standard no permissions error, though, I don't think. |
| 21:45 | Jared: share ls -l /var/lib/koha? | |
| 21:45 | Jared | wajasu: Yes, everything under /var/lib/koha is koha.koha |
| 21:45 | jcamins | ls -lR |
| 21:45 | pastebot | "Jared" at 127.0.0.1 pasted "/var/lib/koha permissions" (65 lines) at http://paste.koha-community.org/180 |
| 21:47 | jcamins | Interesting way of doing ls -lR. :) |
| 21:49 | pastebot | "jcamins" at 127.0.0.1 pasted "Jared (not me): try this" (2 lines) at http://paste.koha-community.org/181 |
| 21:50 | jcamins | Jared: as you probably gathered, my name is also Jared, and it's kind of throwing me. ;) |
| 21:51 | pastebot | "Jared" at 127.0.0.1 pasted "Output of zebraidx commands. zebrasrv still returns error 109" (52 lines) at http://paste.koha-community.org/182 |
| 21:52 | Jared | It's good to know there's another Jared out there. There aren't too many |
| 21:53 | jcamins | When I was younger, I had a class with another Jared. |
| 21:55 | I have no idea what's going on there. | |
| 21:56 | Jared | So, I'm trying to recap. You had be run 4 zebraidx commands in an attempt to rebuild zebra-biblios and zebra-biblios-cfg |
| 21:56 | sorry, zebra-biblios and zebra-biblios-dom | |
| 21:57 | jcamins | The first set was incorrect. |
| 21:57 | I caught that after the fact. | |
| 21:57 | It should have been zebra-biblios-dom all along. | |
| 21:57 | Jared | Some more context for you, I have fedora, have installed all of the perl modules (via cpan) |
| 21:58 | I have yaz-3.0.49-1.fc13.x86_64 instaled as an RPM via yum | |
| 21:58 | I have idzebra-2.0.52 installed via a tarball | |
| 21:59 | and I have koha 3.08.04 installed. When installing, I accepted all of the defaults from Makefile.PL | |
| 22:00 | jcamins | Oh, I bet I know what it is. |
| 22:00 | Jared | I did run the install twice (make uninstall and re-ran gunzip & tar) Same effect both times. |
| 22:00 | jcamins | Did you open up the zebra-biblios-dom.cfg file and make sure that all the paths listed match where your files ended up? |
| 22:00 | In particular, I think there's a modulePath directive that is likely to be wrong. | |
| 22:02 | Jared | yes that path is wrong |
| 22:03 | jcamins | Well, it took a while, but I knew we'd find it eventually. :) |
| 22:06 | Jared | still getting error 109 |
| 22:07 | jcamins | Well, yes, that wasn't what was causing the problem. |
| 22:07 | Jared | sorry, I'm on the phone for a few minutes.... Appreciate your ideas. The module path was definitely wrong |
| 22:15 | Hi, sorry about that. | |
| 22:15 | jcamins | We need a "select all" button in the webinstaller. |
| 22:16 | Jared | Do I need to drop and rebuild the databases and rerun the rebuild_zebra script? |
| 22:16 | jcamins | Or to make the test clickable. |
| 22:16 | Yeah, repeat the init thing. | |
| 22:16 | Jared | ok, will do |
| 22:16 | do authorites-dom and biblios-dom or all 4? | |
| 22:17 | jcamins | zebra-biblios-dom |
| 22:17 | The init and commit commands. | |
| 22:18 | Hey, the password field on master is an actual password box. | |
| 22:18 | Nice! | |
| 22:20 | rangi | 3.8.x too |
| 22:20 | jcamins | Oh, I haven't created any new users in my 3.8.x installs. |
| 22:21 | pastebot | "Jared" at 127.0.0.1 pasted "Rebuild steps" (3 lines) at http://paste.koha-community.org/183 |
| 22:21 | Jared | hmm, so despite the changes, I still get error 109. |
| 22:22 | jcamins | And you got no output? |
| 22:22 | Jared | one other thing I did notice, is that by default, stuff is installed to /usr/local/bin, which is not in the default path on fedora. |
| 22:22 | jcamins | Also, try just those first two lines. |
| 22:25 | mib_05oev3 joined #koha | |
| 22:30 | cait left #koha | |
| 22:31 | Jared | I think the takeaway is that this zebra stuff just doesn't work on fedora/centos/rhel. I'd encourage you to document this somewhere |
| 22:31 | I'll have to setup a debian install and test it again there. | |
| 22:32 | jcamins | Jared: I think we've documented it repeatedly. |
| 22:32 | And it's not that it can't be made to work. | |
| 22:32 | Jared | then my apologies, I missed it. What is the best supported distro...latest debian? |
| 22:33 | jcamins | One developer uses Fedora. |
| 22:33 | Jared: no, that was documented, past tense. I don't think it's so much that you missed it as the "use Debian" recommendation was replaced with "here's how I got an old version of Koha to almost run on Fedora." | |
| 22:33 | Yeah, Squeeze is the way to go. | |
| 22:33 | packages? | |
| 22:33 | wahanui | packages is, like, at http://wiki.koha-community.org/wiki/Debian |
| 22:34 | jcamins | ^^ follow that guide. |
| 22:34 | It's dead simple. | |
| 22:46 | wajasu | http://www.loc.gov/z3950/agenc[…]fns/bib1diag.html |
| 22:46 | i always have to run zebra as root. | |
| 22:47 | jcamins | wajasu: then you have your system set up wrong. |
| 22:47 | If you're satisfied with that, though, I suggest you just leave it. | |
| 22:47 | wajasu | i could not infer it fromthe INSTALL instructions. 4 years ago. |
| 22:48 | the problem was apache was so different from debian | |
| 22:49 | none of the same command line tools that apache has. but i was on archlinux. | |
| 22:49 | i'm about to upgrade to 3.10 once that's released. | |
| 22:50 | and run debian in a VM. | |
| 22:52 | it would take 4hours to assemble dependencies from cpan, and build archlinux packages. perl didn't have but a couple of koha's deps. so i saw the light. debian. | |
| 23:18 | papa joined #koha | |
| 23:27 | libsysguy joined #koha | |
| 23:30 | libsysguy joined #koha | |
| 23:42 | wajasu | jcamins: that nyt report is great. thx |
| 23:42 | jcamins | wajasu: you're welcome. Kind of extraordinary, isn't it, how long it takes for Koha modules to load? |
| 23:43 | wajasu | yup. |
| 23:48 | i got 1.29secs total with koha running in a VM with /home/koha/kohadev in the VM filesystem and /home/koha/kohaclone sshfs mounted to my hostOS. | |
| 23:48 | jcamins | Heck of a lot faster than on my laptop. |
| 23:48 | 2.57s. | |
| 23:49 | wajasu | so the more deps the slower. but if they plackify, they are preloaded, and only what is needed is run in on plackified scripts. right ? |
| 23:50 | jcamins | Right. |
| 23:50 | The BEGIN blocks are run exactly once under Plack. | |
| 23:54 | wajasu | i am really glad you are getting search up to par. i hope with your future search plugins, you make a composable plugin, where one can stack them, and configure it so that (a) if no results, run the second, (b) run in parallel with immediate results from one, and asynchronous ajaxified rsults arriving right behind. |
| 23:55 | but i know there is alot of work to do before that. | |
| 23:55 | jcamins | I am designing the system so it would be possible to do that. |
| 23:56 | wajasu | i could tell. |
← Previous day | Today | Next day → | Search | Index