Time Nick Message 00:10 wizzyrea unfair weather. 00:10 wizzyrea first day back to work with brilliant sunshine after two wet days, one cold. 00:11 wizzyrea three wet, one cold, I mean. 00:11 rangi yeah 00:24 ibeardslee go home, 'sick .. of work' 00:25 * Francesca waves 00:25 rangi hi Francesca 00:25 Francesca sup 00:26 wizzyrea ibeardslee: yeah nah, that can't happen 04:27 * dcook waves 04:28 dcook Don't suppose rangi or gmcharlt are around at the moment regarding a Debian package question? 04:28 dcook Or does anyone recall if we're still supporting Debian Squeeze? 04:29 dcook I suppose http://wiki.koha-community.org/wiki/Koha_on_Debian doesn't specify Squeeze anymore.. 07:00 Joubu hi 07:32 * magnuse waves 07:39 * Francesca waves 07:43 Joubu @later rangi there is no way to disable PR on github 07:43 huginn Joubu: I've exhausted my database of quotes 07:43 Francesca lol 07:45 eythian dcook: wiki page history will tell you when it was removed 🙂 07:46 Joubu @later tell rangi there is no way to disable PR on github 07:46 huginn Joubu: The operation succeeded. 07:46 Joubu @later tell rangi I have updated the description of the project with "Note: This project uses its own bug tracker, see bugs.http://koha-comminity.org to report a bug or submit a patch. " Let me know if you prefer a better wording 07:46 huginn Joubu: The operation succeeded. 07:46 reiveune hello 07:47 Francesca hi 07:47 wahanui privet, Francesca 07:49 magnuse Joubu++ 07:57 alex_a bonjour 07:57 wahanui hola, alex_a 08:00 magnuse @wunder boo 08:00 huginn magnuse: The current temperature in Bodo Vi, Norway is -6.0°C (4:00 AM CET on January 05, 2016). Conditions: Partly Cloudy. Humidity: 33%. Dew Point: -16.0°C. Pressure: 29.99 in 1016 hPa (Falling). 08:02 rangi Joubu++ 08:02 rangi thanks 08:02 rangi at some point i should grab those PR and make them into patches 08:04 Joubu At least 1 is important (it fixes a syntax issue in sql file), the last one 08:04 Joubu I have closed the PR, asking to open a bug report on bz 08:10 gaetan_B hello 08:10 Joubu @later tell khall How did you generate the schema? I get an error "make_schema_at(): Checksum mismatch in './/Koha/Schema/Result/Borrower.pm', the auto-generated part of the file has been modified outside of this loader" 08:10 huginn Joubu: The operation succeeded. 08:12 Joubu @later tell khall certainly comes from 017f62ea3752a459a1f5cafecae85e9fb5bfbdd1 08:12 huginn Joubu: The operation succeeded. 08:17 magnuse kia ora slef! 08:22 mveron Good morning / daytime #koha 08:22 * mveron waves to rangi - thanks for your blog! 08:22 magnuse hiya mveron 08:23 * cait waves 08:23 rangi no problem, thanks for all your hard work 08:23 cait rangi++ mveron++ :) 08:23 mveron hiya magnuse, cait, Joubu everybody... 08:24 Joubu hi mveron 08:24 cait and a Gutes Neues! 08:24 mveron cait: Äbefalls äs guets Nöis 08:24 Joubu mveron: what is your solution on bug 15462N 08:24 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=15462 critical, P5 - low, ---, jonathan.druart, Needs Signoff , Unable to renew books via circ/renew.pl 08:24 Joubu ? 08:24 Joubu This is a tricky one 08:25 mveron Joubu: Did you try my patch? 08:25 Joubu I haven't seen it yet 08:25 Joubu well yes 08:25 Joubu that was the naive approach 08:25 Joubu and it could work 08:25 Joubu but I was not sure enough yet 08:25 Joubu did you catch the reason (the unique key on itemnumber)? 08:27 mveron Joubu: If my patch is wrong or to naive feel free to obsolete it. 08:29 Joubu item->issues has been (automatically) renamed with item->issue after we have added the unique key constraint on item.itemnumber 08:29 Joubu I am not sure to understand why but when this unique key has been added, the has_many rs has been replaced with a might_have 08:30 Joubu and the plural of issue is gone 08:30 Joubu so that would mean we can only access to 1 issue for a given item 08:30 drojf morning #koha 08:30 Joubu which sounds odd 08:31 Joubu hi drojf :) 08:31 mveron Hi drojf :-) 08:31 cait not sure i can follow the 2 of you :) 08:31 cait i think 1:1 between items and issues seems correct, 1:n for old_issues? 08:32 drojf jo Joubu, mveron and cait :) 08:32 mveron cait: I fixed soething thet seems to be a symptom of some bigger issue. 08:32 mveron smoething 08:32 drojf huh, i meant hi. but jo works too 08:32 drojf :D 08:32 mveron something (should put my glasses) 08:32 Joubu yes it only refers issues of course 08:32 drojf @wunder berlin, germany 08:32 huginn drojf: The current temperature in Berlin Tegel, Germany is -9.0°C (9:20 AM CET on January 05, 2016). Conditions: Light Snow. Humidity: 79%. Dew Point: -12.0°C. Windchill: -15.0°C. Pressure: 29.53 in 1000 hPa (Rising). 08:32 drojf eeek 08:32 mveron @wunder Allschwil 08:32 huginn mveron: The current temperature in Grenchen, Switzerland is 4.0°C (9:20 AM CET on January 05, 2016). Conditions: Light Rain Showers. Humidity: 100%. Dew Point: 4.0°C. Pressure: 29.44 in 997 hPa (Rising). 08:32 Joubu there is no fk with old_issues 08:32 cait @wunder Konstanz 08:32 huginn cait: The current temperature in Saint Gallen-Altenrhein, Germany is 5.0°C (9:20 AM CET on January 05, 2016). Conditions: Light Rain Showers. Humidity: 87%. Dew Point: 3.0°C. Windchill: 4.0°C. Pressure: 29.47 in 998 hPa (Rising). 08:33 magnuse @wunder boo 08:33 huginn magnuse: The current temperature in Bodo, Norway is -6.0°C (9:20 AM CET on January 05, 2016). Conditions: Clear. Humidity: 42%. Dew Point: -17.0°C. Windchill: -15.0°C. Pressure: 29.98 in 1015 hPa (Steady). 08:33 liw @wunder Espoo, Finland 08:33 huginn liw: The current temperature in Viherlaakso, Espoo, Finland is -19.8°C (10:29 AM EET on January 05, 2016). Conditions: . Humidity: 90%. Dew Point: -21.0°C. Windchill: -20.0°C. Pressure: 29.68 in 1005.0 hPa (Rising). 08:33 cait it picks the funniest places... all but Konstanz 08:33 * Joubu is trying something 08:33 cait no rain here, but cloudy and rather warm for winter 08:33 mveron brb 08:33 magnuse liw wins 08:41 rangi @later tell tcohen can you give me ssh access to the debian_7 jenkins node, I 08:41 huginn rangi: The operation succeeded. 08:42 rangi @later tell tcohen I'd like to see if I can figure out why some of the tests are failing 08:42 huginn rangi: The operation succeeded. 08:45 rangi @later tell tcohen or can you install URL::Encode that will fix 2 08:45 huginn rangi: The operation succeeded. 08:50 ashimema morning 08:51 lds hello 08:51 cait morning ashimema and lds 08:51 sophie_m hello cait and #koha 08:51 mveron hi ashiema and lds and sophie_m 08:57 LibraryClaire morning #koha :) 09:01 mveron morning LibraryClaire 09:02 mveron Joubu: Could your findings on http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=15462#c4 have impact at other places? 09:02 huginn 04Bug 15462: critical, P5 - low, ---, jonathan.druart, Signed Off , Unable to renew books via circ/renew.pl 09:02 ashimema morning LibraryClaire.. and all others I missed since last saying good morning ;) 09:02 * ashimema needs tea 09:03 * mveron fetches some coffee... 09:04 LibraryClaire hi ashimema. mveron :) 09:05 mveron brb 09:09 Joubu mveron-away: no, seems to only impact this script 09:09 mveron OK, glad to hear :-) 09:09 mveron Joubu++ 09:10 drojf1 mveron: does fetching coffee work with git? 09:11 mveron drojf1: I did not try yet... :-) 09:12 ashimema Morning Joubu 09:16 Joubu hi ashimema 09:16 cait git fetch coffee? :) 09:19 ashimema 'git worktree coffee_break' 09:19 ashimema :) 09:19 drojf1 :) 09:20 drojf i should fetch the irc password from the old notebook … 09:22 ashimema lol 09:26 * mveron thinks about a coffee enhancement for Koha, maybe cron job driven (every 30 minutes a coffee?) 09:26 cait hm maybe we need a sensor too 09:27 drojf in koha-create, we have a lot of sed -e "s/__PLACEHOLDER__/$replacement/g" … but for UPLOAD_PATH it is sed -e "s#_UPLOAD_PATH__#UPLOAD_PATH#g" … is the # some magical sed variant or an error? 09:27 cait to check caffeine levels 09:27 cait so the coffee flow can be customized :) 09:27 liw drojf, sed allows any character to be used as delimiter, / just the usual one, but since a path is likely to contain / that expression uses # instead 09:27 liw drojf, you can even use a space charaqcter, I think 09:28 drojf ah, yes. now that you say that, i remeber i have learned that before. and forgot. i'm old :/ 09:28 drojf liw++ 09:29 drojf i probably asked that for the exact same line when i tested the patch… 09:29 * drojf hides in shame 09:29 liw the good thing about getting old and forgetful is that you get to re-visit all manner of wonderful things as if they were new 09:29 drojf heh 09:30 liw and also you get to show off your knowledge of the esoteric things you didn't forget, of course 09:30 drojf you mean the stuff everyone else has forgotten because it's obsolete? :) 09:31 drojf now i forgot what led me to the question in the first place. i'll try the coffee thing 09:33 Joubu @later tell khall please regenerate the Schema files, some tests are failing because of bug 13624 09:33 huginn Joubu: The operation succeeded. 09:39 nlegrand hey #koha 09:45 sophie_m hi rangi, Would i be possible to add v3.18.13 on koha ? It seems that Liz is on holiday and I need it to upgrade some client. Or can someone else do that ? 09:46 sophie_m add the tag I mean 10:01 Joubu Does anyone understand this test? 10:01 Joubu 153 warning_is { BuildSummary($marc21_subdiv, 99999, 'GEN_SUBDIV') } [], 10:01 Joubu 154 'BuildSummary does not generate warning if main heading subfield not present'; 10:02 Joubu gmcharlt maybe? 10:04 nlegrand I'm looking at translation and I'm a bit lost, how the .po files are generated? 10:06 magnuse nlegrand: by misc/translator/translate 10:08 nlegrand magnuse: ha! thanks, now I know where to start ^^ 10:08 drojf @later tell marcelr would you agree that having the upload folder in a separate backup file would be a good idea? right now (i think) it ends up in the configs.tar.gz when it is in the standard path, or nowhere if it is somewhere else (package installation) 10:08 huginn drojf: The operation succeeded. 10:38 nlegrand I've been messing a bit with holds lately, for our personnal use, not sure it will be usefull for others, and I didn't want to interfer with bug 5609 10:38 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=5609 enhancement, P1 - high, ---, koha-bugs, NEW , Holds Rewrite 10:39 nlegrand I just wanted to know, what's the status of this bug? Is it currently worked on? I guess I will put my nose on it when something comes up since we will soon use holds to asks for books in stacks :) 10:42 cait I think it's an old omnibus bug 10:42 magnuse and all the bugs it depends on seem to be fixed 10:42 cait the depends on seem all resolved by now 10:43 cait I thik it would probably be ok to close it even and then have a new one if someone wants to tackle a bigger holds rewrite again 10:43 magnuse someone from bywater should say if there is more that will be done, perhaps? 10:44 cait @later tell khall is 5609 currently being worked on or could it be closed? 10:44 huginn cait: The operation succeeded. 10:44 magnuse or just close it and say "looks like all the pieces have been done, please reopen if that is not the case" 10:44 nlegrand ha right! 10:46 nlegrand pff I should ask things more often, you answer everything nicely :) 10:46 nlegrand thanks ^^ 11:18 drojf is there no direct link to upload.pl in tools? how are people supposed to set files nonpublic? i don't see that option if i use the upload.pl plugin from cataloguing 12:17 drojf for the record, it's in bug 14686 12:17 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=14686 enhancement, P5 - low, ---, m.de.rooy, NEW , New menu option and permission for file uploading 12:19 tcohen mornin 12:19 tcohen g 12:21 cait hi tcohen and a happy new year :) 12:24 tcohen hi cait! 12:24 tcohen @later tell rangi you already have access to the server. Remember you need to jump from the master jenkins server 12:24 huginn tcohen: The operation succeeded. 12:48 oleonard Hi #koha 12:49 cait morning oleonard :) 12:49 LibraryClaire hi oleonard 13:33 oleonard @wunder 45701 13:33 huginn oleonard: The current temperature in Heatherstone, Athens, Ohio is -10.4°C (8:33 AM EST on January 05, 2016). Conditions: Clear. Humidity: 63%. Dew Point: -16.0°C. Windchill: -10.0°C. Pressure: 30.63 in 1037 hPa (Steady). 13:35 oleonard I wonder if we could work github pull requests into our workflow somehow. It would be great to not have to put off people's offer of help. 13:35 cait true 13:36 cait but not sure how 13:36 oleonard I don't even know what one does with a pull request, so I'm no help there 13:40 Joubu nothing directly I suppose, or you can write a hook to open a bug report on bugs.k-c.org, then fill it and attach the patch 13:40 Joubu s/you/we :) 13:51 cait oleonard: normally you merge them into your codebase - with a click in on a button.. i didn't really get further than that so far :) 14:27 cait rainbow outside my window :) 14:55 ashimema pull requests are lovely.. 14:56 ashimema shame we won't ever use them 14:57 ashimema They basically do the 'report bug -> fix bug -> submit to community -> feedback on submission -> update submission -> merge to community' in one nice streamlined interface 14:57 * ashimema uses github for pretty much every project except koha 14:59 ashimema I did add bug 15465 this morning in relation to github as it happens though 14:59 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=15465 enhancement, P5 - low, ---, nengard, NEW , README for github 14:59 ashimema we are putting people off with our workflow me thinks 14:59 Joubu Is there a way to have both? 14:59 Joubu github and bz 14:59 ashimema erm.. 15:00 ashimema we bascially have bugzilla 'cause we wanted to host it ourselves 15:00 Joubu (There were 6 PR in 1 year on github) 15:00 ashimema which makes sense 15:00 ashimema indeed.. there aren't many cuase the vast majority of us know the github is not our primary place to work 15:01 ashimema it's only entirely new to the project people that tend to submit there 15:02 ashimema hence my thoughts on having a README.md file especially for github pointing them in the right direction before they even try 15:25 cait have to run (bus) bye all 16:07 eythian @wunder ams 16:08 huginn eythian: The current temperature in Schiphol, Badhoevedorp, Netherlands is 7.2°C (5:07 PM CET on January 05, 2016). Conditions: Overcast. Humidity: 90%. Dew Point: 6.0°C. Windchill: 7.0°C. Pressure: 29.27 in 991 hPa (Steady). 16:12 drojf @wunder berlin, germany 16:12 huginn drojf: The current temperature in Holzdorf, Germany is -7.0°C (5:00 PM CET on January 05, 2016). Conditions: Light Snow Grains. Humidity: 86%. Dew Point: -9.0°C. Windchill: -12.0°C. Pressure: 29.53 in 1000 hPa (Rising). 16:12 drojf getiing warmer ;) 16:13 reiveune bye 16:53 gaetan_B bye 17:55 mveron bye 18:30 Nemo_bis Hi 18:46 rangi morning 18:46 rangi @later tell tcohen thanks 18:46 huginn rangi: The operation succeeded. 20:26 rangi @later tell tcohen i don't suppose you could gpg encrypt and email the passphrase for the jenkins ssh key, I have totally forgotten/mislaid it 20:26 huginn rangi: The operation succeeded. 20:40 rangi @later tell fredericd I have change cf_release_notes to be a textarea now, on bugzilla 20:40 huginn rangi: The operation succeeded. 21:20 cait hi #koha :) 21:20 wizzyrea hi cait 21:21 wizzyrea https://github.com/mozilla/autolander 21:21 wizzyrea ^ could be useful 21:21 wizzyrea for github/bugzilla 21:22 rangi hmm 21:23 rangi that might encourage people to use github, which busts all our signoff/qa workflow 21:23 rangi if you could switch a repo on github readonly 21:24 rangi and just use it as a backup 21:24 rangi or just get rid of it ... 21:24 wizzyrea yeah the bad thing is that people kind of know how to use github, or it's an easy access point 21:25 wizzyrea your points are valid though, it would break our workflow as it exists now 21:25 rangi if someone wanted to take over it, and submit their pull requests as patches 21:25 wizzyrea that's kind of what I was thinking this tool might do 21:25 rangi seems mostly the other way 21:26 rangi ie pulls from bugzilla 21:27 rangi but if we could disable that bit, and just made pull requests create bugs/attachments 21:27 rangi that might be useful 21:27 wizzyrea yeah, that's the only part I was interested in 21:28 wizzyrea if it could create bugs from incoming pull requests and immediately close them 21:28 wizzyrea that would be ace 21:28 rangi yup 21:29 rangi actually i could script something like that 21:29 rangi it just needs to be run as a cron, fetch the pull requests from github 21:29 wizzyrea I saw a thing yesterday that could automatically close pull requests 21:29 rangi make bugs 21:29 rangi theres cpan libs for most of that 21:29 rangi ill have a look at the weekend 21:30 wizzyrea hm, it'd have to check to see if there was an existing bug 21:30 rangi yep 21:30 rangi it probably cant do that 21:30 rangi not in any reliable way 21:30 rangi that pretty much takes a human 21:31 rangi because english sux 21:31 wizzyrea :) 21:31 wizzyrea it might be better if we started omitting the "bug" part of the commit messages 21:31 rangi hmm? 21:31 wizzyrea so just "xxxx - description of my very annoying bug" 21:32 wizzyrea instead of "bug xxxx - description of my very annoying bug" 21:32 rangi just the number? 21:32 wizzyrea yeah 21:32 rangi how would that help? 21:32 wizzyrea dunno, fewer patterns to worry about? 21:32 rangi if there's a number in the string thats easy to parse 21:33 rangi its more that i highly doubt there will ever be a bug number 21:33 rangi in the commit 21:33 rangi just a string of text 21:33 wizzyrea oh from github 21:33 rangi so we'd pretty much have to make a new bug for every pull request 21:33 wizzyrea yeah, I was just thinking about that 21:33 rangi but if we tag them 21:33 wizzyrea bit of chicken and egg there 21:34 wizzyrea maybe when you close add the bug number to the close message 21:34 wizzyrea so there's a record 21:35 rangi yep 21:35 rangi also if we mark them as created from github in bugs 21:35 rangi then a human could eyeball and resolve duplicate etc 21:37 wizzyrea https://github.com/gera/gitzilla there's this one too 21:39 rangi ta 21:40 geek_cl hi guys, what about this circulation.pl time run : top pastebin _ http://pastebin.com/fsuPFSfM 21:40 geek_cl 3 hours and counting 21:45 geek_cl that is normal ? 21:45 rangi of course not :) 21:45 rangi whats the load on that machine? 21:45 geek_cl on the roof 21:45 rangi theres your problem then 21:45 geek_cl all cpu (8) up to 100% 21:45 rangi you might want to wind down the max number of connections that apache is letting through 21:46 rangi also, is it swapping? 21:46 geek_cl swapping a little 21:46 geek_cl 3 MiB 21:47 rangi yeah, it just looks like that machine is trying to do too much, and more and more is piling up 21:47 rangi is the db on the same machine? 21:47 geek_cl nop 21:48 rangi i'd check it as well 21:48 geek_cl the db server is a remote machine 21:48 rangi also the connection between the two 21:48 geek_cl i already check the db server and is OK 21:48 rangi whats your max connection for apache? 21:48 geek_cl let me check 21:49 rangi but yeah 2-4 seconds 21:49 rangi is what circulation.pl should take 21:49 rangi if it's taking longer than that, something is going wrong 21:49 rangi what does show processlist 21:49 dcook Wasn't there a bug at one stage where circulation was taking a long time? 21:49 rangi on the mysql server tell you 21:50 rangi not that long 21:50 wizzyrea yeah not that long 21:50 dcook Ahh 21:50 * dcook retreats back into the shadows 21:50 rangi not hours ;) 21:50 geek_cl rangi, http://pastebin.com/T1WNz2dH 21:51 geek_cl i will check processlist 21:51 rangi dcook: it does take longer than it should, theres quite a few wins to be made, specially if more of it is done using the rest 21:51 dcook rangi: Ah, I was reading that as seconds rather than minutes. Yikes. 21:51 geek_cl rangi, too many Sleep process 21:52 rangi i think what i would do is restart apache 21:52 geek_cl 7 , like circulation 21:52 geek_cl ok 21:52 bag is it all from the same IP? 21:52 rangi and then watch the log 21:52 geek_cl rangi, what about max conns of apache 21:52 rangi you have 8 cores eh? and how much ram? 21:53 geek_cl RAM 8 GiB 21:53 rangi yeah, if you aren't OOMing .. you can probably leave that 21:53 geek_cl after apache2 restart, the CPU's breath again 21:53 rangi something else is hanging those circs 21:54 rangi yeah id tail the access log for a while .. and see if you can spot where it comes from if it happens again 21:54 bag find the computer that the traffic is coming from and clear the cache on that computer 21:54 bag (that’s worked for us) 21:54 * geek_cl nice 21:54 geek_cl let me check 21:57 geek_cl i find a internal (customer) IT guy, hitting circulation.pl 22:00 geek_cl ohh. no no, is a IP of the gateway, koha is under a NAT 22:01 nuentoter hello everyone 22:02 wizzyrea hi 22:05 nuentoter have a question, finally was able to get things imported into koha from our old system. If i go under tools>inventory I see my records and if i click them they all have items 22:05 nuentoter i cannot search anything 22:05 rangi how did you install koha? 22:05 nuentoter figured zebra needed a restart 22:05 nuentoter what do you mean how did i install koha? 22:06 rangi did you install it from packages 22:06 rangi or from the tarball 22:06 nuentoter im running it on a virtualbox using debian7 installed through packages 22:06 rangi right 22:06 rangi so the thing you want to make sure you are doing is using the package commands, not by hand 22:07 rangi so to restart zebrasrv it would be 22:07 rangi sudo service koha-common restart 22:07 rangi and to rebuild your index 22:07 rangi sudo koha-rebuild-zebra -v -f <instance name goes here> 22:07 rangi eg 22:07 rangi sudo koha-rebuild-zebra -v -f mylibrary 22:08 geek_cl thanks rangi ;) 22:08 rangi a lot of problems are cause by people running commands as root, like rebuild_zebra.pl .. and messing up the permissions, and then the cronjobs cant run 22:08 nuentoter thats what i did 22:08 nuentoter oops 22:09 eythian wahanui: zebra troubleshooting 22:09 wahanui zebra troubleshooting is see [understanding zebra indexing] and [yaz client] 22:09 rangi yeah, so now the zebrasrv cant actaully read those indexes 22:09 rangi because root owns then, and it runs as koha-instancename 22:10 rangi so you will want to fix those permissions 22:10 rangi if you do a ls -l /var/lib/koha 22:10 rangi what is in there? 22:10 wahanui i heard in there was a css dir 22:10 nuentoter i was logged in to root for an unrelated reason, while still logged in as root i sent the rebuild_zebra.pl command 22:11 rangi yeah, dont ever do that :-) 22:11 rangi koha-rebuild-zebra is what you want to use, but we can fix it, if you fix the permissions 22:12 rangi in /var/lib/koha there should be a dir owned by instance-koha:instance-koha 22:12 rangi eg 22:12 rangi drwxr-xr-x 6 catalyst-koha catalyst-koha 4096 Dec 14 12:09 catalyst 22:12 nuentoter total 4 drwxr-xr-x 5 abelj-koha abelj-koha 4096 Nov 5 20:34 abelj 22:12 rangi cool, and if you ls -l inside there? 22:13 rangi its probably easiest just do 22:14 rangi sudo chown -R abelj-koha:abelj-koha /var/lib/koha/abelj 22:14 rangi to make sure 22:14 rangi then sudo service koha-common restart 22:14 rangi to restart the zebrasrv 22:14 rangi and see if that has got your searching back 22:14 rangi you might want to do a 22:15 rangi sudo koha-rebuild-zebra -f -v abelj 22:15 rangi as well, just for good measure 22:20 nuentoter yay biblio export runnin looks good so far 22:21 nuentoter and search works again 22:21 nuentoter TY rangi!!! 22:21 wizzyrea yay! 22:22 cait rangi++ :) 22:22 rangi no worries 22:22 rangi it's almost always permissions 22:22 nuentoter hahaha you dont know me, i always worry lol 22:23 cait German? :P 22:24 rangi now you've learnt some nz slang as well as fixing your koha :) 22:24 nuentoter french/irish 22:24 cait :) 22:24 cait maybe it's more wide spread in europe :) 22:25 cait worrying too much that is 22:26 nuentoter now next step for me is to get all my record imported........ ugghhhhh 22:26 nuentoter doing batches of 25 books at a time, 35,673 books left to go! YAY 22:26 nuentoter i love my library lolol 22:27 wizzyrea why 25 at a time? 22:27 nuentoter because of the way our current system exports things I have to hand edit every single book. so i do it in batches to prevent sscrew ups 22:28 wizzyrea oh that is rubbish :( 22:28 nuentoter winnebago/spectrum is a pain in my @$$ 22:28 wizzyrea hm 22:28 wizzyrea I know someone who may have some tips for you 22:28 wizzyrea re: winnebago 22:28 nuentoter !!!! i will take any info they have 22:28 wizzyrea see your PM's in 2 secs 22:29 nuentoter most people that converted did so YEARS ago, i inherited this old system :( 22:29 wizzyrea *nod* it's crap 22:37 dcook I have some design questions if people have a moment to offer thoughts 22:37 rangi im going to get nachos 22:37 dcook That's fair enough 22:37 rangi but ill read back 22:38 wizzyrea yeah will read them when I have a minute, go ahead and ask :) 22:38 dcook For this OAI-PMH stuff, I have a daemon that downloads records and writes them to disk. That all works quite well. 22:38 dcook But now I'm thinking about a daemon as an importer as well 22:38 dcook Something that monitors the disk periodically and does X with them 22:38 dcook But... I'd like that daemon to have access to Koha modules 22:39 dcook So while the OAI-PMH downloader (it's actually extensible so it could pretty much do anything) is loosely coupled 22:39 dcook I want the importer to be fairly tightly coupled 22:39 dcook I could just have a cronjob run instead of a daemon, but a daemon is going to be a lot less problematic 22:40 dcook So I could just start up the daemon with PERL5LIB set... 22:40 dcook or use FindBin 22:40 wizzyrea how does the zebra one do it? 22:40 dcook That's a good point. We only use the cronjob... I hadn't thought about that 22:40 dcook Tamil's daemon reads the database using koha-conf.xml 22:41 dcook I don't know how the community one works, but that's a good point 22:41 dcook Actually, the zebra one doesn't need any modules I think? 22:41 dcook Just DB access 22:41 * wizzyrea admits to not knowing 22:41 dcook Hmm rebuild_zebra.pl does use a few C4 modules 22:41 dcook wizzyrea: That's really helpful 22:42 dcook Because if there's a precedent... I'll use that 22:42 dcook I find the whole module path thing to be annoying in all projects.. 22:42 dcook Ideally, it would be kind of nice to just be able to use /usr/local or something... 22:42 dcook Although then our gits wouldn't work.. 22:43 dcook Although I guess for git-dev stuff we could use PERL5LIB.. 22:43 * dcook wonders how all projects ever do this 22:43 nuentoter i dont think zebra actually accesses the DB 22:43 wizzyrea it about has to, to get the records to index 22:43 dcook Well, not Zebra itself 22:43 wizzyrea the indexer does though 22:43 dcook rebuild_zebra.pl writes records out to a directory that Zebra manages 22:44 dcook wizzyrea: what do you mean by indexer? 22:44 wizzyrea rebuild_zebra 22:44 wahanui it has been said that rebuild_zebra is already there but only every 10 mins.... 22:44 dcook Ah yeah 22:44 wizzyrea zebra reads it's own databases 22:44 dcook Yeah, rebuild_zebra.pl uses C4 modules to do stuff 22:44 wizzyrea but the link is rebuild_zebra 22:44 dcook Well, Zebra reads from disk to build its databases 22:44 dcook rebuild_zebra writes to disk so that Zebra can read from it 22:45 dcook With my project, my downloader writes to disk and the importer reads from disk 22:45 dcook So sort of backwards to what we're doing with Zebra 22:45 wizzyrea note I"m using 'database' in a loose form 22:45 wizzyrea not strictly in the relational-database form 22:45 dcook Good point 22:45 wahanui I know! The blade went right through that child! 22:46 dcook We sure have some odd ones in #koha :p 22:46 wizzyrea hehe 22:46 dcook Lesse... I should just fire up my Debian VM 22:46 dcook My life would be so much easier if we just used Debian for dev and prod 22:47 dcook I did take a few lessons from Zebra when doing this 22:47 nuentoter what do you use if not debian? 22:47 dcook I wrote a little "icarus-client" in a sort of shout out to "yaz-client" 22:47 dcook openSUSE 22:47 wahanui openSUSE is, like, not used by many developers, and will likely be difficult to get Koha working with 22:47 dcook hehe 22:47 wizzyrea it's like you like punishment 22:48 dcook Ikr? 22:48 dcook Slowly effecting change around here. That would be something that would make me really happy 22:48 dcook O_O 22:49 dcook These bywater folks are blowing my mind right now 22:49 wizzyrea oh? 22:49 nuentoter I've used linux for quite a few years personally, but never did a whole lot with it tbh, it was just a free os to surf the web with cuz windows sucks 22:49 bag happy birthday thatcher 22:49 dcook barton and thatcher are the same person? :p 22:49 wizzyrea lol no they are not 22:49 dcook nuentoter: Yeah, I started using Linux for work, and now I use it personally too 22:49 barton no... 22:49 dcook And I've used... a fair few distros 22:49 * dcook still likes Debian best 22:50 wizzyrea I like debian and it's derivatives best. 22:50 wizzyrea everyone else wants to be like them :P 22:50 dcook hehe 22:50 dcook Yeah, we're a Debian/Ubuntu house 22:50 dcook And work is all openSUSE/SUSE 22:50 nuentoter I have used a handful and on my personal laptop i use #! only because its slick and quick with no bells and whistles 22:50 dcook Except for one Ubuntu server I have.. 22:50 nuentoter i use debian on desktop at home 22:50 dcook nuentoter: Yeah, I thought about #! for my old netbook 22:51 dcook Went with lubuntu in the end though 22:51 dcook Then it got stolen so it didn't matter much 22:51 dcook wizzyrea: koha-indexer has /usr/share/koha/lib hard-coded into PERL5LIB 22:51 dcook And not in the best of ways.. 22:51 nuentoter for older hardware #! is nice, not the new #!++ which is debian 8 based 22:52 wizzyrea \o/ 22:52 * dcook ponders 22:52 dcook I'll take that as incentive to just require users to have PERL5LIB set ahead of time I guess.. 22:52 dcook Or maybe use FindBin 22:52 nuentoter the whole openbox setup is what hooked me, no icons, no menus just simple 22:52 wizzyrea yeah, that's a precedent alright. rangi might have opinions. 22:52 dcook wizzyrea: It's a Debian-specific precedent, but yeah, I'd be curious to hear what rangi would say 22:53 dcook What is INC most of the time.. 22:53 dcook I was just looking yesterday 22:53 dcook I'm cheating with something.. 22:53 dcook Oh, I do have another one.. 22:53 wizzyrea you could create a startup option where you use hardcode unless you specify something different 22:53 dcook I have stuck some modules into "bin" :O 22:53 dcook wizzyrea: Yeah, that's true 22:54 wizzyrea i mean, there's probably a reason not to do that. 22:54 nuentoter anyway, gotta close up the library and go home for the evening, thank you kind folks! 22:54 dcook laters nuentoter 22:54 wizzyrea good luck nuentoter :) 22:54 dcook wizzyrea: Yeah, I'm trying to think of the ideal way of doing things 22:55 dcook In the case of the downloader, it could be completely separate from Koha 22:55 dcook So in theory... the downloader could be replaced down the line with something else 22:55 dcook And the importer could handle other data providers 22:56 dcook In theory, I'd like the importer to be given a bunch of data and go to town on it... and then you could check on the progress later 22:56 dcook So you upload a MARC file, tell it to upload, and then go off and do something else without worrying about any timeouts or anyhting 22:56 dcook I mean... we already do have BackgroundJob which works in the background, although I haven't studied that extensively 22:56 dcook Anyway, just babbling now 22:57 dcook I don't really like the idea of a closely-coupled daemon... but surely other projects must do it too? 22:57 dcook So that you can exploit your project's existing code.. 22:57 dcook I get the whole "have your daemon do one thing" but.. 22:58 dcook I suppose that would be more possible if Koha were smaller 22:58 dcook Anyway, thanks for that wizzyrea :) 22:58 wizzyrea yep, rangi will probably have better opinions 22:59 dcook Hmm, maybe I'll wait for his opinion before I start on that 23:02 dcook Actually, now that I think about it, another idea I had wouldn't need that.. 23:02 dcook Rather, the "importer" daemon would download "import profiles" from a Koha web service 23:02 dcook Each profile would have it monitoring a different place (e.g. a directory) 23:03 dcook Then the daemon would simply send the record off asynchronously to a Koha import web service 23:03 dcook I need a more complex import web service than any we currently have though 23:03 dcook Something that filters a record, performs matching, and then does X 23:03 dcook For that... the only thing the daemon would need is username/password and a web service URI 23:04 dcook The downloader daemon uses a "task format" while this import daemon would use a "import profile format" 23:04 dcook That would be the only sort of dependency 23:04 dcook That whatever task provider or import profile provider use that format that the daemon understands 23:05 dcook The import profiles could in theory also be moved into other parts of Koha... 23:06 dcook So that all imports could perform record filtering, matching, and X 23:06 dcook X being particular to that "kind" of import 23:06 dcook X probably just being "authority","bibliographic","items" (and maybe "holdings") 23:07 dcook And the daemon only needs to read from that "import profile" web service on start or reload 23:08 dcook I suppose that web service would actually sort of be a "discovery" document 23:09 dcook Well, no, not quite.. 23:09 dcook Because the daemon would need to know more than just "Oh hey... these are the import endpoints we support" 23:12 pianohacker we need to add a feature to huginn that tracks the longest streak of someone talking to themselves, dcook ;) 23:13 dcook hehe 23:13 dcook I'm pretty sure I'd occupy all the top spots ;) 23:13 pianohacker also, hi! :) 23:13 dcook But I'm actually only talking to myself and whoever wants to listen/comment :p 23:13 dcook also hi :) 23:13 dcook New year going well so far? 23:26 pianohacker dcook: absolutely. Looks to be much easier than the last :) 23:26 rangi dcook: do you know much about gearman 23:27 rangi http://gearman.org/ 23:27 rangi http://search.cpan.org/~dormando/Gearman-1.12/lib/Gearman/Client.pm 23:27 rangi http://search.cpan.org/~dormando/Gearman-1.12/lib/Gearman/Worker.pm 23:27 rangi etc 23:27 dcook pianohacker: Awesome 23:27 dcook rangi: Nopes 23:28 rangi i think instead of reinventing the queue/work distribution wheel, we could slowly move bits of koha out and have gearman looking after who does what and when 23:29 dcook I like this idea 23:29 rangi file comes in for import, farm it off to the import worker 23:29 rangi keep on trucking along 23:29 dcook Well, the Gearman model is what I'm doing with the downloader already 23:29 dcook Web UI sends a job to the job server, job server spawns a process to deal with it 23:29 rangi yep 23:30 dcook child process queues up the record for somethign else down the pipeline 23:30 rangi gearman lets you create workers in whatever language you like etc too 23:30 dcook Yeah, I'd be down with Gearman at a glance 23:30 rangi handles the reporting back 23:30 rangi all that stuff thats a pita to deal with 23:30 dcook Yeah, the worker in the downloader case is a OAIPMH module I made, so that would be great 23:31 dcook Yeah, the more I read, the more I like this idea 23:32 dcook I have a project at home that this would be great for as well.. 23:32 rangi the thing I like is its a tried and tested thing, so one less bit of code we have to worry about 23:32 dcook ^ 23:32 dcook Exactly 23:32 dcook So much that 23:33 dcook In the case of the importer, I'm not 100% sure what I want to do yet 23:33 dcook I don't know if I want an importer reading in files and processing them.. 23:33 dcook Or if I want something reading in files and sending them to a Koha API 23:33 dcook Handing them off to a Koha API would probably be faster and more generalizable.. 23:34 dcook But then the transfer agent thing needs to know extra details to tell the API.. 23:34 dcook Mainly (how do you want to filter, match, and ultimately add/update/ignore this) 23:34 dcook While we could put that in a configuration file, it would be nice to have it in the database so that end users could make those decisions 23:35 dcook But then there's needs to be a link between the file being read and all that import info.. 23:35 rangi yeah but the worker doesnt need to know its in the db 23:35 dcook True. It just needs to know it in some way. 23:35 rangi the client hands it off to the jobserver the jobserver passed that info to the worker, with the data 23:36 dcook Hmm 23:36 rangi heres some data, and heres the config 23:36 dcook That's what I'm doing with the downloader 23:36 pianohacker dcook/rangi: I've been wanting to make the cronjobs managed (configurable from web, etc.), and this sounds like it could be useful for that as well 23:36 dcook I wasn't sure if I wanted to go that route with the importer 23:36 dcook pianohacker: Yeah, I've been thinking about that too 23:36 rangi pianohacker: yup 23:36 rangi the task scheduler would be the first thing to do 23:36 dcook ^ 23:36 pianohacker it's on the todo list, somewhere down there :) 23:36 rangi its small, self contained, known problem 23:37 pianohacker and doesn't currently work :) 23:37 dcook The only thing is I have a deadline for this one :p 23:37 rangi rewrite to use gearman instead of at 23:37 dcook The downloader is a task scheduler at this point 23:37 dcook Gearman would be great 23:37 dcook So here's the flow I have so far 23:37 dcook Client tells task/job server "Here's a task" 23:38 dcook The task/job server spawns a worker to deal with that task 23:38 dcook In this scenario, it's downloading records via OAI-PMH 23:38 dcook If it's a repeating task, it loops forever following a certain interval 23:38 dcook It writes those records out to disk atm 23:38 dcook I suppose I could use that same task/job server to handle the importing.. 23:39 dcook And then when time isn't of the essence, we replace my task/job server with gearman 23:39 dcook The OAI-PMH worker is specific to a module already so that could be pretty much drag and drop.. 23:39 * dcook is tempted to post his task/job server code separate to the OAI code... 23:40 dcook rangi: Are there packages for gearman? 23:40 rangi yep 23:40 rangi https://packages.debian.org/search?keywords=gearman 23:41 dcook Oh wow... it goes back a ways 23:41 dcook Hmm pre version 1 23:41 dcook rangi: Have you used it yet? 23:41 rangi yeah, its been around for a while 23:41 rangi yep 23:41 rangi tumblr and yelp others use it 23:41 dcook Yeah, I was seeing that 23:41 dcook I was just wondering which version they're on 23:41 pianohacker is it a swagger2 situation where we'd really want latest? 23:42 dcook Yeah, that's what I'm worried about 23:42 rangi id go for stable 23:42 rangi over shiny and new 23:42 rangi specially in a job queue 23:42 dcook Totes 23:42 dcook I'm just wondering what is stable 23:43 dcook Hmm no news in a few years 23:43 dcook That could be a good sign I suppose 23:43 dcook It should be fairly rock solid as it's a fairly simple concept 23:43 rangi yep 23:44 dcook Hmm, no packages for openSUSE. Boo.. 23:44 dcook Although someone apparently has already done some legwork: https://gist.github.com/CauanCabral/5967374 23:45 rangi i reckon it's worth trying out anyway 23:45 dcook Agreed 23:46 dcook For now, I'll probably just keep chugging with what I have, as I'm running out of time, but I like it 23:47 dcook But that makes me wonder if I should use my existing task server for both downloading and importing.. 23:47 dcook Actually, how would that importing work? 23:47 dcook You can pass data to the task server 23:47 dcook But you couldn't necessarily take advantage of your Koha Perl modules 23:47 rangi you hand of jobs to a worker suited to do them 23:47 dcook I suppose you could tell it to send the queued records to a Koha API 23:48 rangi obv a worker to do the import, should be able to talk to the db 23:48 dcook Hmm, I'll have to look at those Perl modules to see how they implement that 23:48 rangi either directly, or via some api 23:49 rangi /svc/bib/new 23:49 rangi eg 23:49 dcook Yeah, atm I'm thinking api 23:50 dcook Interesting... you have to spawn your own workers manually.. 23:50 dcook There has to be a way around that as it says it scales.. 23:50 dcook Unless you pre-spawn a certain number.. 23:52 dcook Since you're starting your own workers, you could provide all sorts of info.. 23:53 dcook Yeah, you have to manage your own clients and workers... but that's all right I guess. 23:53 dcook I autospawn workers and cap out at a configurable limit 23:54 dcook Probably less overhead with the Gearman way of doing it.. 23:54 dcook Also a lot more control 23:55 dcook I wonder a bit about what happens when Gearman goes down 23:55 dcook Well Gearmand 23:56 dcook Looks like it uses a relational database to manage jobs.. 23:56 dcook rangi: Would you have one Gearmand per client or just one for all Koha clients? 23:56 dcook Wait that didn't make sense 23:57 dcook One gearmand per system or one per Koha client 23:57 * dcook wonders how the Gearman client stores information about ongoing jobs.. 23:57 rangi that'd be up to you 23:58 rangi all koha needs to know is what ip to talk to 23:58 dcook Let's say you want to stop a job midway through though... does Gearmand return enough info to the client? 23:59 rangi itll return what the worker gives it 23:59 dcook Hmm 23:59 dcook I wonder how that would work..