2010-05-27T00:13:35 *** mck9 has left #evergreen 2010-05-27T02:13:39 *** brendan_bywater has quit IRC 2010-05-27T02:23:15 *** brendan_bywater has joined #evergreen 2010-05-27T02:58:20 *** dbwells_ has joined #evergreen 2010-05-27T03:01:48 *** dbwells has quit IRC 2010-05-27T03:02:00 *** dbwells_ is now known as dbwells 2010-05-27T04:19:16 *** pmplett has joined #evergreen 2010-05-27T06:13:46 *** pmplett has quit IRC 2010-05-27T07:50:26 *** mck9 has joined #evergreen 2010-05-27T08:01:18 *** rsinger has joined #evergreen 2010-05-27T08:19:32 *** sfortin has joined #evergreen 2010-05-27T08:36:24 *** Melissa has joined #evergreen 2010-05-27T08:38:21 *** Melissa has left #evergreen 2010-05-27T08:54:46 *** alxp has joined #evergreen 2010-05-27T09:16:50 *** bshum has joined #evergreen 2010-05-27T09:22:21 *** Dyrcona has joined #evergreen 2010-05-27T09:40:08 some Mozilla folks want to know if we'd be interested in using their logo (and getting a link-back from their Powered By landing page): http://www.mozilla.org/projects/powered-by.html 2010-05-27T09:41:28 the language says on "our site", so I guess evergreen-ils.org, and not necessarily the actual staff client 2010-05-27T09:46:16 phasefx: seems good at first glance - obviously we'd want a smaller version of the logo, like http://www.flock.com/ has in the lower-right hand cornder 2010-05-27T09:47:31 of course, we use lots of technology stacks, and they probably support Powered by Logos too. Could make a nascar-style page, but for Mozilla, you need permission from them to put their logo near others :) 2010-05-27T09:51:06 hmm - I see that 2010-05-27T10:07:21 *** Callender has quit IRC 2010-05-27T10:11:31 *** Callender has joined #evergreen 2010-05-27T10:15:45 can anyone refresh my memory on where i set up trancendence for a set of bibs? (id like to add a new "source" to trancend) 2010-05-27T10:19:39 *** cerpy has joined #evergreen 2010-05-27T10:20:38 *** youdonotexist has joined #evergreen 2010-05-27T10:22:33 mrpeters-isl: config.bib_source 2010-05-27T10:28:52 gmcharlt: is it possible to give a newly created marc record (from scratch) one of those sources? 2010-05-27T10:28:55 so it trancends 2010-05-27T10:29:33 mrpeters-isl: not in the UI, no... phasefx? ;) 2010-05-27T10:29:44 oooh that's going to be unpopular! 2010-05-27T10:31:57 mrpeters-isl: what are you trying to transcend? 2010-05-27T10:32:04 "state documents" 2010-05-27T10:32:10 they are online documents that have no holdings 2010-05-27T10:32:45 mrpeters-isl: but they are available to everyone in the consortium, so there's no need to use the 856 tag auri feature... 2010-05-27T10:32:49 http://evergreen.lib.in.us/opac/en-US/skin/craftsman/xml/rdetail.xml?r=17763043&d=0&hc=1&adv=ocn630565320 &rt=tcn 2010-05-27T10:32:54 mrpeters-isl: no group has cared enough about transcendent records to fund the dev of adding a source picker to the record editing/adding/whatever interfaces :( 2010-05-27T10:33:09 miker_: i hear ya. we may be coming calling! 2010-05-27T10:33:22 mrpeters-isl: do you have the metadata for them elsewhere? you could just use pymarc or something to generate marc records based on official metadata, then import those records in vandelay, which would let you pick your source. 2010-05-27T10:33:24 please! I'd love to see that get done 2010-05-27T10:33:58 jeff: a less than optimal workaround 2010-05-27T10:34:25 and of course, i could trancend them with SQL but it'd get to be a big PITA and would be better to jsut have the catalogers pick the source when they add a new record 2010-05-27T10:34:32 mrpeters-isl: hey, your way has catalogers creating marc records by hand. who are you calling "less than optimal"? ;-) 2010-05-27T10:35:16 * phasefx would love to be able to select the bib source from New MARC and from import via Z39.50 :) 2010-05-27T10:35:28 it all depends on the nature of the collection... if it's established and well-described elsewhere vs if it's constantly growing and isn't yet cataloged/described. 2010-05-27T10:40:01 My goal is to figure out how to copy the values of one MARC field to another tab in the Staff Client. (not the 00x fields though). People here would like to have Ctrl + some key to do that. Point me in the right direction please. I'm a lilttle clueless. 2010-05-27T10:42:36 *** kmlussier has joined #evergreen 2010-05-27T10:44:57 cerpy: there's work underway to make this easier; right now you're stuck copy/pasting one subfield at a time through the clipboard 2010-05-27T10:45:36 s/underway/coming up/ 2010-05-27T10:46:34 if you're an OCLC user, there's work being done for integrating Connexion 2010-05-27T10:47:33 (connexion work is separate from "improved" (aka optional alternate) marc editor, to be clear) 2010-05-27T10:48:06 integrating biblios is also an option for those who would want to fund that 2010-05-27T10:48:14 OK, thanks y'all 2010-05-27T10:48:40 you can also import and export records, and use marcedit to manipulate them while they're out of Evergreen 2010-05-27T11:10:32 *** cerpy has quit IRC 2010-05-27T11:25:29 *** jamesrf has joined #evergreen 2010-05-27T11:46:43 *** artunit_ has joined #evergreen 2010-05-27T11:46:49 *** dbwells_ has joined #evergreen 2010-05-27T11:47:49 *** artunit has quit IRC 2010-05-27T11:47:52 *** artunit_ is now known as artunit 2010-05-27T11:47:57 *** dchristens has joined #evergreen 2010-05-27T11:48:27 *** dbwells has quit IRC 2010-05-27T11:48:29 *** dbwells_ is now known as dbwells 2010-05-27T11:52:48 Hi folks. We're going to be upgrading from 1.6.0.0 to 1.6.0.4, and at the same time doing some OS updates. Would it make more sense to apply the OS updates and then incremental EG updates, or export EG data from Postgres, apply OS updates, build latest EG, and import data? What do you think? 2010-05-27T11:55:38 dchristens: I'd suggest OS first, but why would you need to do anything to postgres? I would /not/ suggest doing OS and Postgres and Evergreen upgrades all at the same time 2010-05-27T11:55:49 too many moving parts, IMO 2010-05-27T11:55:58 (my concern is that EG might be compiled against libraries that would be upgraded as part of the OS updates) 2010-05-27T11:56:12 miker_: yep, that worries me. 2010-05-27T11:57:59 I assume this is just an update, not a full version upgrade (like, just updates to some karmic packages, not an upgrade from karmic to lucid ... to use ubuntu as an example) 2010-05-27T11:58:17 miker_: correct 2010-05-27T11:59:12 in that case, updates should be safe ... there should be no ABI incompatabilities within a single release of an OS (for redhat or ubuntu or debian style releases) 2010-05-27T11:59:24 gentoo is a different story, of course ;) 2010-05-27T12:00:31 :-) 2010-05-27T12:01:34 So I should be pretty safe doing incremental updates to EG after the OS updates.... good - that makes me happy :-D 2010-05-27T12:02:22 you should, yes. just don't upgrade pg across major versions without backups ;) 2010-05-27T12:02:50 miker_: that way lies madness! 2010-05-27T12:04:09 miker_++ 2010-05-27T12:07:01 miker_: you have no sense of adventure whatsoever, sir 2010-05-27T12:07:17 :) 2010-05-27T12:09:54 "Adventure is just bad planning" - Roald Amundsen 2010-05-27T12:10:42 or an old atari game 2010-05-27T12:10:50 :-) 2010-05-27T12:14:03 is except Foo as Bar: invalid syntax in some versions of python but not others? 2010-05-27T12:27:59 python 2.6 2010-05-27T12:28:05 is what is needed, I think 2010-05-27T12:30:43 phasefx: wish I could help - the only python I know is Monty.... 2010-05-27T12:32:02 * phasefx is playing with citeproc-js, which uses python for the test suite 2010-05-27T12:34:08 *** ghid0rah has joined #evergreen 2010-05-27T12:34:36 Hello, has anyone gotten evergreen running on Ubuntu 10.04? 2010-05-27T12:36:51 ghid0rah: not in production. using the karmic deps it should be straightforward, though 2010-05-27T12:42:57 *** youdonotexist has quit IRC 2010-05-27T12:45:27 *** youdonotexist has joined #evergreen 2010-05-27T12:47:12 *** brendan_bywater has quit IRC 2010-05-27T13:04:28 *** pmplett has joined #evergreen 2010-05-27T13:10:53 *** brendan_bywater has joined #evergreen 2010-05-27T13:19:25 miker_: what's the largest number of bibs you've tried to load with in-db ingest in a single transaction? 2010-05-27T13:26:43 Dyrcona: a little over 11k 2010-05-27T13:27:21 miker_: I'm doing 15 simultaneous batches of 50,000 each. Should I be doing smaller batches? 2010-05-27T13:27:47 miker_: I started this first set yesterday at about 3:00 pm EDT. 2010-05-27T13:28:22 mkier_: 50,000 used to work well with 1.4 and 1.6, before in-db ingest. 2010-05-27T13:28:23 batch size shouldn't matter ... watching top, are you seeing all 15 do stuff all the time, or do they seem to be blocking? 2010-05-27T13:28:37 top on the server? 2010-05-27T13:28:40 yes 2010-05-27T13:28:50 lemme check. 2010-05-27T13:29:00 the goal is to see all of the COPY streams using lots of CPU 2010-05-27T13:29:38 they were yesterday until sometime in the night. now the load average has really dropped. 2010-05-27T13:30:29 hrm... and the ingest.metarecord_mapping.skip_on_insert flag is enabled? 2010-05-27T13:30:32 my top 15 processes are all postrgres and the top 6 or so of those are using between 10 and 20 % CPU. 2010-05-27T13:30:43 lots of IO wait? 2010-05-27T13:30:55 when i look with PG Admin, i see them running but I can't tell what they're doing. 2010-05-27T13:31:10 I think the IO part is mostly over. 2010-05-27T13:31:14 mind pasting the top screen? 2010-05-27T13:31:50 I'm trying to,a ctually, but its hard to select.. :) 2010-05-27T13:31:55 IO continues for the whole time with in-db ingest ... not just during the initial COPY into bre 2010-05-27T13:32:14 you could q to exit and it'll stop refreshing ;) 2010-05-27T13:32:51 Dyrcona pasted "top" at http://paste.lisp.org/display/100577 2010-05-27T13:33:46 miker_: yes, i did that. 2010-05-27T13:33:47 :) 2010-05-27T13:35:45 *** wiggles has joined #evergreen 2010-05-27T13:36:07 miker_: I am considering going back to the code that I used for 1.4/1.6 that did the ingest outside of the database. I think this is taking too long. 2010-05-27T13:37:07 they've all use about the same amount of CPU time, so things are happening. I wonder if it's just trying to process deferred fkey tirggers 2010-05-27T13:37:14 omg 2010-05-27T13:37:24 srsly? 2010-05-27T13:37:30 Does anyone have op? Can we boot lisppaste? 2010-05-27T13:37:49 looks like they stopped it 2010-05-27T13:38:00 they hopefully being paste.lisp.org 2010-05-27T13:38:02 Too late, I'm ignoring lisppaste3. 2010-05-27T13:38:25 heh .. I'm not, and no more happening 2010-05-27T13:38:48 so, anyway. what should I do? 2010-05-27T13:39:00 Dyrcona: so, another option would be to use INSERT instead of COPY, and do it outside a transaction 2010-05-27T13:39:13 just 15 piles of bare inserts 2010-05-27T13:39:16 omg 2010-05-27T13:39:29 i thought insert would be slower, that's why I'm doing the copy. 2010-05-27T13:40:08 that would also require a lot more extensive changes to my code. i'd have to do threaded java or something. 2010-05-27T13:40:37 * phasefx will wipe that spam from the logs later 2010-05-27T13:40:45 must have a rate limit 2010-05-27T13:40:50 well, if my theory about a huge pile (50k * ~100) fkeyss is correct, it would spread the load instead of doing them all at the end 2010-05-27T13:41:26 Dyrcona: oh, you're starting a copy inside a program and then pushing copy data in a loop? 2010-05-27T13:41:37 and only evergreen http://paste.lisp.org/list 2010-05-27T13:41:37 i have something like 930,000 bibs to load 2010-05-27T13:41:45 miker_: no. 2010-05-27T13:42:08 if you're generating files, just generate insert statements instead of copy data lines, no? 2010-05-27T13:42:39 miker_: I'm now writing a copy data file for each input file and a sql that fires the copy inside a transaction, also for each input file. 2010-05-27T13:42:53 right 2010-05-27T13:43:45 just write "INSERT INTO ... VALUES ( quoted stuff here )" instead of "tabbed stuff here" to the file 2010-05-27T13:43:51 miker_: then I wrote a little C program which I named "bibbomb" to load them simultaneously using fork() and execl(). i can control how many loads happen simultaneously via a command line option. 2010-05-27T13:44:03 miker_: yes. i can do that. 2010-05-27T13:44:06 well 2010-05-27T13:44:30 miker_: do you think not using a single transaction would help? 2010-05-27T13:44:44 if you've got the bibs suitably scrubbed that marc2bre can read them, you don't have to write /any/ code 2010-05-27T13:44:48 miker_: single transaction per file, i means. 2010-05-27T13:45:42 miker_: i do most of the scrubbing after conversion to marc21slim, it's easier to clean up the xml than the raw MARC. 2010-05-27T13:45:44 there's a point of diminishing returns with fkey-heavy data when it comes to the "use one transaction to be faster" rule, so it's possible 2010-05-27T13:46:19 *** dbs has joined #evergreen 2010-05-27T13:46:30 marc2bre can take marcxml, and it's hella faster with that instead of marc21 2010-05-27T13:46:31 the real advantage to NOT using a single transaction would be that bad records wont' prevent other from loading. 2010-05-27T13:47:02 Dyrcona: that, and if you've got so many fkey checks per transaction that it's faster to break them up 2010-05-27T13:47:21 mrpeters-isl: why not use the 856 $9 subfield in the records that you want to be transparent, with $9 set to CONS or SYS or whatever your consortial OU shortname is? 2010-05-27T13:49:20 miker_: i always found my own bib load solution faster than marc2bre, et al., until in-db ingest came along. 2010-05-27T13:49:51 miker_: also marc2bre would choke on many of our records as dumped from our ILS. 2010-05-27T13:50:10 Dyrcona: right, don't use it until you've scrubbed the records 2010-05-27T13:50:17 Dyrcona: likely due to the NFD vs. NFC vs. JavaScript fun 2010-05-27T13:50:22 if marc2bre chokes on a record, many other parts will too 2010-05-27T13:50:31 parts of evergreen 2010-05-27T13:52:56 dbs: I don't think marc2bre uses any js 2010-05-27T13:53:26 miker_: I was thinking of fingerprinting, which you're right was in ingest 2010-05-27T13:54:34 well, i'll just leave the current mess running. 2010-05-27T13:55:11 i'm going on vacation for two weeks starting Tuesday (Monday being a holiday), so I'll see what's going on when I get back. 2010-05-27T13:57:26 heh 2010-05-27T13:58:04 i think Tom is going to update everything though towards the end of the second week, so it may all get killed if it hasn't finished by then. 2010-05-27T13:58:34 miker_ might kill himself if it isn't finished by then :) 2010-05-27T13:59:05 Dyrcona: I wish I could help more, but without seeing what's actually happening (or knowing the tools your using) I got nuthin right now 2010-05-27T13:59:54 Yeah. A zip file with your framework with a few sample files of MARC records would probably be helpful. 2010-05-27T14:01:49 it's unfortunate, because I was able to load 5k records in about 1 minute on a 2-core athalon desktop ... 2010-05-27T14:01:58 but I understand your frustration 2010-05-27T14:02:13 well, i think its that i'm loading so many records in a single transaction. 2010-05-27T14:02:39 50,000 on average with the last group having the left over 35,000 or so. 2010-05-27T14:03:00 checkpoint_segments = 3? 100? 2010-05-27T14:03:05 it could be, which is why I suggested bare inserts, but more files would work too, and gain the COPY benefits. I was using 1k records per file 2010-05-27T14:03:11 that worked fine with ingest outside of the database, but now that it is in-db... 2010-05-27T14:03:20 dbs: heh, that's another thing to look at, yep 2010-05-27T14:03:47 dbs: i did see a message about checkpoints in the postgres logs. 2010-05-27T14:03:54 a copy of postgresql.conf could be useful info :) 2010-05-27T14:04:53 gotta love debian and its "children" : /etc/postgresql/8.4/main 2010-05-27T14:05:10 everything so specific... 2010-05-27T14:07:27 looks like checkpoint_segments is set to the default. the line is commented out. that's 3, isn't it? 2010-05-27T14:08:27 probably ... you want a high number for ~1M record batch load ;) ... 100 (or 500) wouldn't be out of the question 2010-05-27T14:14:38 *** scottmcd has joined #evergreen 2010-05-27T14:21:58 well i just killed the parent and all of its children. (UNIX is so violent.) 2010-05-27T14:23:35 lol 2010-05-27T14:24:53 So, is the database on EG tend to be more disk, RAM, and/or CPU heavy? 2010-05-27T14:25:07 yes. 2010-05-27T14:26:24 so, i'm going to change checkpoin_segments to 100, and try with batches of 10,000 records each. 2010-05-27T14:26:41 you think I should also go with insert instead of copy? 2010-05-27T14:27:17 *** dbs has quit IRC 2010-05-27T14:27:19 Dyrcona: shrinking the batches would be a good first step 2010-05-27T14:27:42 scottmcd: if you have enough RAM, they're CPU heavy 2010-05-27T14:27:50 I'm looking at a 32 core, 128GB, 8 spindle database box and I'm trying to decide if the RAID controller will have sufficient IOPS to support the CPU and RAM. 2010-05-27T14:27:53 if you don't, they're IO-bound 2010-05-27T14:28:14 scottmcd: with your datasize, you'll have all of your hot data in RAM 2010-05-27T14:28:27 which is FAST! 2010-05-27T14:28:31 sweet 2010-05-27T14:29:09 get lots and lots of disk space! thats our bottleneck right now 2010-05-27T14:30:06 *** dbs has joined #evergreen 2010-05-27T14:30:09 scottmcd: I'd be surprised if your initial load of the db was more than 100G (now, it grows as data is generated and bibs are added, etc, but only about 1/2 of the full data size is "hot" for a similarly sized install at 200G after 4 years of growth) 2010-05-27T14:30:28 if that makes sense :) 2010-05-27T14:31:03 It does. 2010-05-27T14:31:04 scottmcd: and the data that grows fastest for you (circ/hold data) is the smallest 2010-05-27T14:32:55 *** dbs has quit IRC 2010-05-27T14:32:57 Right now, our NAS/backup server is spec'd with 2 1Gb ports. Will we need more bandwidth than that? 2010-05-27T14:33:12 *** dbs has joined #evergreen 2010-05-27T14:34:22 IMO, no. logs are a trickle, and backups are bursty and off-hours 2010-05-27T14:35:58 our logs are pretty hefty! 2010-05-27T14:36:16 mrpeters-isl: yeah, but not 1Gb/s heafty ;) 2010-05-27T14:36:32 syslog is on a seperate box, anyway. 2010-05-27T14:36:36 ~= 100MB/s 2010-05-27T14:36:47 almost 500GB this month 2010-05-27T14:38:41 mrpeters-isl: that's 200k/s ... even compressing that to 8h/day instead of 24, it's 0.6MB/s 2010-05-27T14:38:57 it's big in aggregate, but it's not a torrent 2010-05-27T14:39:10 yeah, i wasn't meaning to imply that 2010-05-27T14:39:15 just that you want lots of space to keep them around :) 2010-05-27T14:40:24 well, and compress them, and cull the useless ones (most (by size) become worthless very quickly) 2010-05-27T14:40:40 Our config is at 1.8TB or 2.1TB, depending on whether we decide to use a hot spare. 2010-05-27T14:40:51 for syslog 2010-05-27T14:40:54 yeah, ive been having to keep them around for 30 days to do some estimated "searches per month" counts by grepping them 2010-05-27T14:40:55 keep activity.log forever, keep pg logs for maybe a week 2010-05-27T14:41:04 always use a hot spare 2010-05-27T14:41:29 mrpeters-isl: is that 30 days uncompressed? 2010-05-27T14:41:43 yeah 2010-05-27T14:41:48 we have mega space on the apple 2010-05-27T14:42:39 yeah ... so, with compression and culling you'd be looking at /maybe/ 5G/mo long term storage 2010-05-27T14:42:55 scottmcd: and, listen to atheos on the hot spare :) 2010-05-27T14:44:12 miker_: Easy enough. :P 2010-05-27T14:45:10 and if you ever want to hear a good war story on why, find me at the next Evergreen conference. I start shaking and trembling when I tell the story, so it must be shared over beers 2010-05-27T14:46:26 Heh. 2010-05-27T14:46:36 lol 2010-05-27T14:49:04 new policy, Raid 1+0 w/hot spare no exceptions 2010-05-27T14:50:06 atheos: how about the policy for shutting down servers during the middle of the day? :) 2010-05-27T14:50:18 dbs: we'd never do such a thing :) 2010-05-27T14:50:30 at least, not on purpose! 2010-05-27T14:50:37 you just had to say that didn't you dbs ! 2010-05-27T14:50:48 * mrpeters-isl waits for phone to ring with network failure messages 2010-05-27T14:51:13 our member libraries have become extremely tolerant of this (runs and hides) 2010-05-27T14:52:33 my failsafe response for calming any staff member down over evergreen issues is "remember SIRSI". Paifies them every time 2010-05-27T14:52:40 hahaha 2010-05-27T14:52:55 atheos: that doesn't work in PINES anymore, unfortunately 2010-05-27T14:52:58 :-) 2010-05-27T14:53:26 it won't long up here either, we have been experiencing a high turnover rate this year. I have no idea where people are going in this economy 2010-05-27T14:53:36 3 half-days of downtime erases a year of stable uptime immediately 2010-05-27T14:53:45 insert "work" between won't & long 2010-05-27T14:53:50 OUCH chrissharp123 2010-05-27T14:54:23 # PINES = low fault tolerance 2010-05-27T14:54:42 bah, that's what standalone is for :) 2010-05-27T14:54:58 Of course, I'm jinxing myself, but we haven't had any unexpected downtime since November 2009 2010-05-27T14:55:11 speaking of standalone....it would be relaly nice if the due date + x days pulled from the circ matrix 2010-05-27T14:55:21 we have a +30 days modifier, but no circ modifiers that allow for that long of circ 2010-05-27T14:55:22 at 99.86% uptime and rising... 2010-05-27T14:55:30 yet, we need a +21 days modifier and dont :( 2010-05-27T14:55:31 dbs - any migrations over that period of time? 2010-05-27T14:55:33 the funny thing about that is that when we were down, we had been stable so long that people had forgotten how to use standalone :-) 2010-05-27T14:55:38 atheos: one 2010-05-27T14:55:43 nice going 2010-05-27T14:56:25 lots of human sacrifices to the sys admin gods paid off 2010-05-27T14:56:35 hmm. perhaps that one isn't live yet. I've been away for all of May 2010-05-27T14:56:55 we're about to order enough hardware to match what we currently have, so if all goes well, we can get this a bit more stable 2010-05-27T14:57:10 where would one modify those "Today + 14 Days" type messages in standalone? 2010-05-27T14:57:20 maybe i can add a + 21 option easilly 2010-05-27T14:57:32 phasefx: i bet you know! 2010-05-27T14:57:52 yeah, I think so, but you'll have to package a new client 2010-05-27T14:58:01 crap! i was afraid of that! 2010-05-27T14:58:12 could we lobby for Today + 21 days to be included in the next staff client release? 2010-05-27T14:58:31 maybe even + 28 days 2010-05-27T14:58:36 you have 1 week, 2 weeks, why not 3 and 4 :) 2010-05-27T14:59:21 watch someone complain that the drop-down is getting too large :) 2010-05-27T14:59:33 :( 2010-05-27T14:59:39 2 more wont hurt! take out 30 days haha 2010-05-27T14:59:43 but send a patch for offline_checkout.xul and lang.dtd, look for common.date_helper 2010-05-27T14:59:57 that's the great thing about change. doesn't matter what it is, better, worse, this, that. someone will complain about it 2010-05-27T15:00:02 is that a local file, or server side? 2010-05-27T15:00:25 * dbs kind of regrets putting that into an entity, as the translated versions probably don't work particularly well (unless postgres is really, really smart) 2010-05-27T15:00:26 lang.dtd is both, offline_checkout.xul is local, but they're all in the same source repository 2010-05-27T15:02:03 http://svn.open-ils.org/trac/ILS/browser/branches/rel_1_6/Open-ILS/xul/staff_client/chrome/content/circ/offline_checkout.xul that the one? 2010-05-27T15:02:14 with a bit more dev work, we could make it data-driven 2010-05-27T15:02:28 yeah, that file controls what you see in the offline checkout interface 2010-05-27T15:02:52 the dtd entities &common.date_helper.3_days; etc. need to be defined in lang.dtd 2010-05-27T15:02:53 *** scottmcd_ has joined #evergreen 2010-05-27T15:03:02 phasefx: just locally cache the copy info for each barcode, and the patron info for each patron, and circ rules. EASY. 2010-05-27T15:03:38 dbs: that's a bit too data driven for me :D 2010-05-27T15:03:41 ok, not finding lang.dtd 2010-05-27T15:04:11 some folks can't even load the entire offline patron list as it is, without getting script delay alerts 2010-05-27T15:04:26 mrpeters-isl: it's in opac/locale 2010-05-27T15:04:34 thank you! 2010-05-27T15:04:49 phasefx: which is clearly why the answer is to just install a self-contained copy of EG on each bookmobile, then just write code to sync circ transactions between EG databases 2010-05-27T15:05:18 *** scottmcd__ has joined #evergreen 2010-05-27T15:05:22 gmcharlt: an offline/disconnect mode for FulfILLment nodes :) 2010-05-27T15:05:31 :) 2010-05-27T15:06:11 so lang.dtd must go both server side, and then where client side? 2010-05-27T15:06:13 Or use VoltDB. Which is magical in so many ways, but most importantly in transmitting data without power 2010-05-27T15:06:38 *** scottmcd has quit IRC 2010-05-27T15:07:00 C:\Program Files\Evergreen Staff Client\build\chrome\locale\en-US ? 2010-05-27T15:07:05 mrpeters-isl: when the staff client is "compiled", it gets its own copy of lang.dtd 2010-05-27T15:07:35 *** scottmcd__ is now known as scottmcd 2010-05-27T15:07:35 *** StephenGWills has left #evergreen 2010-05-27T15:07:40 *** scottmcd_ has quit IRC 2010-05-27T15:07:51 a lot of this stuff is done the way it is for historical reasons. We could refactor things, and let the chrome parts of the client have its own dtd file 2010-05-27T15:08:30 way back when, there was this notion that we could move files between local/remote as desired 2010-05-27T15:09:06 mrpeters-isl: see http://evergreen-ils.org/dokuwiki/doku.php?id=scratchpad:packaging_the_staff_client_for_windows for how to build and package a client (post-patch) 2010-05-27T15:09:43 :) was gonna ask about that next 2010-05-27T15:10:20 im getting 3x "14 days" options now in offline 2010-05-27T15:10:30 doh, ignore me 2010-05-27T15:10:47 but if you're building from subversion (instead of a tarball), you'll need to run ./autogen.sh (not /openils/bin/autogen.sh) before ./configure 2010-05-27T15:14:03 wow, that was painless 2010-05-27T15:17:40 inno script helps ALOT 2010-05-27T15:18:39 *** dbs has quit IRC 2010-05-27T15:22:22 phasefx: stupid question. where is Inno putting my final exe? i can't find it! 2010-05-27T15:23:17 in a directory named Output 2010-05-27T15:23:48 ah yes 2010-05-27T15:25:32 hmm i get an error "cannot parse application.ini" when starting 2010-05-27T15:25:47 but now its gone 2010-05-27T15:25:48 interesting 2010-05-27T15:26:03 what version of xulrunner are you using? 2010-05-27T15:26:24 1.93 or whatever was in your script 2010-05-27T15:26:35 you'll want to use 1.9.0.17 for 1.6 evergreen 2010-05-27T15:26:43 oooh good to know 2010-05-27T15:27:30 mrpeters-isl: on that wiki page, skip down to 1.6.0.4 2010-05-27T15:28:10 believe it or not, it WAS on their site 2010-05-27T15:28:14 http://download02.mozilla.org/pub/mozilla.org/xulrunner/nightly/1.9.0.17-candidates/build1/ 2010-05-27T15:28:23 but ill use yours since we know it worsk 2010-05-27T15:28:26 yeah, they brought it back afterward 2010-05-27T15:28:58 or rather, that they've juggled directories and hostnames some, and I probably thought it was gone 2010-05-27T15:29:19 no 1.9.0.17 in https://releases.mozilla.org/pub/mozilla.org/xulrunner/releases/ , for example 2010-05-27T15:31:11 *** dbs has joined #evergreen 2010-05-27T15:31:50 mrpeters-isl: btw, trunk has a new print strategy a library can use if Mozilla can't print to their printers. Might not be perfect (I think some printers may truncate long lines, etc.) 2010-05-27T15:35:12 cool! 2010-05-27T15:35:19 where should i submit this patch to? 2010-05-27T15:35:22 if anyone were to want it 2010-05-27T15:35:49 * dbs plans to continue being non-productive by shifting over to Fedora 13 as his new environment. Should be good for fleshing out our distro support, anyway 2010-05-27T15:37:18 dbs: vacation to recover from your vacation :) 2010-05-27T15:37:56 mrpeters-isl: OPEN-ILS-DEV would be great, thanks! You'd definitely want the community to adopt it, otherwise, you'd be maintaining those changes and rolling your staff clients forever 2010-05-27T15:38:49 mrpeters-isl: don't forget to include a DCO (I forget the link) 2010-05-27T15:39:03 phasefx: could we make it an OU setting; an array of time deltas that gets populated at client connection time? with basic defaults? 2010-05-27T15:39:42 I don't see why not, other than we don't have good support for managing fancy org unit settings like that 2010-05-27T15:40:20 the last time I wanted to work on that, we ended up making a dedicated interface for the setting in question for expediency 2010-05-27T15:40:42 circ.offline_intervals = ["Today + 14 days", "Today + 28 days", "Today + 1 month"] -- yeah, maybe someday somebody will come up with a data-editor widget 2010-05-27T15:40:48 dbs: we're already delivering some options from online connections, like surveys and patron groups 2010-05-27T15:41:45 circ.offline_intrvals = [ { "Today + 14 days" : 14 }, { "Today + 28 days" : 28 } ] 2010-05-27T15:41:57 or something 2010-05-27T15:42:23 someone could put [ { "DVD's" : 14 } ] 2010-05-27T15:43:26 making end-user staff work with JSON from a GUI kind of sucks :) 2010-05-27T15:44:34 ok, patch submitted to dev 2010-05-27T15:44:43 thanks for the help, phasefx 2010-05-27T15:45:36 mrpeters-isl: rock, I'll give it a test and pull it in sometime this week/weekend 2010-05-27T15:45:54 cool! compiled client included, too 2010-05-27T15:46:39 scary :) hope you changed the innoscript file some, don't want folks to use that and have it get confused with stock clients when talking to support folks 2010-05-27T15:47:03 it is changed to read "Custom Offline Clieint" 2010-05-27T15:47:11 mrpeters-isl++ 2010-05-27T15:47:13 AppVerName=Evergreen Staff Client v1.6.0.0 (Custom Offline Mode) 2010-05-27T15:47:23 i suppose i should have made that show up on the application though 2010-05-27T15:47:51 ill recompile with that in place 2010-05-27T15:47:56 1.6.0.0 has other bugs in the local client that have since been fixed, but no biggie 2010-05-27T15:48:06 yeah, we're just all on 1.6.0.0 in indiana for now 2010-05-27T15:48:10 ++ for mpresters patch. 2010-05-27T15:48:23 mrpeters-isl: getting your feet wet is what's important to me :) 2010-05-27T15:48:39 my typing fingers don't work today. 2010-05-27T15:48:41 yeah! 2010-05-27T15:48:49 more folks who can mess with the client, the better for me and for Evergreen 2010-05-27T15:49:13 when migrations stop, that's definetly somewhere i want to start playing around 2010-05-27T15:50:01 *** dbs_ has joined #evergreen 2010-05-27T15:52:39 oh my 2010-05-27T15:53:08 just ignore lisppaste3 2010-05-27T15:53:17 *** dbs has quit IRC 2010-05-27T15:53:25 people are lame 2010-05-27T15:53:25 *** lisppaste3 has quit IRC 2010-05-27T15:53:35 well, that's an option... :) 2010-05-27T15:54:17 someone from ESI should ask for op in the channel. i'm not sure what Freenode's procedures are. 2010-05-27T15:54:33 yeah i agree an OP would be good 2010-05-27T15:55:01 doesn't need to be from ESI, but I'll poke 2010-05-27T15:57:10 dbs_: it may need to be you, you're the founder of the channel 2010-05-27T15:57:44 *** emrikol has joined #evergreen 2010-05-27T16:02:31 phasefx: is the ReferenceError: font_helper is not defined something fixed in subsequent 1.6 staff clients? 2010-05-27T16:02:39 comes up when first entering offline checkout 2010-05-27T16:03:04 yes 2010-05-27T16:03:16 cool. i dont think it hurts anything, just a nag :) 2010-05-27T16:03:33 fwiw, I think you could use a 1.6.0.4 client (or rel_1_6_0) against your 1.6.0.0 server with no trouble 2010-05-27T16:03:50 good to know. i was a bit nervous about trying 2010-05-27T16:03:50 the intention is for compatability within a major release 2010-05-27T16:04:44 though an old 1.6 client isn't going to have easy access to booking once that comes out in 1.6.x 2010-05-27T16:05:00 i think once the major release comes out we'll upgrade 2010-05-27T16:05:35 *** jamesrf has quit IRC 2010-05-27T16:05:38 just didn't see much benefit in dedicating the resources to upgrading from 1.6.0.0 to one of the minor releases since it would require new staff client installs (was chaos, as simple as it SHOULD be) for some of our less technical libraries 2010-05-27T16:07:08 it's going to get more fun when we change xulrunners.. you used to be able to just install over old staff client installations, but when moving to xulrunner 1.9.2, there can be stuff left over from a younger installation that prevents xulrunner from operating. So we'll need to do something to erase such problem files if they exist 2010-05-27T16:07:22 ew yeah that will be tons of fun 2010-05-27T16:07:53 but we want to follow xulrunner/firefox now that they're in sync 2010-05-27T16:08:24 could even start pushing the client as a firefox extension 2010-05-27T16:23:15 *** wiggles has left #evergreen 2010-05-27T16:25:51 *** dbs_ has quit IRC 2010-05-27T16:39:31 *** alxp has quit IRC 2010-05-27T16:49:54 *** kmlussier has quit IRC 2010-05-27T17:01:16 *** weather15 has joined #evergreen 2010-05-27T17:01:23 Hello Everybody 2010-05-27T17:03:41 It seems that when I ran the makefile postgresql 8.4 was installed now when I go to create the database I recieve this error: createdb: database creation failed: ERROR: new encoding (UTF8) is incompatible with the encoding of the template database (SQL_ASCII) HINT: Use the same encoding as in the template database, or use template0 as template. what should I do? 2010-05-27T17:05:02 *** jamesrf has joined #evergreen 2010-05-27T17:05:15 I came across this: createdb -E UTF8 -T template0 foo is this what I should run? 2010-05-27T17:05:59 weather15: you may need to use pg_dropcluster and pg_createcluster, or initdb 2010-05-27T17:06:19 exactly how do I do that? 2010-05-27T17:06:19 *** bshum has quit IRC 2010-05-27T17:06:38 weather15: I don't know, and if you do it, it'll erase all data you might have in postgres 2010-05-27T17:06:57 I don't have any in it yet as i'mjust trying to install 2010-05-27T17:07:50 found this: Usage: /usr/bin/pg_dropcluster [--stop] 2010-05-27T17:08:30 no harm in trying that createdb line you found first 2010-05-27T17:08:44 okay will try that 2010-05-27T17:09:03 you're using ubuntu 9.10? 2010-05-27T17:09:21 Yes, I believe i need to run this instead though: createdb evergreen -E UTF8 -T template0 foo 2010-05-27T17:09:53 read this and comments: http://jacobian.org/writing/pg-encoding-ubuntu/ there is some mention of template0 in a link from the comments 2010-05-27T17:10:40 I guess I will run his solution code 2010-05-27T17:11:02 since I have no data yet 2010-05-27T17:11:50 okay, I created a new cluster as he said 2010-05-27T17:12:30 Created the Datbase! 2010-05-27T17:13:28 Perhaps that should be added to the wiki installinstructions 2010-05-27T17:13:33 In fm_IDL.xml what does this mean? 2010-05-27T17:14:29 moodaepo: I believe it means that field isn't represented in the database, but gets populated elsewhere (middle layer, etc.) 2010-05-27T17:14:44 moodaepo: that the status field on the object is not persisted in the database ... it's either computed or a has_many linked field 2010-05-27T17:15:06 * phasefx got half of it right 2010-05-27T17:15:42 (or a might_have linked field, but there are very few of those) 2010-05-27T17:16:03 miker_ phasefx: that's what I was guessing, trying to build a report which needs to pull holds which are fulfilled by new items with copy status "In process" so thought I could use that but no luck 2010-05-27T17:16:36 oh wait other way around: items with "In process" which fulfill holds 2010-05-27T17:16:39 weather15: yeah, a blurb for Ubuntu 9.10 (or maybe some generic sanity check of encoding and workaround) would be good 2010-05-27T17:17:21 moodaepo: items that are in process and mentioned as the fulfilling copy on a hold? how'd that happen? :) 2010-05-27T17:18:15 Database Tables Issue: http://pastebin.com/2E5KJDfF is this good or bad? 2010-05-27T17:18:17 miker_: don't ask me I've just been told do do it : ) btw I wanted to take you up on testing 1.6 (i.e. 1.6.1) but after we figure somethings out 2010-05-27T17:18:36 probably next week unless someone gets it done by then 2010-05-27T17:19:05 moodaepo: thanks, every bit helps 2010-05-27T17:19:06 weather15: I'd say not good 2010-05-27T17:19:26 okay what should I do? 2010-05-27T17:19:28 weather15: I think those are harmless, and would dissappear if you re-ran the command a second time 2010-05-27T17:19:37 okay will try and re run 2010-05-27T17:20:34 *** Dyrcona has quit IRC 2010-05-27T17:20:42 now I get a wholeunch more 2010-05-27T17:20:59 of errors? 2010-05-27T17:21:49 yes, this is as far back as my putty session would go: http://pastebin.com/rXce8x3D 2010-05-27T17:23:41 yeah, that's pretty bad. You can redirect that output into a file, by appending > db.out to the command. Then you can do less db.out to view the output in its entirey. The first ERROR in the file should be the one we focus on 2010-05-27T17:24:00 okay willpost in minute 2010-05-27T17:24:37 *** dchristens has quit IRC 2010-05-27T17:25:42 miker_: Now that I think about it how about this answer the reason there are holds that might be fulfilled by this new item which has status "In process" is because it's a new copy and holds are bib level? Does that make sense or am I wrong about holds being bib level? 2010-05-27T17:27:08 such items won't be captured for a hold, because items don't stay In Process when checked in (which is the method of capturing) 2010-05-27T17:27:22 but they can be targeted 2010-05-27T17:28:00 moodaepo: if In Process is marked as holdable for you, what phasefx said 2010-05-27T17:28:22 or rather, the items can be captured, but you won't find captured items with an In Process status 2010-05-27T17:28:41 if you don't want In Process itmes to be targeted, just make In Process not holdable 2010-05-27T17:28:43 not under natural conditions, anyway 2010-05-27T17:28:44 looks like this is the first error: psql:012.schema.vandelay.sql:1: ERROR: schema "vandelay" does not exist 2010-05-27T17:28:58 weather15: are you building from a tarball? 2010-05-27T17:29:10 weather15: those are fine. you'll get one per schema 2010-05-27T17:29:20 weather15: what's the next ERROR? some of those errors will be harmless, from a DROP SCHEMA command 2010-05-27T17:29:39 but those ERROR's on INSERT are bad 2010-05-27T17:31:30 i'm still going through the file but all the errors seem just like the one before 2010-05-27T17:32:29 this is the last error: psql:extend-reporter.sql:17: ERROR: schema "extend_reporter" does not exist there all just like this one 2010-05-27T17:33:16 so I guess all is good then 2010-05-27T17:33:42 phasefx since we want them holdable how do I target them? 2010-05-27T17:33:51 weather15: what does: SELECT * FROM biblio.record_entry; return? 2010-05-27T17:34:16 if a single row with an id of -1, then your paste must be from a previous run 2010-05-27T17:34:36 weather15: if grep ERROR db.out | sort | uniq -c shows that each of those schema errors only occur once, then you're probably okay 2010-05-27T17:35:07 *** brendan_bywater has quit IRC 2010-05-27T17:35:39 output of that secound command: http://pastebin.com/NXRc4dgX 2010-05-27T17:36:07 looks good to me 2010-05-27T17:36:27 well, there's not one for each table 2010-05-27T17:36:42 odd, Okay should I still run SELECT * FROM biblio.record_entry; 2010-05-27T17:36:58 are you comfortable with using psql? you'll need to use miker_'s SELECT command from within it 2010-05-27T17:37:17 have used mysql so I will attempt 2010-05-27T17:40:09 what credientials can I use to login to postgres? 2010-05-27T17:40:41 the same ones you used with with your eg_db_config.pl command 2010-05-27T17:41:11 I get this: psql: FATAL: Ident authentication failed for user "evergreen" 2010-05-27T17:41:20 you created a database, and a user/password when you were the postgres user. Use those 2010-05-27T17:43:11 I ran this: psql -d evergreen -U evergreen -W entered password and got this: psql: FATAL: Ident authentication failed for user "evergreen" 2010-05-27T17:43:13 *** lisppaste3 has joined #evergreen 2010-05-27T17:43:19 weather15: psql -U evergreen -h your-host/localhost -d evergreen -W 2010-05-27T17:43:50 tahnks 2010-05-27T17:43:53 *thanks 2010-05-27T17:44:49 * phasefx is heading out for now, "good luck! 2010-05-27T17:45:00 result of mikers command: http://pastebin.com/2XYheUQB 2010-05-27T17:45:59 weather15: looks good. all should be well 2010-05-27T17:46:08 Great! Thanks! 2010-05-27T17:46:16 *** chandler has joined #evergreen 2010-05-27T17:46:37 Hey folks. Sorry about the issue with lisppaste3. Apparently some folks think they're awfully clever. 2010-05-27T17:47:12 I've added a few more roadblocks; if it doesn't work, feel free to just kick the bot (it only rejoins if I ask it to). 2010-05-27T17:48:43 chandler++ 2010-05-27T17:49:58 *** sfortin has quit IRC 2010-05-27T17:50:12 *** brendan_bywater has joined #evergreen 2010-05-27T18:01:25 Next problem can't login 2010-05-27T18:02:15 what do you do when you get this: http://pastebin.com/G6mYHfrm 2010-05-27T18:05:50 I ran that tester script and got this looks good:http://pastebin.com/1H5UAf7W 2010-05-27T18:10:06 weather15: is memcached running? 2010-05-27T18:10:50 give me a minute 2010-05-27T18:11:11 Starting memcached: memcached disabled in /etc/default/memcached. 2010-05-27T18:11:26 wonder how that happened or what to do about it 2010-05-27T18:11:41 What, exactly, does memcached cache? 2010-05-27T18:11:43 some ubuntu installs disable it for some reason 2010-05-27T18:12:00 Found Solution: ENABLE_MEMCACHED=yes ~ 2010-05-27T18:12:58 Memcached is now running lets try again 2010-05-27T18:14:04 Error stil apears try restarting opensrf? 2010-05-27T18:14:19 yeah 2010-05-27T18:15:25 Great got output similiar to whats on the wiki. Does thismean Ihavea working install? 2010-05-27T18:16:59 nice to see the Evergreen OPAC running on my server and not the dev server 2010-05-27T18:17:40 is there anything elese I should do to see if the system is ready for small production? 2010-05-27T18:18:43 Have at tops 1000 materials 2010-05-27T18:20:06 there's nothing specific, but i guarantee you'll encounter things 2010-05-27T18:20:41 not surprising Ihope there all fixable 2010-05-27T18:21:40 I saw you guys keep a list of know evergreen libraries you can add us 2010-05-27T18:23:35 i can add you what's your library? 2010-05-27T18:27:44 *** weather15_ has joined #evergreen 2010-05-27T18:27:50 I'm back 2010-05-27T18:28:10 *** weather15 has quit IRC 2010-05-27T18:28:47 we are Greater Laverdiere Family Library Network a small library organized for my family. We have 3 branches as well 2010-05-27T18:30:49 ok do you have a website? 2010-05-27T18:31:39 yes an informational website or evergreen site? 2010-05-27T18:32:12 well whatever you want on that page, a link to an info site or any comments? 2010-05-27T18:32:47 *** youdonotexist has quit IRC 2010-05-27T18:33:01 http://library.laverdieres.com comment: Evergreen is great nomatter what sixe the library 2010-05-27T18:33:15 sorry I'm typing to fast today 2010-05-27T18:33:32 http://library.laverdieres.com comment: Evergreen is great no matter what size the library 2010-05-27T18:34:00 *** chandler has left #evergreen 2010-05-27T18:34:03 er no i meant more of a comment about your library or where you're located or something like that 2010-05-27T18:34:49 Where located in the United states in Rhode Island and Massachusetts 2010-05-27T18:35:52 ok its done :) 2010-05-27T18:36:10 Okay 2010-05-27T18:39:02 As faras SSL goes is there any particular cheap SSL authority Icould look into? 2010-05-27T18:39:21 *** scottmcd has quit IRC 2010-05-27T18:59:16 *** jamesrf has quit IRC 2010-05-27T19:04:30 weather15_: We use StartCom Ltd. 2010-05-27T19:04:35 * moodaepo going home 2010-05-27T19:05:46 *** pmplett is now known as pmpafk 2010-05-27T19:13:08 I was thinking of using startcom 2010-05-27T19:13:19 do they work with all browser without cert errors 2010-05-27T19:13:28 *** jamesrf has joined #evergreen 2010-05-27T19:13:44 *** weather15_ has quit IRC 2010-05-27T21:06:22 *** brendan_bywater has quit IRC 2010-05-27T21:07:14 *** jamesrf has quit IRC 2010-05-27T21:18:19 *** mck9 has left #evergreen 2010-05-27T21:20:52 *** pmpafk is now known as pmplett