2010-01-28T00:14:45 *** jamesrf has joined #evergreen 2010-01-28T00:18:46 *** lisppaste3 has quit IRC 2010-01-28T00:19:12 *** lisppaste3 has joined #evergreen 2010-01-28T00:21:36 *** dbs has quit IRC 2010-01-28T00:34:56 *** jamesrf has quit IRC 2010-01-28T01:51:55 *** pmp_afk has quit IRC 2010-01-28T02:08:52 *** pmp_afk has joined #evergreen 2010-01-28T02:19:50 *** mck9 has left #evergreen 2010-01-28T04:09:33 *** pmp_afk has quit IRC 2010-01-28T06:12:44 *** artunit_ has joined #evergreen 2010-01-28T06:21:50 *** artunit has quit IRC 2010-01-28T07:07:34 *** eguest309 has joined #Evergreen 2010-01-28T07:17:14 *** eguest309 has left #Evergreen 2010-01-28T07:24:09 *** sfortin has joined #evergreen 2010-01-28T07:45:46 *** alxp has joined #evergreen 2010-01-28T07:58:38 *** mck9 has joined #evergreen 2010-01-28T08:11:12 *** pmp_afk has joined #evergreen 2010-01-28T08:15:35 *** eguest309_ has joined #Evergreen 2010-01-28T08:15:36 *** pmp_afk has quit IRC 2010-01-28T08:16:52 *** eguest309_ has quit IRC 2010-01-28T08:56:50 *** Melissa has joined #Evergreen 2010-01-28T09:07:53 *** Melissa has left #Evergreen 2010-01-28T09:11:46 *** StephenGWills has joined #evergreen 2010-01-28T09:12:15 mornin y'all 2010-01-28T09:13:55 before I get too crazy chasin this down. Are there any known problems installing a 1.6 staff client over the top of 1.4 staff client and having a network error reported by the new 1.6 client? 2010-01-28T09:14:12 *** jenny has joined #evergreen 2010-01-28T09:29:09 can someone suggest a direction to take? My staff client is throwing a Network Failure at me. After calling open-ils.actor.rg_unit_setting.values.ranged.retrieve I get a server error with malformed JSON response? 2010-01-28T09:29:59 org_unit_setting NOT rg_unit_setting 2010-01-28T09:31:36 StephenGWills: it's probably not a staff client problem, but a server problem. i'd check to make sure all services started up 2010-01-28T09:31:45 ok 2010-01-28T09:33:52 a ps wax | grep open yields a cornucopia of OpenSRF drones incluing .actors and setting ones. am not sure which specific services would be involved in this negotiation? 2010-01-28T09:36:08 I do have server problems however. clark-kent is complaining about reporter.classic_current.circ not existing.... I'll pooke in the logs for a while. 2010-01-28T09:36:12 thanks berick 2010-01-28T09:41:37 *** dbs has joined #evergreen 2010-01-28T09:45:10 *** artunit_ is now known as artunit 2010-01-28T09:46:00 *** r123 has joined #evergreen 2010-01-28T09:58:30 when I do a stop_all should instances of openurl.pl shutdown? 2010-01-28T09:59:45 nope - that's apache 2010-01-28T10:32:46 *** bshum has joined #evergreen 2010-01-28T10:36:55 all services appear to be running. I have a vague recall of a similar error with an earlier staff client but don't remember: eror from run method: maformed JSON at offset ^\w\d+$ in JSON.pm - this during logon 2010-01-28T10:38:23 client is running on windows and I just ran the installer for 1.6 on top of a system that was running a 1.4.0.6 staff client. 2010-01-28T10:44:21 StephenGWills: So you didn't uninstall 1.4 and install 1.6? 2010-01-28T10:44:52 no... I just ran the intaller for 1.6 2010-01-28T10:44:56 installer* 2010-01-28T10:45:29 I am trying to do an upgrade from 1.4 to 1.6 2010-01-28T10:46:59 I remember doing an upgrade but I uninstalled and installed the client or installed it in a separate directory...then again berick suggested to you it's probably a server side issue 2010-01-28T10:47:20 i actually ran it right over the web without even downloading the 1.6 package onto windows first 2010-01-28T10:48:05 so I can test the client against the demo 1.6 then to prove it is server side. 2010-01-28T10:48:13 just need to get the 1.6 411 2010-01-28T10:48:34 was that sitka or indiana that put up the 1.6 demo site? 2010-01-28T10:48:43 off to DIG, bbilb 2010-01-28T10:55:14 fudgies, the 1.6 test system doesn't seem to be documented on the demo page :( 2010-01-28T10:55:22 anybody got a pointer? 2010-01-28T10:58:31 got it 2010-01-28T11:00:13 client works against indiana's server... it's my server. :( 2010-01-28T11:16:34 StephenGWills: Which of our servers are you pointing at? 2010-01-28T11:17:23 dig.evergreen.lib.in.us and thank you for have it avail! :) 2010-01-28T11:17:30 sure, no problem 2010-01-28T11:17:36 just was curious :) 2010-01-28T11:17:53 now if i can just do a diff of yours against mine starting at / I'll be greatful! ;) 2010-01-28T11:19:01 StephenGWills: pm 2010-01-28T11:25:04 *** sfortin has quit IRC 2010-01-28T11:25:35 *** sfortin has joined #evergreen 2010-01-28T11:28:41 StephenGWills: I've had a few phantom 'Network Failures' since we came up, and when I run out of basic ideas I find restarting the whole stack to be helpful. memcached -> ejabberd -> Evergreen -> apache 2010-01-28T11:29:47 ok. i have been restarting pieces but not the whole schebang. tx. 2010-01-28T11:31:50 *** mrpeters-isl has quit IRC 2010-01-28T11:31:59 *** mrpeters-isl has joined #evergreen 2010-01-28T11:38:15 always worth checking to see if any resources are being overloaded (cpu, mem, io, file handles) 2010-01-28T11:51:47 *** brendan_ga has quit IRC 2010-01-28T11:51:56 *** brendan_ga has joined #evergreen 2010-01-28T11:54:20 after a systematic restart following the trouble shooting script on open-ils.org including all the ps | grep blah stuff.... no dice. still pulling the network failure followed by the skull and cross bones page :( 2010-01-28T11:58:40 did the log poking turn up anything? 2010-01-28T12:02:02 i'm gonna restart the upgrade and figure out what i did wrong. i must have missed something, probably around the eg_db_config steps. 2010-01-28T12:02:43 StephenGWills: Are you working in a Virtual Machine? 2010-01-28T12:04:35 no. metal ubuntu 8.10 box 2010-01-28T12:17:29 clear cache on the client? destroy the client profile? 2010-01-28T12:20:56 *** bshum has left #evergreen 2010-01-28T12:27:15 *** brendan_ga has quit IRC 2010-01-28T12:40:39 *** jamesrf has joined #evergreen 2010-01-28T12:42:48 *** brendan_ga has joined #evergreen 2010-01-28T12:50:12 *** artunit has quit IRC 2010-01-28T12:52:54 how does staff client know what workstation I am? I thought I unintalled the client but on reinstall it still knew what workstation I was. 2010-01-28T12:53:26 clearing cache on client did not help and i don't think the eg_db_config is restartable 2010-01-28T12:54:53 StephenGWills: On Windows, check C:\Documents and Settings\\Application Data\OpenILS\open_ils_staff_client 2010-01-28T12:55:12 profiles are in there 2010-01-28T12:55:32 specifically, a file called ws_info stores the client's registered workstation 2010-01-28T12:56:39 it's located in profiles\{profile foo}\chrome off of the directory dbwells mentioned 2010-01-28T12:56:46 also stuff in C:\Documents and Settings\\Local Settings\Application Data\OpenILS\open_ils_staff_client - I haven't kept the difference straight. 2010-01-28T12:57:36 I generally rename OpenILS to OpenILS.bak in those two directories when I want a completely clean profile 2010-01-28T13:01:20 *** sfortin has quit IRC 2010-01-28T13:02:02 *** sfortin has joined #evergreen 2010-01-28T13:04:36 *** r123 has quit IRC 2010-01-28T13:18:01 *** lisppaste3 has quit IRC 2010-01-28T13:18:11 ok, tx 2010-01-28T13:18:22 *** mrpeters-isl has quit IRC 2010-01-28T13:25:05 no joy, i was able to register but once i tried to login I got the network failure and skull/crossbones again :( 2010-01-28T13:26:12 *** lisppaste3 has joined #evergreen 2010-01-28T13:26:53 *** artunit has joined #evergreen 2010-01-28T13:30:03 i have a vague recollection that the last time i had an error like this, I was missing a stacks record in a table..... 1.6 is sensitive to another missing record in actors that 1.4 was more forgiving of, perhaps? 2010-01-28T13:37:38 Or maybe you're missing a new dependency in 1.6? dunno. hard to tell - you're not giving us much in the way of logs, etc 2010-01-28T13:52:58 sorry dbs, i posted very early this morning with a log error of: After calling open-ils.actor.rg_unit_setting.values.ranged.retrieve I get a server error with malformed JSON response? 2010-01-28T13:53:40 i have discovered just now that I do not have any entries in my org_unit_setting table with org_unit == to the org_unit I told it my client is from... 2010-01-28T13:54:05 hah 2010-01-28T13:54:10 will futz with it this way a while. 2010-01-28T13:54:54 remember to autogen.sh again after adding that entry 2010-01-28T14:02:50 that cleared the JSON error. 2010-01-28T14:04:32 yay!!! 2010-01-28T14:04:55 StephenGWills++ Persistence++ : ) 2010-01-28T14:06:04 this db is hosed though. I'd probably be better off with a fresh 1.6 install and reload my lib data onto it. LOL! 2010-01-28T14:06:07 thanks guys 2010-01-28T14:12:04 * dbs installs ieHTTPHeaders to figure out what the heck IE is doing, and why it's not firing off the desired osrf-http-translator request to open-ils.resolver 2010-01-28T14:36:07 *** b_bonner has joined #evergreen 2010-01-28T14:41:17 *** bshum has joined #evergreen 2010-01-28T14:41:42 *** bshum has quit IRC 2010-01-28T14:58:37 *** StephenGWills has quit IRC 2010-01-28T15:08:09 *** mrpeters-isl has joined #evergreen 2010-01-28T15:41:30 *** alxp has quit IRC 2010-01-28T15:49:37 w00t: http://coffeecode.net/archives/210-Conifer-garners-two-awards-from-the-Ontario-Library-Association.html 2010-01-28T15:50:26 conifer++ 2010-01-28T15:55:46 *** sfortin has quit IRC 2010-01-28T16:12:09 *** jamesrf has quit IRC 2010-01-28T16:15:47 dbs++ 2010-01-28T16:24:27 given a bunch of circs with stop_fines of LOST and checkin_time of null, they could be paid for, or unpaid. cross-ref the circ id with money.billable_xact.id and see if there's an xact_finish? 2010-01-28T16:25:32 either that or mmbxs xact_finish / balance_owed... 2010-01-28T16:43:07 dbs++ 2010-01-28T16:49:34 *** jamesrf has joined #evergreen 2010-01-28T17:26:19 *** jamesrf has quit IRC 2010-01-28T17:31:28 *** jenny has left #evergreen 2010-01-28T17:33:50 *** jamesrf has joined #evergreen 2010-01-28T17:35:28 *** dbs has quit IRC 2010-01-28T17:44:29 *** pmplett has joined #evergreen 2010-01-28T18:45:10 hey there 2010-01-28T18:45:17 is there any harm in disabling mod_proxy with evergreen? 2010-01-28T18:47:25 I'd expect for added content not to work 2010-01-28T18:49:02 hmm thanks 2010-01-28T18:57:46 *** _dkyle_ has quit IRC 2010-01-28T18:58:12 *** _dkyle_ has joined #evergreen 2010-01-28T19:42:48 *** jamesrf has quit IRC 2010-01-28T19:43:24 *** pmplett is now known as pmp_afk 2010-01-28T19:49:20 *** dbs has joined #evergreen 2010-01-28T20:05:51 *** pmp_afk is now known as pmplett 2010-01-28T20:17:57 *** pmplett has quit IRC 2010-01-28T20:19:11 *** pmplett has joined #evergreen 2010-01-28T20:58:37 *** DougF has joined #evergreen 2010-01-28T21:01:23 I've been playing aroudn with a clean install of Evergreen 1.6 on debian 5. I'm not very good at navigating the documentation yet, and I can seem to find a guide on "how to setup evergreen for you own use". Are there scripts that will delete the sampe data auto-imported during the install procedure? Is there a guide someone can point me to for doing this? 2010-01-28T21:02:36 are you comfortable with sql? You could rip stuff out that way. Otherwise, you can use GUI's through the staff client for most things, under Admin -> Server Settings 2010-01-28T21:05:50 i tried the gui.. but it seems pretty flaky or maybe i just dont understand what i see... most times when i click delete it put a * in front on the left panel... sometimes it jsut remves the item. Doesnt seem to be consistent 2010-01-28T21:06:40 I'm somewhat comfortable with sql.. but I thought there might be premade perl scripts for some of this 2010-01-28T21:07:35 it's pretty bare bones, the stock install. The big thing to change are the org units. The rest is more generic and less patron facing 2010-01-28T21:08:46 one thing about changing org units, is that you have to do a server-side command afterward to generate some static javascript encoding those org units 2010-01-28T21:09:01 that'd be /openils/bin/autogen.sh as user opensrf 2010-01-28T21:10:45 are there any known issues with the gui? I have gone in to the sql tables and verified when the 'delete" does work the the records are removed.. and I did figure the autogen out after re-readint eh install instuctions a coupel of dozen times! ;) 2010-01-28T21:12:08 I've seen viewport/size glitches were it looks like part of the lists are cut off, depending on the version of xulrunner 2010-01-28T21:12:46 there is an older interface for manipulating some of this, /cgi-bin/config.cgi You have to enable access to it in your apache config 2010-01-28T21:16:15 if you're willing, you could report bugs here: https://bugs.launchpad.net/evergreen 2010-01-28T21:16:52 im no sure i know enuf to be able to call smething a "bug". still just learning at this point 2010-01-28T21:18:28 th other thing i wasnt sure of.. is when u register a workstation. it looks as if the org unit come it to play with that too.. but i wasnt sure. 2010-01-28T21:19:23 yes, two things happening there. One, the workstation is associated with an org unit, and two, part of the org unit shortname is embedded into the workstation name 2010-01-28T21:20:04 I'd say "usability" bugs are fair game for pointing out 2010-01-28T21:21:30 are the referntial integrity contraints enforced in the database so I can't break the database by manually deleting things? Ie.. will it notify me if what im trying to do breaks the integrity 2010-01-28T21:21:48 we can haz referential constraints 2010-01-28T21:22:13 (i.e. yes, we've got RI up the wazoo) 2010-01-28T21:22:14 ok.. so if i change the org unit after i have a workstation registerd.. does it matter? or should I "unregidter" the workstation if that is even possible and rereg after the org changes are done 2010-01-28T21:22:51 reregister the workstation with a different OU by using a variation on the hostname (e.g. uppercase vs lowercase) 2010-01-28T21:23:03 or delete the ws_info file from the client 2010-01-28T21:23:11 or delete the workstation in the database table 2010-01-28T21:23:24 it'll force you to re-register 2010-01-28T21:23:43 all workarounds for the real usability bug 2010-01-28T21:23:50 * jeff grins 2010-01-28T21:24:25 or use -profilemanager in the target of your shortcut (or command-line invocation) 2010-01-28T21:25:07 * phasefx is tempted to have that be the default, but should really survey folks first 2010-01-28T21:25:17 different UI/UX "bugs" if you're a dev vs a sysadmin vs a user :) 2010-01-28T21:25:27 another gotcha to be aware of is that you have to have a work_ou assigned to any non-admin user that is registering the workstation (edit the permissions of the user in the view patron interface, or via SQL in permission.usr_work_ou_map) 2010-01-28T21:25:28 phasefx: having what be the default? 2010-01-28T21:26:04 jeff: -profilemanager with the stock shortcut for the windows executable 2010-01-28T21:26:05 i like that our users don't have to reg workstations. 2010-01-28T21:26:56 because we use per-machine rather than per-user staff client config 2010-01-28T21:27:25 ok.. that will give me enuf to play around with for a bit. i'll try as many option as i can to see what feel most comfotable to me. Thanks a bunch... and as a follow up to the other day.. i id get evergreen on debian 5 x64 to install ok if i gave the virtual machine at least 2gb ram. I was only using 1 gb before but i had thought the mach spc only said i needed 512mb 2010-01-28T21:28:04 DougF: the registered workstation library is where the client gets its sense of location, btw, rather than say, the home library of the logged in staff member 2010-01-28T21:28:18 DougF: used to be only 512MB, but that was with no GUI and before we added a bunch more services 2010-01-28T21:29:31 ok.. thanks... i'll play some more.. i'm sure i'll be back 2010-01-28T21:29:49 great! 2010-01-28T21:30:53 dbs: I'd be in favor of having the constraint on actor.workstation be a combination of name + owning_lib rather than just name, which lead to the current shortname behavior 2010-01-28T21:31:49 phasefx: makes sense to me, so you could have 'circdesk' at both BR1 and BR2 2010-01-28T21:32:07 not sure what folks using reporting would think. Could make a view to give them a combined column 2010-01-28T21:32:32 if they rebelled against adding in another column for lib :) 2010-01-28T21:33:44 I'm the only person in our library creating reports at the moment (gah), so no harm done to me! 2010-01-28T21:33:52 I know in PINES libraries sometime do change their "policy names". I wonder if they re-register all their workstations 2010-01-28T21:33:56 make staff client usernames be dot-sep, so br1.circdesk is the canonical name, an 'unqialified' "circdesk" gets the ws_ou prepended to fully qualify it before auth 2010-01-28T21:34:24 *** brendan_ga has quit IRC 2010-01-28T21:34:34 ours are sys-br-user for generic role accounts 2010-01-28T21:34:41 ah, I forgot about that wrinkle 2010-01-28T21:35:04 which wrinkle? 2010-01-28T21:35:06 could have the client start storing and passing the workstation id 2010-01-28T21:35:14 that it passes in the "name" as part of auth 2010-01-28T21:35:54 so we can't have duplicate names if we're using name as an auth argument 2010-01-28T21:36:29 oh, right. with your "could have 'circdesk' at both BR1 and BR2" i was thinking it was username and not workstation name you were talking about. 2010-01-28T21:37:24 I bet we're stuck with the current scheme. I don't relish the thought of bundling more "migration code" into the client 2010-01-28T21:37:47 heh. my first patch was fixing a bug in such migration code. 2010-01-28T21:37:52 :D 2010-01-28T21:40:19 still, i think the workstation name process could be improved especially from the deployment standpoint 2010-01-28T21:40:38 or maybe i just want to automate too many things 2010-01-28T21:41:05 jeff: did you see the firefox extension stuff I was laying with? 2010-01-28T21:41:13 playing with, even :D 2010-01-28T21:43:23 we could push workstation handling more to the server. You login first without a workstation, and it makes suggestions to you. The local client uses whatever you choose for segregating stuff on the file system 2010-01-28T21:44:53 say, the first time the client runs, it generates a UUID, and it passes that to the server. The server will check to see if any workstations are associated with that UUID, and offer them up if so, or, based on policy, maybe create a new workstation, or allocate a pre-created workstation to it 2010-01-28T21:45:21 and all the client ever does is send that UUID 2010-01-28T21:45:56 if this was all on a web stack, the UUID could be in a cookie 2010-01-28T21:46:21 and if a workstation needs to be "moved" to a newly imaged machine or something, easy as pie 2010-01-28T21:47:59 *** brendan_ga has joined #evergreen 2010-01-28T22:16:26 *** rsinger_ has joined #evergreen 2010-01-28T22:28:49 *** rsinger has quit IRC 2010-01-28T22:28:49 *** rsinger_ is now known as rsinger 2010-01-28T23:31:24 howdy phasefx