| # | Time | Nick | Message |
|---|
| # | 03:57:19 | JMCraig has joined #evergreen |
| # | 04:00:50 | agJohn has quit IRC |
| # | 04:01:00 | JMCraig is now known as agJohn |
| # | 04:09:37 | senator has quit IRC |
| # | 04:09:43 | senator has joined #evergreen |
| # | 05:54:35 | sabata has joined #evergreen |
| # | 05:54:49 | sabata | testing |
| # | 05:55:06 | sabata | hello |
| # | 06:00:42 | sabata | hello can someone help me |
| # | 06:08:01 | sabata | @sabata hello |
| # | 06:08:01 | pinesol` | sabata: Error: "sabata" is not a valid command. |
| # | 07:08:07 | sabata | hello |
| # | 07:08:23 | sabata | I am having problems with starting opensrf |
| # | 07:16:54 | alxp has joined #evergreen |
| # | 07:56:45 | sabata has quit IRC |
| # | 08:11:56 | sabata has joined #evergreen |
| # | 08:12:29 | sabata | hello anybody here |
| # | 08:12:54 | sabata | i need assitance with my opensrf server |
| # | 08:15:36 | granitize has joined #evergreen |
| # | 08:17:51 | sabata | hi granitize can you help me |
| # | 08:18:32 | granitize | sabata: heh I can try. |
| # | 08:18:47 | sabata | thank you |
| # | 08:19:14 | sabata | i am having problems with opensrf |
| # | 08:19:20 | sabata | i am extremely newbie |
| # | 08:19:31 | sabata | I did this http://www.open-ils.org/dokuwiki/doku.php?id=troubleshooting:checking_for_errors |
| # | 08:19:54 | sabata | until i got to step 8 where it talks about autogen |
| # | 08:20:19 | sabata | and the response I got what no response ... going to sleep |
| # | 08:20:58 | sabata | i had compiled a email with attachements and all but i did know how to use the mail list |
| # | 08:21:18 | sabata | with all the messages, but since then i think i broke it even worse |
| # | 08:22:04 | granitize | are you su - opensrf @ step 8 |
| # | 08:22:47 | sabata | no i think i was opensrf |
| # | 08:23:04 | sabata | or at least thats what i read |
| # | 08:23:07 | granitize | that's good. |
| # | 08:23:23 | granitize | I've not seen going to sleep. |
| # | 08:23:37 | tspindler has joined #evergreen |
| # | 08:23:55 | granitize | did you get the srfsh# prompt? |
| # | 08:23:57 | sabata | this is also the first time i am using IRC |
| # | 08:24:14 | granitize | you're doing fine! :-) |
| # | 08:24:42 | sabata | the first time i did the install i did the request math add 2 2 and it respond 4 |
| # | 08:24:47 | granitize | sabata: sorry - That's step 9 |
| # | 08:25:05 | sabata | after installing evergreen it never worked again |
| # | 08:25:35 | sabata | oh |
| # | 08:25:39 | sabata | no i never got the step 9 |
| # | 08:25:49 | granitize | there's also a perl script Open-ILS/src/support-scripts/settings-tester.pl |
| # | 08:25:55 | sabata | it showed ... Updating fieldmapper then said going to sleep |
| # | 08:26:21 | sabata | is that the syntax to run |
| # | 08:27:02 | granitize | It's the path in the uncompressed download. |
| # | 08:29:24 | sabata | found it |
| # | 08:29:37 | sabata | should i run it with the full path name |
| # | 08:29:59 | sabata | as which user root or opensrf |
| # | 08:31:03 | sabata | ok it ran |
| # | 08:31:12 | sabata | what do i do with the output |
| # | 08:31:24 | granitize | Read it and see if there are abny errors. |
| # | 08:31:31 | granitize | any errors |
| # | 08:32:51 | sabata | jabber x4 connected, database x5 succesfully, database drivers x6 OK, and another OK OK |
| # | 08:32:59 | sabata | no errors |
| # | 08:34:44 | granitize | any errors in the logs? |
| # | 08:34:53 | granitize | I'm having troubles too. :-) |
| # | 08:35:14 | granitize | Ican get to a record but can't search when using a non-local database. |
| # | 08:36:08 | sabata | shoo |
| # | 08:36:23 | sabata | that sound far ahead in the future for me |
| # | 08:36:35 | granitize | heh... |
| # | 08:37:01 | granitize | When you get it working it feels great! |
| # | 08:37:40 | sabata | i backedup my logs be4 chatting, now i only got 2 logs public.localhost and its the friend private and only warnings now errors |
| # | 08:40:03 | sabata | um trying the troubleshoot thing again, becos for me its almost home time now ZA |
| # | 08:44:17 | sabata | again the same error |
| # | 08:44:20 | sabata | opensrf@ubuntu:/openils/var/log$ /openils/bin/autogen.sh -c /openils/conf/opensrf_core.xml -u Updating Evergreen organization tree and IDL using '/openils/conf/opensrf_core.xml' Updating fieldmapper No Response from settings server...going to sleep |
| # | 08:44:40 | sabata | that is onscreen |
| # | 08:46:27 | granitize | at 8? |
| # | 08:47:25 | sabata | yes @ 8 |
| # | 08:48:01 | sabata | i mean its just 1 more step and I am home free |
| # | 08:48:16 | sabata | why wont it just let be pass by |
| # | 08:48:18 | sabata | lol |
| # | 08:49:32 | granitize | is opensrf running? |
| # | 08:51:48 | tspindler has quit IRC |
| # | 08:51:48 | sabata has quit IRC |
| # | 08:51:49 | dbs has quit IRC |
| # | 08:51:50 | granitize has quit IRC |
| # | 08:51:52 | tildeequals has quit IRC |
| # | 08:51:55 | dbwells has quit IRC |
| # | 08:51:55 | jeff has quit IRC |
| # | 08:51:55 | denials_ has quit IRC |
| # | 08:51:57 | shadowspar has quit IRC |
| # | 08:51:57 | phasefx has quit IRC |
| # | 08:51:57 | bradl has quit IRC |
| # | 08:51:57 | senator has quit IRC |
| # | 08:51:58 | atz has quit IRC |
| # | 08:51:59 | mjgiarlo has quit IRC |
| # | 08:52:00 | leed has quit IRC |
| # | 08:52:00 | jeffdavis has quit IRC |
| # | 08:52:01 | moodaepo has quit IRC |
| # | 08:52:01 | cbandito has quit IRC |
| # | 08:52:03 | lisppaste has quit IRC |
| # | 08:52:03 | eby has quit IRC |
| # | 08:52:03 | brendan2 has quit IRC |
| # | 08:52:03 | sylvar has quit IRC |
| # | 08:52:05 | tsbere has quit IRC |
| # | 08:52:06 | artunit has quit IRC |
| # | 08:52:07 | gmcharlt has quit IRC |
| # | 08:52:07 | phasefx_ has quit IRC |
| # | 08:52:07 | berick has quit IRC |
| # | 08:52:08 | kbeswick has quit IRC |
| # | 08:52:09 | wjr has quit IRC |
| # | 08:52:10 | mtate has quit IRC |
| # | 08:52:10 | Dmagick has quit IRC |
| # | 08:52:11 | eeevil has quit IRC |
| # | 08:52:13 | csharp has quit IRC |
| # | 08:53:06 | tspindler has joined #evergreen |
| # | 08:53:06 | granitize has joined #evergreen |
| # | 08:53:06 | sabata has joined #evergreen |
| # | 08:53:06 | senator has joined #evergreen |
| # | 08:53:06 | dbs has joined #evergreen |
| # | 08:53:06 | kbeswick has joined #evergreen |
| # | 08:53:06 | lisppaste has joined #evergreen |
| # | 08:53:06 | moodaepo has joined #evergreen |
| # | 08:53:06 | atz has joined #evergreen |
| # | 08:53:06 | mjgiarlo has joined #evergreen |
| # | 08:53:06 | shadowspar has joined #evergreen |
| # | 08:53:06 | tildeequals has joined #evergreen |
| # | 08:53:06 | leed has joined #evergreen |
| # | 08:53:06 | artunit has joined #evergreen |
| # | 08:53:06 | eby has joined #evergreen |
| # | 08:53:06 | mtate has joined #evergreen |
| # | 08:53:06 | jeffdavis has joined #evergreen |
| # | 08:53:06 | Dmagick has joined #evergreen |
| # | 08:53:06 | phasefx has joined #evergreen |
| # | 08:53:06 | cbandito has joined #evergreen |
| # | 08:53:06 | dbwells has joined #evergreen |
| # | 08:53:06 | brendan2 has joined #evergreen |
| # | 08:53:06 | jeff has joined #evergreen |
| # | 08:53:06 | sylvar has joined #evergreen |
| # | 08:53:06 | tsbere has joined #evergreen |
| # | 08:53:06 | denials_ has joined #evergreen |
| # | 08:53:06 | wjr has joined #evergreen |
| # | 08:53:06 | gmcharlt has joined #evergreen |
| # | 08:53:06 | bradl has joined #evergreen |
| # | 08:53:06 | eeevil has joined #evergreen |
| # | 08:53:06 | berick has joined #evergreen |
| # | 08:53:06 | phasefx_ has joined #evergreen |
| # | 08:53:06 | csharp has joined #evergreen |
| # | 08:53:28 | granitize | as opensrf /openils/bin/osrf_ctl.sh -c /openils/conf/opensrf_core.xml -l -a stop_c |
| # | 08:53:45 | granitize | then /openils/bin/osrf_ctl.sh -c /openils/conf/opensrf_core.xml -l -a start_c |
| # | 08:53:58 | granitize | you get this right? Starting OpenSRF C (host=localhost) |
| # | 08:55:13 | sabata | yes it ran |
| # | 08:55:34 | granitize | any errors ? grep ERR /openils/var/log/*.log |
| # | 08:56:04 | sabata | opensrf@ubuntu:/openils/var/log$ grep ERR /openils/var/log/*.log /openils/var/log/osrfsys.log:opensrf 2010-11-26 15:54:07 [ERR :2257:osrf_system.c:401:] Child process 2258 (app opensrf.math) killed by signal 15 |
| # | 08:58:54 | dbs | sabata: the "killed by signal 15" is probably when you issed stop_c |
| # | 08:59:14 | sabata | ok |
| # | 08:59:51 | dbs | what do you see when you type: |
| # | 08:59:53 | dbs | perl -MNet::Domain -e 'print Net::Domain::hostfqdn() . "\n";' |
| # | 09:00:11 | granitize | big guns in now! :) |
| # | 09:00:26 | sabata | really i thought that was you |
| # | 09:00:32 | granitize | snort |
| # | 09:00:34 | sabata | i see the name now hello dbs |
| # | 09:00:36 | sabata | perl -MNet::Domain -e 'print Net::Domain::hostfqdn() . "\n";' |
| # | 09:00:42 | sabata | ubuntu.internal.dwaf.gov.za |
| # | 09:00:45 | sabata | sorry |
| # | 09:01:27 | dbs | okay. can you do "osrf_ctl.sh -l -a stop_all" |
| # | 09:01:42 | dbs | then "ps wax | grep -i open" to see if any opensrf processes are still running? |
| # | 09:02:12 | sabata | opensrf@ubuntu:/openils/var/log$ ps wax | grep -i open 1337 ? Ssl 0:00 /usr/lib/gvfs//gvfs-fuse-daemon /home/opensrf/.gvfs 2451 pts/0 S+ 0:00 grep --color=auto -i open |
| # | 09:02:23 | dbs | looks good |
| # | 09:02:35 | sabata | shoo u r a big gun |
| # | 09:02:38 | sabata | lol |
| # | 09:02:48 | dbs | now clear the logs: "rm -f /openils/var/log/*log" |
| # | 09:03:54 | dbs | and start up again: "osrf_ctl.sh -l -a start_router; sleep 5; osrf_ctl.sh -l -a start_perl; sleep 10; osrf_ctl.sh -l -a start_c" |
| # | 09:03:54 | sabata | done |
| # | 09:04:09 | dbs | The "sleep" commands shouldn't be necessary, but oh well, can't hurt |
| # | 09:04:45 | dbs | once you finish that, "ps wax | grep -i open | wc -l" |
| # | 09:04:55 | dbs | (to count the number of OpenSRF processes) |
| # | 09:05:30 | sabata | 101 |
| # | 09:05:44 | dbs | wow! sounds like you have a lot running :) |
| # | 09:06:10 | sabata | is that good or bad |
| # | 09:06:20 | dbs | I have 102 right now, on a development server, so that's good |
| # | 09:06:38 | sabata | so i am one short .. jokes |
| # | 09:06:40 | dbs | let's try logging in: |
| # | 09:06:49 | sabata | ok |
| # | 09:06:59 | dbs | srfsh |
| # | 09:07:11 | sabata | yes |
| # | 09:07:12 | dbs | then inside srfsh, "login admin open-ils" |
| # | 09:07:13 | sabata | i am in |
| # | 09:07:55 | sabata | ko what do you need on this screen i should give u i see succesful x2 |
| # | 09:08:13 | sabata | Name: osrfMethodException Status: No authentication seed found. open-ils.auth.authenticate.init must be called first Status: 404 |
| # | 09:08:15 | dbs | can you paste the results to http://paste.lisp.org/new/evergreen ? |
| # | 09:08:24 | glen_ has joined #evergreen |
| # | 09:08:30 | dbs | ah, sounds like it could be memcached |
| # | 09:08:48 | dbs | ps wax | grep memcached |
| # | 09:08:53 | sabata | ok should i still paste |
| # | 09:09:05 | dbs | sure, can't hurt |
| # | 09:09:13 | sabata | opensrf@ubuntu:/openils/var/log$ ps wax | grep memcached 1134 ? Sl 0:00 /usr/bin/memcached -m 64 -p 11211 -u memcache -l 127.0.0.1 |
| # | 09:10:06 | lisppaste | Sabata pasted "dbs request a paste" at http://paste.lisp.org/display/117046 |
| # | 09:10:28 | sabata | oh wow |
| # | 09:10:29 | dbs | sabata: then annotate that paste with the results of "grep -C 10 memcache /openils/conf/opensrf.xml" |
| # | 09:10:35 | sabata | it gt here before me |
| # | 09:10:59 | dbs | yeah, lisppaste is neat :) |
| # | 09:11:25 | lisppaste | Sabata annotated #117046 "annotated paste" at http://paste.lisp.org/display/117046#1 |
| # | 09:13:26 | dbs | huh. looks okay. You could try stopping all of the opensrf services, changing that "localhost:11211" to "127.0.0.1:11211" in opensrf.xml, and then starting the opensrf services again - i have (rarely) seen that be an issue |
| # | 09:14:12 | dbs | Before doing that though, perhaps stop the opensrf services, then restart memcached ("/etc/init.d/memcached restart") and start up the opensrf services again |
| # | 09:14:46 | dbs | (and I'm not a big gun, I've just been through umpteen installations before and have bumped into most of the bumps!) |
| # | 09:16:44 | sabata | yoh |
| # | 09:16:49 | sabata | ok um done |
| # | 09:17:06 | sabata | ok u not a big gun |
| # | 09:17:10 | sabata | lol |
| # | 09:18:04 | sabata | and ps wax | grep -i open | wc -l says 101 |
| # | 09:19:04 | lisppaste | Sabata annotated #117046 "another annotate" at http://paste.lisp.org/display/117046#2 |
| # | 09:19:31 | dbs | looks good! |
| # | 09:19:34 | _dkyle_ has joined #evergreen |
| # | 09:19:36 | sabata | yay |
| # | 09:19:39 | dbs | now try "autogen.sh -u" |
| # | 09:20:14 | sabata | opensrf@ubuntu:/openils/conf$ autogen.sh -u Updating Evergreen organization tree and IDL using '/openils/conf/opensrf_core.xml' Updating fieldmapper No Response from settings server...going to sleep |
| # | 09:21:16 | dbs | booooo |
| # | 09:21:32 | sabata | yeah big boooooooo |
| # | 09:21:58 | dbs | can you try the srfsh login again? |
| # | 09:22:21 | sabata | no prompt |
| # | 09:22:45 | sabata | its hanging/sleeping after that |
| # | 09:23:00 | sabata | ok did ctrl+c |
| # | 09:23:23 | dbs | hrm |
| # | 09:23:24 | sabata | ok i have logged in |
| # | 09:23:25 | dbs | okay. |
| # | 09:23:31 | dbs | one more thing to try. |
| # | 09:23:36 | sabata | yeah |
| # | 09:23:37 | dbs | shut down the opensrf services |
| # | 09:24:02 | sabata | like this osrf_ctl.sh -l -a stop_all |
| # | 09:24:04 | dbs | we've seen a few strange network routing cases where requests seem to take forever |
| # | 09:24:06 | dbs | (yep) |
| # | 09:24:30 | sabata | did |
| # | 09:26:35 | dbs | then open /usr/local/share/perl5/OpenSRF/AppSession.pm in an editor |
| # | 09:27:39 | sabata | it says new file in the bottom of nano |
| # | 09:27:53 | dbs | hrm, that's not good |
| # | 09:27:58 | sabata | GNU nano 2.2.4 File: /usr/local/share/perl5/OpenSRF/AppSession.pm |
| # | 09:28:06 | dbs | maybe your distro installs opensrf in a different location than mine |
| # | 09:28:35 | dbs | type "find /usr/local -name "AppSession.pm" |
| # | 09:29:29 | sabata | "/usr/local/share/perl/5.10.1/OpenSRF/AppSession.pm" |
| # | 09:29:53 | sabata | its open |
| # | 09:30:29 | dbs | around line 836 there should be a "recv_timeout => 30"; change the "30" to a "300" |
| # | 09:30:50 | dbs | (I think that's the right file and line - there are an awful lot of "timeout" values hardcoded in OpenSRF) |
| # | 09:31:16 | sabata | recv_timeout => 30, |
| # | 09:31:38 | dbs | http://www.open-ils.org/irc_logs/evergreen/2010-09/%23evergreen.29-Wed-2010.log suggests we should actually be editing /usr/local/share/perl/5.10.1/OpenSRF/Utils/SettingsClient.pm |
| # | 09:32:23 | dbs | so let's close this one up, and edit /usr/local/share/perl/5.10.1/OpenSRF/Utils/SettingsClient.pm instead |
| # | 09:32:55 | dbs | lines 84 and 90 have "timeout => 10" values, up those from 10 to 120 |
| # | 09:33:14 | dbs | then start up opensrf again and try "autogen.sh -u" one more time |
| # | 09:35:30 | sabata | heyyy its moved |
| # | 09:35:46 | sabata | oh damn |
| # | 09:36:08 | sabata | excitement killed it i did ctrl-c to copy |
| # | 09:36:14 | dbs | hah |
| # | 09:36:16 | sabata | redid it again |
| # | 09:36:49 | lisppaste | Sabata annotated #117046 "rerun" at http://paste.lisp.org/display/117046#3 |
| # | 09:37:50 | dbs | I'm wondering whether we should just go ahead and increase that timeout by default in OpenSRF, since we've seen the same problem a few times. There's clearly some weird underlying network problem, but we haven't been able to figure out what it is yet. |
| # | 09:38:05 | dbs | Hmm. Some progress, at least |
| # | 09:38:22 | dbs | might be time to check the logs now to see if there's more information |
| # | 09:39:07 | sabata | damn lost command |
| # | 09:40:06 | lisppaste | Sabata annotated #117046 "error paste" at http://paste.lisp.org/display/117046#4 |
| # | 09:42:28 | dbs | yeah, those don't appear to be recent errors |
| # | 09:43:48 | sabata | i am seeing the key errors only now |
| # | 09:44:05 | sabata | i only had 2 errors b4 joining us |
| # | 09:44:23 | sabata | hey where is granitize my buddy |
| # | 09:44:32 | sabata | lol |
| # | 09:46:52 | dbs | do you have a /openils/var/web/opac/common/js/en-US directory? |
| # | 09:47:09 | sabata | hey i just looking at cp: cannot stat `/openils/var/web/opac/common/js//en-US/OrgTree.js': No such file or directory |
| # | 09:47:12 | sabata | lol |
| # | 09:47:24 | sabata | no folder like that |
| # | 09:47:33 | sabata | only en-GB |
| # | 09:47:52 | sabata | i think its the time i was installing language options |
| # | 09:48:01 | sabata | i opted for GB |
| # | 09:48:19 | sabata | how or where can i change that reference |
| # | 09:48:29 | dbs | ahhhhhh |
| # | 09:48:46 | dbs | there's a hard-coded reference to en-US in autogen.sh |
| # | 09:49:03 | dbs | there's an assumption that en-US is always installed |
| # | 09:49:18 | sabata | i am glad i could help |
| # | 09:49:33 | granitize | sabata: Watching it all |
| # | 09:49:33 | sabata | where is autogen by the way |
| # | 09:49:44 | dbs | /openils/bin/autogen.sh |
| # | 09:49:45 | sabata | with pop corn and all |
| # | 09:50:03 | dbs | I don't even know how you would get rid of en-US! |
| # | 09:50:14 | granitize | Nice digging dbs and sabata… language! who knew. |
| # | 09:51:23 | sabata | thank you |
| # | 09:51:48 | sabata | i notice they have 2 forward slashes `/openils/var/web/opac/common/js//en-US/OrgTree.js' |
| # | 09:51:50 | dbs | you would have had to remove "en-US" from the config.i18n_locale in the database |
| # | 09:52:02 | dbs | yeah, the double forward slashes don't make a difference |
| # | 09:52:20 | sabata | lol i actaully deleted everything else when i installed GB |
| # | 09:52:21 | dbwells has quit IRC |
| # | 09:52:38 | dbs | hah |
| # | 09:53:19 | dbs | INSERT INTO config.i18n_locale (code, marc_code, name, description) VALUES ('en-US', 'eng', 'English (US)', 'American English'); -- ought to do the trick |
| # | 09:53:31 | dbs | (from psql, naturally) |
| # | 09:53:40 | sabata | so after $JSDIR should i delete the extra slash or not "cp "$JSDIR/en-GB/OrgTree.js" "$FMDOJODIR/"" |
| # | 09:53:44 | dbs | now I have to go back to my real work |
| # | 09:54:34 | dbs | I honestly don't know if anything will work without en-US in place - we're beyond normal support scenarios and into experimental mode now :) |
| # | 09:55:10 | sabata | pgAdmin is lifesaver |
| # | 09:55:14 | sabata | done |
| # | 09:56:03 | sabata | i am nearly out of here too |
| # | 09:56:04 | dbs | try "autogen.sh -u" again |
| # | 09:56:14 | sabata | knocking off shortly |
| # | 09:56:37 | glen_ has quit IRC |
| # | 09:56:51 | sabata | damn restarted opensrf |
| # | 09:56:59 | sabata | now running authgen |
| # | 09:57:13 | sabata | busy... |
| # | 09:58:33 | lisppaste | Sabata annotated #117046 "retry" at http://paste.lisp.org/display/117046#5 |
| # | 10:07:15 | dbwells has joined #evergreen |
| # | 10:11:50 | dbs has quit IRC |
| # | 10:15:14 | granitize | Anyone have any ideas why I can get to a record with the ID# but searches fail whilst using a non-local db? |
| # | 10:15:42 | granitize | jeffdavis: I need to use the -l because all that is non-local is postgres. |
| # | 10:18:02 | atz_ has joined #evergreen |
| # | 10:20:56 | atz has quit IRC |
| # | 10:21:28 | sabata | hey granitize um outta here |
| # | 10:22:00 | sabata | thank you kindly for everyones contribution |
| # | 10:22:47 | sabata has quit IRC |
| # | 10:28:55 | agJohn | granitize: My favorite way to get that behavior is to get distracted right when I should be running the meta-data portion of the script. |
| # | 10:29:38 | agJohn | (Or to be more specific, I process the records into SQL, add 'em, and forget to add the meta-data.) |
| # | 11:19:23 | granitize | agjohn: huh? |
| # | 11:20:10 | granitize | I migrated a live db to a non=local postgres implementation and configured the evergreen server to use that instance. |
| # | 11:20:34 | granitize | Can you be more specific? |
| # | 11:20:55 | agJohn | Ah, sorry. So everything was working when it was local? |
| # | 11:21:19 | granitize | the "local" db was a different db. :-) |
| # | 11:21:30 | granitize | but yes, |
| # | 11:21:45 | agJohn | OK. I'm not sure I understand what you mean by "migrated".... |
| # | 11:22:05 | granitize | the new db is a copy of my live db that does work |
| # | 11:22:14 | agJohn | Gotcha. |
| # | 11:22:26 | agJohn | Hmmm. |
| # | 11:22:52 | granitize | b/c I am testing non-local db as my live is running out of dic=sk space.. poor planning! |
| # | 11:23:07 | granitize | yes… hmmm |
| # | 11:23:26 | granitize | Where are the memory requirments? |
| # | 11:23:43 | granitize | on the db server or on the apache/opsnils box. |
| # | 11:23:55 | agJohn | Right. Well, I always run Evg on a separate box/VM from Postgres so I'm guessing we can figure it out. |
| # | 11:24:43 | agJohn | Memory requirements are highly dependent on the load on the Evg box; on the Postgres box, on the size of the DB. |
| # | 11:24:58 | granitize | nod |
| # | 11:25:20 | agJohn | But let's back up; you say you can do an id# search. What's the program function you're using for this? |
| # | 11:25:36 | granitize | 48GB on the evergreen box and 3GB on the postgres server… 800,000 records and additional indexes. |
| # | 11:26:11 | granitize | catalog. |
| # | 11:26:21 | agJohn | Your Pg server's memory-starved, I would guess. It may be that it's trying to do the search and just not finishing. |
| # | 11:27:05 | agJohn | However, let's see what we can figure out. What program function works and what does not? |
| # | 11:28:23 | granitize | http://137.149.200.12/opac/en-CA/skin/default/xml/rdetail.xml?r=118770 |
| # | 11:28:50 | agJohn | OK. And a keyword search (of any kind) does not work, right? |
| # | 11:28:57 | granitize | correct. |
| # | 11:29:07 | granitize | Shelf browser works |
| # | 11:29:08 | agJohn | How'd you transfer the data from the live environment to the Pg test setup? |
| # | 11:29:26 | granitize | Using a backup and the WAL files. |
| # | 11:29:42 | granitize | I didn't do it actually - Had a db admin do it. |
| # | 11:29:58 | granitize | pgadmin connects fine. |
| # | 11:30:16 | granitize | book covers don't work. |
| # | 11:30:55 | granitize | This URL has a book cover live. |
| # | 11:30:55 | granitize | http://137.149.200.12/opac/en-CA/skin/default/xml/rdetail.xml?r=792270 |
| # | 11:31:24 | agJohn | (One problem at a time, eh?) I've never tried that so I don't know if that'll work. What I do know works is to create a DB on the target server (using template0--very important), make a backup of the DB using pg_dump (I prefer the "custom" format) and restore using pg_restore. |
| # | 11:32:06 | agJohn | But, if you backed up all the DB files, I would guess that would work. My guess is that the test environment is missing the tsearch stuff. |
| # | 11:32:47 | granitize | Sorry… you asked what was working and not… :-) |
| # | 11:32:54 | granitize | What is tsearch? |
| # | 11:33:29 | gmcharlt | tsearch is the full text search engine used by Postgres |
| # | 11:33:52 | granitize | I see it… I'll check. |
| # | 11:34:12 | granitize | How can I check? |
| # | 11:34:33 | agJohn | Right; sorry for the confusion. I meant w/ regard to the issues you'd raised (in terms of what functions in the software). |
| # | 11:34:35 | agJohn | RE tsearch and related stuff, here's some specific steps in the install instructions that you need to ensure the DB admin person did: |
| # | 11:34:36 | agJohn | http://open-ils.org/dokuwiki/doku.php?id=server:1.6.1:install (Step III) |
| # | 11:34:38 | agJohn | more coming.... |
| # | 11:35:29 | agJohn | Sorry, that's 2. III (odd outlining rules in play here). |
| # | 11:36:15 | granitize | sorry… crisis… I'll be back. |
| # | 11:36:21 | agJohn | Also, these steps from 3.I. (use your DB name in place of evergreen): |
| # | 11:36:24 | agJohn | psql -f /usr/share/postgresql/<pgsql_version>/contrib/tablefunc.sql evergreen |
| # | 11:36:26 | agJohn | psql -f /usr/share/postgresql/<pgsql_version>/contrib/tsearch2.sql evergreen |
| # | 11:36:28 | agJohn | psql -f /usr/share/postgresql/<pgsql_version>/contrib/pgxml.sql evergreen |
| # | 11:42:01 | agJohn | granitize: One simple check you can do is to run this query (via psql or pgAdmin): |
| # | 11:42:02 | agJohn | select tsearch2() |
| # | 11:42:04 | agJohn | You'll get an error message, but the content of the error message is critical, if it includes this: |
| # | 11:42:05 | agJohn | function tsearch2() does not exist |
| # | 11:42:07 | agJohn | then you know your Postgres install is missing critical elements--this seems a bit unlikely given your migration approach. |
| # | 11:43:17 | agJohn | You can also do a test keyword query. Using the pattern in the paste below: |
| # | 11:44:45 | lisppaste | agJohn pasted "Test of keyword search (for dogs)" at http://paste.lisp.org/display/117056 |
| # | 11:45:28 | agJohn | If you get errors from that SQL expression of a keyword search for "dogs", then you've got some kind of setup issue. |
| # | 11:47:59 | agJohn | Now, the other thing that's quite possible is that it's trying to do the search, but it's taking so long that the opac-response-assembly code simply times out and gives up (you'll just see the spinner running on the web browser). |
| # | 11:48:01 | agJohn | So, if the above does no return fairly quickly, you can then go to an ssh or telnet session on the Pg server and run the "top" command. You should see (high on the list) a postgres process (may show as postmaster). It'll be using a significant fraction of your cpu%. |
| # | 11:55:08 | agJohn | If that's the case, then you have a config/total RAM problem. If the query does eventually come back, you can try heating up the cache on your Pg box and see if it'll get to a decent performance level. Use the scripts in the following paste. Watch the progress of the cache-heating script via top--within a matter of a minute or two, it should get to the point that the postgres process is... |
| # | 11:55:09 | agJohn | ...using very nearly 100% of the CPU. If it does not, then you've got too little RAM--or, at minimum, you need to adjust the configuration of Postgres to use more shared buffers (if you contact me in a private session, I could help you with that--or send me an email via alphagconsulting.com). |
| # | 11:55:44 | lisppaste | agJohn pasted "Keyword Cache Heater script" at http://paste.lisp.org/display/117058 |
| # | 12:06:48 | granitize | agjohn: back and reading your suggestions… Thanks |
| # | 12:11:19 | granitize | select tsearch2() returns ERROR: could not find tsearch config by locale |
| # | 12:11:20 | granitize | So... |
| # | 12:11:31 | granitize | The steps from 3.1 above will enable this? |
| # | 12:21:05 | granitize | relation "pg_ts_dict" already exists |
| # | 12:23:56 | granitize | pgxml.sql not on server. |
| # | 12:31:37 | agJohn | Sounds like you have a somewhat mixed bag. When you get to the point that you're getting an error more like: |
| # | 12:31:38 | agJohn | tsvector_update_trigger: not fired by trigger manager; |
| # | 12:31:40 | agJohn | Go ahead and try to keyword search SQL (pasted abvoe as .../117056) |
| # | 12:32:57 | agJohn | granitize: (Above comment was kind of jumping the gun.) To get the pgxml.sql scripts onto the server, you'll need to add the contrib package for Postgres |
| # | 12:33:59 | agJohn | That's going to be in the script that installs Postgres. If you need help figuring it out, let us know what OS you're on. |
| # | 12:41:32 | dbs has joined #evergreen |
| # | 12:41:32 | dbs has joined #evergreen |
| # | 12:42:36 | granitize | running on Debian.Debian |
| # | 12:43:48 | granitize | I'll look at the script that installs postgres. |
| # | 12:48:36 | granitize | I think it was just a base install - So I have some work to do. |
| # | 12:48:37 | granitize | Thanks |
| # | 12:51:02 | granitize | Sorry… postgres server is running SUSE Linux Enterprise Server 10 (x86_64) |
| # | 12:53:18 | granitize | Can SUSE support what's needed? |
| # | 13:01:14 | dbs | granitize: it should, postgresql for evergreen doesn't need much |
| # | 13:01:30 | agJohn | granitize: Not familiar with how to set things up on SUSE, but I don't see why it couldn't work--in principle. Does SUSE use RPM? If I'm not misremembering that, then look at the extra packages that get installed for RedHat or CentOS in the install scripts. |
| # | 13:01:43 | dbs | you probably just need the additional contrib packages and the MARC::Record, MARC::File::XML, and JSON perl packages I pointed out last night |
| # | 13:02:12 | granitize | What about the missing pgxml.sq |
| # | 13:02:13 | agJohn | Or, you might just go to the Postgres web page and look for what it says about SUSE and the contrib package. |
| # | 13:02:43 | agJohn | pgxml.sql should be part of the contrib package(s). |
| # | 13:04:02 | agJohn | (What dbs said....) |
| # | 13:05:01 | granitize | I loaded those. and restarted I thought. |
| # | 13:05:06 | dbs | SLES and OS X, nothing like being off the beaten path :) |
| # | 13:05:13 | granitize | hahaha |
| # | 13:05:22 | granitize | ya… want to join the dark side? |
| # | 13:05:29 | granitize | we have beer and cookies. |
| # | 13:07:42 | granitize | I don't don't know what you mean by contrib packages… sorry |
| # | 13:10:16 | agJohn | The Postgres install consists of the basic package & some "contributed" to the project, but not part of the base system (e.g. postgresql.8.4.<blah, blah>.rpm and one or more packages with name(s) like postrgresql.contrib.8.4...rpm)--you're missing the pgxml.sql script because you do not have the pgxml contrib package installed). You'll need to be able to run that script. So, you need the... |
| # | 13:10:18 | agJohn | ...contrib package that includes it. |
| # | 13:12:06 | agJohn | I don't recall if this is part of a particular package--and I don't remember for sure if SUSE uses RPM-type packages, but in either case, your DB admin person didn't install everything you need. If you want to look through the install scripts for postgres, you may be able to figure out what packages are missing (assuming some substantial familiarity with the package-manager of SUSE). |
| # | 13:12:30 | agJohn | Well, that was dumb--yes, it's part of a particular package--I don't remember WHICH package. |
| # | 13:14:30 | agJohn | What dbs said about it likely being workable on SUSE is right--the Posgres install for Evergreen is not nearly so complex as the prerequisites for the Evergreen side of things; I've run against Postgres on RedHat and CentOS with a Debian or Ubuntu Evergreen server. So, I'm moderately confident you can make it work, but you'll need to know how to load packages on SUSE. |
| # | 13:16:07 | agJohn | (And I wish I could tell you how to do that, but I don't know myself.... But, you might find SUSE packages and info about loading them on the Postgres site (or one it links to).) |
| # | 13:16:12 | dbs | SUSE is RPM, btw. Used to test and develop against it ages ago |
| # | 13:18:00 | agJohn | So, okay. Look through the Pg install scripts that are referenced in the Evg install instructions and you (your DBA) should be able to figure out what packages are needed; verify they're all there, add missing ones, and you should be able to run the missing script. |
| # | 13:18:26 | agJohn | (Once you have all the contrib packages that are needed.) |
| # | 13:20:11 | granitize | Right on. |
| # | 13:20:29 | dbs | granitize: Perhaps: find /usr -name 'pgxml.sql' |
| # | 13:21:19 | dbs | it's in /usr/share/pgsql/contrib on fedora, not that that helps |
| # | 13:24:20 | granitize | not on the SUSE box. |
| # | 13:24:35 | granitize | :( |
| # | 13:26:47 | dbs has quit IRC |
| # | 13:29:13 | agJohn | Another option which may be simpler is you could try just running the pgxml.sql file from your working setup--the problem is, if there's code that goes with it, you'll need the rest of the package for it to work. But you could try running it and seeing if you get errors. |
| # | 13:29:55 | agJohn | That is, just copy the file over from your working setup to the test Pg server and try running it. If you get errors, then you'll have to run down the package that includes it and install the rpm file for it. |
| # | 13:30:17 | granitize | Hunting for the packages in YAST now. |
| # | 13:31:11 | agJohn | (Our man dbs would probably know if there's more than just the SQL script involved, but he's just "left" it would seem.) |
| # | 13:39:20 | granitize | An additional complication it seems is that the OS is actually SLEZ with an expired activation key… more on the soap opera later! Thanks for your input… all good. |
| # | 14:29:58 | tspindler has quit IRC |
| # | 14:44:17 | granitize has left #evergreen |
| # | 14:58:37 | alxp has quit IRC |
| # | 15:28:35 | atz_ has quit IRC |
| # | 15:29:04 | atz_ has joined #evergreen |
| # | 15:46:00 | tildeequals has quit IRC |
| # | 16:59:05 | atz__ has joined #evergreen |
| # | 17:01:19 | atz_ has quit IRC |
| # | 17:31:38 | atz__ has quit IRC |
| # | 17:32:07 | atz__ has joined #evergreen |
| # | 18:15:14 | tildeequals has joined #evergreen |
| # | 18:20:06 | tildeequals_ has joined #evergreen |
| # | 18:21:56 | tildeequals has quit IRC |
| # | 18:29:11 | tildeequals_ has quit IRC |
| # | 18:46:20 | agJohn | Anyone know anything about this 1.6.1.4 Marc Editor error (just upgraded from 1.6.0.6): |
| # | 18:46:22 | agJohn | FIXME, MARC Editor, my_init: SyntaxError: missing = in XML attribute |
| # | 18:52:04 | agJohn | The XML of the record is fine; just checked it w/ an external validator.... Shoot, the upgrade was going so well.... |
| # | 18:55:51 | atz has joined #evergreen |
| # | 18:57:35 | atz__ has quit IRC |
| # | 19:47:07 | atz | agJohn: you validated against XML or MARCXML? I'm guessing the former. |
| # | 22:14:34 | atz_ has joined #evergreen |
| # | 22:16:50 | atz has quit IRC |
April 10-13