| # | Time | Nick | Message |
|---|
| # | 00:40:19 | tildeequals has quit IRC |
| # | 01:20:32 | jamesrf has quit IRC |
| # | 02:01:10 | dbs has quit IRC |
| # | 02:01:25 | dbs has joined #evergreen |
| # | 02:06:38 | atz_ has quit IRC |
| # | 02:07:12 | atz_ has joined #evergreen |
| # | 07:28:11 | sfortin has joined #evergreen |
| # | 08:09:40 | artunit has quit IRC |
| # | 08:09:53 | artunit has joined #evergreen |
| # | 08:48:41 | rsinger has quit IRC |
| # | 09:04:04 | Meliss has joined #evergreen |
| # | 09:06:18 | eeevil | call for testing: http://svn.open-ils.org/trac/ILS/changeset/18901 on any version (note the bug link) ... TIA! |
| # | 09:09:11 | rsinger has joined #evergreen |
| # | 09:23:08 | kmlussier has joined #evergreen |
| # | 09:32:50 | jenny has joined #evergreen |
| # | 09:32:54 | cbandito has joined #evergreen |
| # | 09:51:50 | mrpeters-isl | eeevil: applying it to our migration server now and i will check it out for you |
| # | 09:52:26 | eeevil | thanks! |
| # | 09:54:13 | mrpeters-isl | done! |
| # | 09:54:26 | mrpeters-isl | we get no results (maybe we don't have any items by madonna! lol) |
| # | 09:54:44 | mrpeters-isl | author:"Madonna" title:-sex -music was the query....whcih appears to not have the extra white spaces |
| # | 09:55:57 | eeevil | mrpeters-isl: thanks ... you can replace Madonna with Rowling and sec and music with foo and bar to test |
| # | 09:56:04 | mrpeters-isl | 10-4 |
| # | 09:56:16 | mrpeters-isl | the key is that extra space in front of " foo" right? |
| # | 09:56:48 | mrpeters-isl | hmmm no results for author:Rowling title:-foo -bar which is hard to imagine |
| # | 09:57:09 | mrpeters-isl | http://mig.evergreen.lib.in.us/opac/en-US/skin/default/xml/advanced.xml if you want to confirm |
| # | 09:57:28 | eeevil | rigth |
| # | 09:57:36 | eeevil | any extra spaces, actually |
| # | 09:57:41 | mrpeters-isl | oh, but matches exactly "Rowling" |
| # | 09:57:46 | mrpeters-isl | that will be why we get no results |
| # | 09:58:20 | mrpeters-isl | shoot...using "Contains" 'Rowling' gives no results |
| # | 09:59:35 | mrpeters-isl | http://mig.evergreen.lib.in.us/opac/skin/default/js/adv_global.js shows the patch...no need to restart services or anything right? |
| # | 10:00:16 | eeevil | just shift-reload, perhaps |
| # | 10:00:19 | dbs | clear cache in your browser |
| # | 10:00:32 | dbs | (not that I trust shift-reload to clear cache anymore) |
| # | 10:00:55 | mrpeters-isl | ill use a new browser just in cas |
| # | 10:01:10 | eeevil | looks like your harry potter records contain either (stemmed-to) foo or bar |
| # | 10:02:16 | mrpeters-isl | well, it at least seems like the patch worked since the search string trims out the white space properly, right? |
| # | 10:05:13 | mrpeters-isl | on another subject - runing through 1.6.1-2.0-upgrade-db.sql -- SELECT SETVAL('asset.uri'::TEXT, (SELECT MAX(id) + 1 FROM asset.uri)); -- gives an error "ERROR: "uri" is not a sequence" (I've got this set to stop on error) is this something that should be ignored or an actual error |
| # | 10:08:48 | dbs | mrpeters-isl: it should be SELECT SETVAL('asset.uri_id_seq'::TEXT... |
| # | 10:09:05 | dbs | mrpeters-isl: this in beta4? |
| # | 10:09:10 | mrpeters-isl | beta 3 still... |
| # | 10:09:26 | mrpeters-isl | should i just go right from 1.6.1 to beta 4? |
| # | 10:09:38 | dbs | yeah |
| # | 10:09:42 | eeevil | mrpeters-isl: if you want to test b4, yes |
| # | 10:10:11 | mrpeters-isl | ok. the biggest problem is we had a bunch of the triggered event stuff backported to our 1.6 database so i'm having a heck of a time making the jump to any of the 2.0's |
| # | 10:10:29 | mrpeters-isl | i had the beta 3 upgrade script trimmed down to take care of most of those |
| # | 10:10:36 | mrpeters-isl | but i can do that again |
| # | 10:10:56 | eeevil | mrpeters-isl: diff the original b3 and b4 scripts to show you what to /add/ to your trimmed script |
| # | 10:11:23 | mrpeters-isl | yep thats the plan after i sew them back together :) |
| # | 10:15:25 | dbs | heh. the weirdest thing about the flexibilitylab document in that obviously errant email to the general list is that I actually understand most of what I've scanned. benefit or downfall of having a kinesiologist spouse :) |
| # | 10:21:27 | yboston has joined #evergreen |
| # | 10:25:33 | Barb has joined #evergreen |
| # | 10:31:14 | mrpeters-isl | ok, even with -- Push the auri sequence in case it's out of date |
| # | 10:31:14 | mrpeters-isl | -- Add 2 as the sequence value must be 1 or higher, and seed is -1 |
| # | 10:31:14 | mrpeters-isl | SELECT SETVAL('asset.uri_id_seq'::TEXT, (SELECT MAX(id) + 2 FROM asset.uri)); |
| # | 10:31:14 | mrpeters-isl | I'm still getting the uri not a sequence error :( |
| # | 10:32:10 | mrpeters-isl | wait..ignore me |
| # | 10:32:18 | mrpeters-isl | thats an old error |
| # | 10:38:51 | natschil has joined #evergreen |
| # | 10:41:00 | phasefx | a contribution to the community hive mind: errors with bib merging may be due to a disabled or mangled protect_bib_rec_delete rule on biblio.record_entry. I've seen that twice now |
| # | 10:41:33 | dbs | phasefx++ # taking care of dangling or unanswered threads |
| # | 10:45:05 | Granitize1 has joined #evergreen |
| # | 10:51:41 | phasefx | the formats menu in the OPAC, I figure it's easy to trim stuff from it, but where should I look for adding things? In the opac xml, am I just able to specify combinations of codes from config.item_type_map? |
| # | 10:56:08 | eeevil | phasefx: and item_form |
| # | 10:58:22 | phasefx | ah, so separate the two with a dash |
| # | 10:58:31 | phasefx | eeevil++ |
| # | 11:05:22 | natschil has quit IRC |
| # | 11:10:44 | mrpeters-isl | b4 upgrade craps out on CREATE INDEX metabib_author_field_entry_source_idx ON metabib.author_field_entry (source); |
| # | 11:11:09 | mrpeters-isl | ERROR: relation "metabib_author_field_entry_source_idx" already exists |
| # | 11:11:14 | mrpeters-isl | then kills the xact |
| # | 11:24:07 | parsr has joined #evergreen |
| # | 11:24:08 | b_bonner has joined #evergreen |
| # | 11:24:33 | b_bonner has joined #evergreen |
| # | 12:01:13 | parsr | OpenSRF - current version is 1.6.1 but recall seeing talk of patch coming - patch is not yet available, correct? |
| # | 12:01:35 | parsr | Also, sample scenario say you have OpenSRF 1.6.1 & EG 1.6.1.4 - Would you ever do just an upgrade to OpenSRF (say 1.6.1 to 1.6.2) and not touch openils (or do you need to recompile your Evergreen against any new OpenSRF)? |
| # | 12:02:16 | parsr | i.e. can you just upgrade OpenSRF by itself if patch available? |
| # | 12:10:42 | tsbere throws some SQL thought for stuff from earlier this morning at mrpeters-isl: SELECT SETVAL('asset.uri_id_seq'::TEXT, (SELECT GREATEST(MAX(id),0)+1 FROM asset.uri)); |
| # | 12:11:05 | eeevil | parsr: it would be safest to recompile EG ... until we can find the properly shaped tuits to implement library versioning in the autotools bits |
| # | 12:13:49 | dbs | eeevil: library versioning is in OpenSRF trunk |
| # | 12:14:28 | dbs | tsbere: could do that for maximum conservation of one sequence ID, but... have you looked at the ingest logic for uris? |
| # | 12:14:38 | dbs | 1 sequence ID is nothing compared to 1000 |
| # | 12:15:04 | tsbere | dbs: I saw the +2 workaround in the scrollback. That is about it. <_< |
| # | 12:15:14 | eeevil | dbs: really?! |
| # | 12:15:16 | eeevil | dbs++ |
| # | 12:15:28 | dbs | eeevil: heh. two line patch or something :) |
| # | 12:16:30 | eeevil | dbs: the -version-info bit? |
| # | 12:17:01 | dbs | yup |
| # | 12:17:36 | eeevil | nice ... one small line for autotools, one giant leap for opensrf apps |
| # | 12:17:42 | tsbere | dbs: Also, I have never looked at any ingest logic. If anyone here has, it is Dyrcona. |
| # | 12:17:48 | dbs | I probably should have included a comment that summarizes how -version-info works, because it's pretty much an accident that 2.0.0 matches version 2.0.0 |
| # | 12:19:28 | b_bonner has quit IRC |
| # | 12:27:04 | dbs | current 1.6 behaviour increments callnumber and asset.uri by 1000 for every record that contains a located URI, IIRC |
| # | 12:28:28 | dbs | hmm, that can't be right, we're only at 14533906 for asset.uri |
| # | 12:28:47 | dbs | and 6293965 for acn |
| # | 12:32:30 | tsbere | Any chance of pointing me at the SQL files that define that stuff? Or it is in the perl layer? |
| # | 12:32:57 | dbs | In 1.6, it's perl (OpenILS/Application/Ingest.pm) |
| # | 12:33:32 | dbs | in trunk, the behaviour is much nicer, and defined in SQL (030.metabib I believe in the indexing_ingest_or_delete trigger) |
| # | 12:35:53 | tsbere | Looks reasonable in the sql form, but I don't think I could do going over it justice for a week or two (assuming I started learning how that all works today, instead of doing other things that need to be done) |
| # | 12:39:20 | dbs | hmm, up to asset.uri.id = 14779396 now |
| # | 12:40:05 | dbs | in the process of ingesting about 5,000 records with URIs. fun fun |
| # | 12:41:35 | tsbere finds himself wondering if he will need to re-write his receipt printer program for the legacy ils to work with evergreen in some fashion :( |
| # | 12:46:58 | sfortin has quit IRC |
| # | 12:47:31 | tsbere | I just got off the phone with a library that doesn't think their receipt printers are recent enough to have windows drivers. :( |
| # | 12:47:34 | b_bonner has joined #evergreen |
| # | 12:49:38 | dbs | whoa |
| # | 12:49:54 | dbs | they probably have linux drivers though... :) |
| # | 12:50:54 | tsbere | Just got a correction via email: They might be able to find drivers, if the brand and model identifiers weren't all missing. The thing is so old the labels and such are all worn to nothingness. |
| # | 12:53:35 | mtcarlson has joined #evergreen |
| # | 13:16:54 | mjg_ has joined #evergreen |
| # | 13:18:53 | parsr has quit IRC |
| # | 13:19:53 | jenny has quit IRC |
| # | 13:30:45 | mrpeters-isl | any thoughts on why the 1.6.1 to 2.0 upgrade from beta 4 is dying out at ERROR: relation "metabib_author_field_entry_source_idx" already exists? I have this index in even my 1.6.0.0 server, so of course it exists? |
| # | 13:31:24 | mrpeters-isl | perhaps I need to DROP INDEX IF EXISTS in the script before it tries to create those 4 indexes? |
| # | 13:31:50 | eeevil | perhaps. I suggest dropping it and trying again |
| # | 13:32:13 | mrpeters-isl | eeevil: ok, but it's normal that those exist prior to 2.0 right? |
| # | 13:32:25 | mrpeters-isl | and this isn't some customization that i'm forgetting about? |
| # | 13:33:37 | mrpeters-isl | ok now this is funky - evergreen=# DROP INDEX metabib_author_field_entry_source_idx; |
| # | 13:33:37 | mrpeters-isl | ERROR: index "metabib_author_field_entry_source_idx" does not exist |
| # | 13:34:12 | mrpeters-isl | its almost like the transaction is trying to do this twice |
| # | 13:34:15 | mrpeters-isl | CREATE INDEX metabib_title_field_entry_value_idx ON metabib.title_field_entry (SUBSTRING(value,1,1024)) WHERE index_vector = ''::TSVECTOR; |
| # | 13:34:33 | mrpeters-isl | then ... CREATE INDEX metabib_title_field_entry_source_idx ON metabib.title_field_entry (source); just a bit later |
| # | 13:34:48 | mrpeters-isl | err...i meant |
| # | 13:34:48 | mrpeters-isl | CREATE INDEX metabib_author_field_entry_source_idx ON metabib.author_field_entry (source); |
| # | 13:35:54 | mrpeters-isl | ah value != source |
| # | 13:36:00 | eeevil | yeah |
| # | 13:37:43 | eeevil | if we're going to do that, and I don't know that we should, we should drop them before (or inside with IF EXISTS) the transaction |
| # | 13:38:27 | mrpeters-isl | whats puzzling me is where they got dropped in the first place |
| # | 13:38:38 | eeevil | they don't |
| # | 13:38:41 | mrpeters-isl | because if i try to drop them now, it says they aren't there |
| # | 13:39:04 | eeevil | trac is complaining about the file being to big to annotate, so doing that from the shell now |
| # | 13:39:22 | mrpeters-isl | heh yeah this is a monster script |
| # | 13:40:30 | mrpeters-isl | postgresql-8.4-main.log:2010-12-03 12:09:22 EST STATEMENT: CREATE INDEX metabib_author_field_entry_source_idx ON metabib.author_field_entry (source); |
| # | 13:40:30 | mrpeters-isl | postgresql-8.4-main.log:2010-12-03 12:50:02 EST ERROR: relation "metabib_author_field_entry_source_idx" already exists |
| # | 13:40:31 | mrpeters-isl | postgresql-8.4-main.log:2010-12-03 12:50:02 EST STATEMENT: CREATE INDEX metabib_author_field_entry_source_idx ON metabib.author_field_entry (source); |
| # | 13:40:35 | mrpeters-isl | those should be relevant lines from the postgres logs |
| # | 13:41:00 | mrpeters-isl | i guess it took some time to error, since this started at 12:09 and died at 12:50 |
| # | 13:41:50 | eeevil | or from two runs of your chopped up script? |
| # | 13:42:36 | mrpeters-isl | probably more likely |
| # | 13:42:42 | mrpeters-isl | i'm using 1 single script now though |
| # | 13:42:56 | mrpeters-isl | can't remember what time i started the most recent run...sometime before lunch |
| # | 13:43:04 | eeevil | mrpeters-isl: you know, I bet those are not stock indexes ... they were added to your db by hand, not by default schemas |
| # | 13:43:27 | eeevil | not stock in 1.6 I mean |
| # | 13:43:32 | mrpeters-isl | eeevil: very possible. so i'm wondering if now that they are gone, the upgrade might just run properly |
| # | 13:43:44 | eeevil | if you've dropped them, it should |
| # | 13:43:53 | mrpeters-isl | ok. i will give it another pass |
| # | 13:44:46 | mrpeters-isl | it's vaguely familiar that gmcharlt might have added (or advised us to add) those extra indexes a while back |
| # | 13:48:10 | djfiander has joined #evergreen |
| # | 13:49:07 | afterl has joined #evergreen |
| # | 13:49:38 | djfiander | NCIP is making my brain melt. |
| # | 13:50:01 | dbs | djfiander: was it the brown NCIP? I heard that stuff was bad. |
| # | 13:50:49 | afterl_ has joined #evergreen |
| # | 13:50:57 | djfiander | no, this is NCIP classic. |
| # | 13:51:24 | afterl_ | dbs is funny |
| # | 13:52:12 | dbs | fwiw, there appears to be no indexes other than the index vector created on metabib.author_field_entry in stock rel_1_6; that said, we can chalk it up as a best learning practice to DROP IF EXISTS anything droppable before creating in the future, I guess |
| # | 13:52:59 | dbs | Unless, perhaps, somebody defines a local object "foo" with a vastly different meaning than the newly introduced one - in which case, erroring is good. |
| # | 13:53:02 | dbs is so conflicted |
| # | 13:53:02 | mrpeters-isl | heh all i do is cause trouble |
| # | 13:54:19 | bshum has joined #evergreen |
| # | 13:54:26 | afterl has quit IRC |
| # | 13:55:47 | afterl_ has quit IRC |
| # | 14:00:50 | afterl has joined #evergreen |
| # | 14:06:45 | afterl has quit IRC |
| # | 14:13:16 | [1]Marduk has joined #evergreen |
| # | 14:15:05 | mrpeters-isl | CURSE! ERROR: relation "metabib_author_field_entry_source_idx" already exists |
| # | 14:15:24 | mrpeters-isl | evergreen=# DROP INDEX metabib_author_field_entry_source_idx; |
| # | 14:15:24 | mrpeters-isl | ERROR: index "metabib_author_field_entry_source_idx" does not exist |
| # | 14:15:29 | mrpeters-isl | so which is it! ugh! |
| # | 14:16:05 | phasefx | probably a path thing. may need to specify the table along with the index for the drop |
| # | 14:16:11 | dbs | probably DROP INDEX metabib.metabib_author... |
| # | 14:16:33 | dbs | (needs to be schema-qualified) |
| # | 14:16:40 | mrpeters-isl | ah yes that did the trick |
| # | 14:16:50 | phasefx wrong, dbs right. unga bunga |
| # | 14:18:11 | afterl has joined #evergreen |
| # | 14:18:13 | mrpeters-isl crosses fingers this time |
| # | 14:18:26 | bshum | mrpeters-isl++ # digging at the 1.6.1-2.0 upgrade script |
| # | 14:19:07 | mrpeters-isl | hah i'm not sure i'm doing much good! |
| # | 14:20:01 | mrpeters-isl | except for causing dbs to be conflicted :P |
| # | 14:20:09 | bshum | I've been there ;) |
| # | 14:20:18 | bshum | Speaking of metabib stuff |
| # | 14:21:02 | bshum | I've been poking at this blog post by dbs about adding additional indexes and weights for relevance: http://coffeecode.net/archives/218-Adjusting-relevancy-rankings-in-Evergreen-1.6,-some-explorations.html |
| # | 14:21:31 | bshum | Have you poked at this more lately dbs? I just implemented part of this to increase title matches for general keyword searches. |
| # | 14:22:24 | bshum | Or anyone else for that matter. |
| # | 14:22:46 | bshum | I've been wondering if we'd need to modify triggers or whatnot to generate the additional metabib entries |
| # | 14:22:51 | bshum | For new bibs we import |
| # | 14:22:53 | bshum | Or create |
| # | 14:26:05 | Maryll has joined #evergreen |
| # | 14:26:27 | dbs | Should get generated automatically on ingest or reingest IIRC |
| # | 14:27:17 | [1]Marduk has quit IRC |
| # | 14:28:00 | afterl has quit IRC |
| # | 14:28:08 | bshum | dbs: Oh really? That's awesome... wasn't sure if I needed extra steps. |
| # | 14:28:45 | bshum | Thanks :) |
| # | 14:28:56 | bshum | I'll test it some more and see what happens. |
| # | 14:33:44 | eeevil | bshum: anything updated or inserted after the definition is added uses the definition it sees. also, changing the weight value does /not/ require reingesting |
| # | 14:34:32 | eeevil | bshum: only changing the xpath (or search_field/facet_field) would, for existing index defs |
| # | 14:36:21 | bshum | eeevil: Ah alright. Neat. |
| # | 14:37:30 | Maryll has quit IRC |
| # | 14:37:42 | bshum | Well, wandering out for the weekend. Have a good one everyone |
| # | 14:37:44 | dbs has quit IRC |
| # | 14:37:58 | Maryll has joined #evergreen |
| # | 14:38:09 | bshum has quit IRC |
| # | 14:39:49 | Maryll has left #evergreen |
| # | 14:54:13 | mrpeters-isl | gotta ask another one - CREATE INDEX metabib_series_field_entry_source_idx ON metabib.series_field_entry (source); is giving postgresql-8.4-main.log:2010-12-03 14:48:03 EST DETAIL: Key (source)=(5300001) is not present in table "record_entry". |
| # | 14:54:28 | mrpeters-isl | source = bre.id correct? |
| # | 14:55:40 | mrpeters-isl | thus, if i don't have a bre.id = 5300001 why is it keying on one with that source? |
| # | 14:56:10 | mrpeters-isl | ah doh you know what |
| # | 14:56:20 | eeevil | correct ... there was, at one, a missing fkey there. looks like it still is missing on your db ;) |
| # | 14:56:20 | mrpeters-isl | with those earlier DROP INDEXES i bet i need to truncate all of those tables |
| # | 14:56:30 | eeevil | no, don't truncate those |
| # | 14:56:33 | mrpeters-isl | no? ok |
| # | 14:56:50 | mrpeters-isl | SELECT * FROM metabib.series_field_entry WHERE source=5300001; returns a row, but there is no corresponding bre |
| # | 14:56:56 | mrpeters-isl | so just delete from those tables where source NOT IN bre? |
| # | 14:57:10 | eeevil | just: DELETE FROM metabib.series_field_entry WHERE source NOT IN (select id from biblio.record_entry); |
| # | 14:57:14 | eeevil | right |
| # | 14:57:17 | mrpeters-isl | awesome ok |
| # | 14:57:30 | jeff has quit IRC |
| # | 14:57:41 | mrpeters-isl | so, this is probably something noone else will run into? |
| # | 14:57:51 | eeevil | unsure... |
| # | 14:58:36 | mrpeters-isl | ok i dont want to cause confusion based on things that could be my fault |
| # | 15:00:54 | mrpeters-isl | new topic - we're getting searches that hang up in our database for 15, 20... minutes at a time |
| # | 15:00:59 | mrpeters-isl | it appears to be when an & is used |
| # | 15:01:08 | mrpeters-isl | 22430 | 00:18:03.607991 | SELECT * |
| # | 15:01:08 | mrpeters-isl | : FROM search.staged_fts( |
| # | 15:01:08 | mrpeters-isl | : 162::INT, |
| # | 15:01:08 | mrpeters-isl | : 2::INT, |
| # | 15:01:08 | mrpeters-isl | : $${"author":{"fts_query":["to_tsquery('author','debbie&macomber')"],"word":["debbie","macomber"],"fts_rank":["rank(author.index_vector, to_tsquery('author','debbie&macomber'))"],"phrase":[]},"title":{"fts_query":["to_tsquery('title','someday&soon')"],"word":["someday","soon"],"fts_rank":["rank(title.index_vector, to_tsquery('title','someday& |
| # | 15:01:09 | mrpeters-isl | soon'))"],"phrase":[]}}$$::TEXT, |
| # | 15:01:30 | mrpeters-isl | if not checked, these just run on and on and on (and this is in our production enviornment) |
| # | 15:01:54 | mrpeters-isl | just curious if anyone else has seen similiar behavor |
| # | 15:03:44 | eeevil | mrpeters-isl: without the full query, and looking at it in your env, I couldn't say. |
| # | 15:04:05 | jenny has joined #evergreen |
| # | 15:04:25 | mrpeters-isl | ok np |
| # | 15:04:51 | mrpeters-isl | i've just been killing them off with pg_cancel_backend when they get out of hand and run for over 20 min |
| # | 15:11:37 | jeff has joined #evergreen |
| # | 15:20:49 | Granitize1 | any more talk of indexing solr with evergreen? |
| # | 15:21:03 | Granitize1 | er… evergreen with solr |
| # | 15:24:59 | eeevil | nothing concrete, just "it'd be neat if..." |
| # | 15:25:21 | eeevil | Granitize1: there are Hard(tm) problems to be solved there |
| # | 15:27:41 | eeevil | if we want to provide a true replacement indexing backend, and not just a bolt-on separate interface |
| # | 15:30:33 | b_bonner has quit IRC |
| # | 15:34:10 | mrpeters-isl | ERROR: insert or update on table "metabib_search_alias" violates foreign key constraint "metabib_search_alias_field_fkey" |
| # | 15:34:11 | mrpeters-isl | DETAIL: Key (field)=(16) is not present in table "metabib_field". |
| # | 15:34:14 | mrpeters-isl | not familiar with this table.... |
| # | 15:34:17 | mrpeters-isl | new in 2.0? |
| # | 15:34:40 | Granitize1 | eeevil: bolt on would go a long way to facilitating a meta-search with other indexes… for the meta-search lovers out there. |
| # | 15:35:36 | Granitize1 | eeevil: guess a schedules export/import into an existing index would be possible. |
| # | 15:36:16 | b_bonner has joined #evergreen |
| # | 15:36:53 | eeevil | Granitize1: bolt-on is outside the scope of EG proper, unless the command-line export tools need work to facilitate it ... I think they're smart enough today, though |
| # | 15:36:56 | mrpeters-isl | I assume the INSERT INTO config.metabib_search_alias (alias,field_class,field) VALUES ('bib.subjectoccupation','subject',16); is where I'm failing...since my config.metabib_field only goes up to 15? |
| # | 15:37:33 | eeevil | mrpeters-isl: possibly ... there's a whole pile of new defs you should have from earlier in the upgrade script |
| # | 15:37:57 | mrpeters-isl | hmmm i have 17, 18, 19... |
| # | 15:37:59 | mrpeters-isl | but no 16 |
| # | 15:38:18 | mrpeters-isl | comes right after -- resolves performance issue noted by EG Indiana in the upgrade script |
| # | 15:38:52 | mrpeters-isl | is it possible #16 was left out accidentally? or should i have it from an earlier release than 1.6.1 |
| # | 15:39:17 | eeevil | I think you should have 16 already ... IIRC, that's the "all subjects" index |
| # | 15:39:35 | mrpeters-isl | hmm |
| # | 15:39:56 | mrpeters-isl | not on production server either... |
| # | 15:40:03 | mrpeters-isl | so it must come in after 1.6.0.90 |
| # | 15:40:09 | mrpeters-isl | err 1.6.0.0 |
| # | 15:42:30 | b_bonner has quit IRC |
| # | 15:44:46 | mrpeters-isl | INSERT INTO config.metabib_field ( id, field_class, name, label, format, xpath ) VALUES |
| # | 15:44:47 | mrpeters-isl | (16, 'subject', 'complete', oils_i18n_gettext(16, 'All Subjects', 'cmf', 'label'), 'mods32', $$//mods32:mods/mods32:subject//text()$$ ); from 950.data.seed-values.sql perhaps solves this? |
| # | 15:48:09 | dbs has joined #evergreen |
| # | 15:48:48 | b_bonner has joined #evergreen |
| # | 15:52:54 | eeevil | mrpeters-isl: yes, but that came before 2.0 |
| # | 15:52:59 | mrpeters-isl | hmm |
| # | 15:53:13 | mrpeters-isl | of concern that we don't have it in our prodution database then... |
| # | 15:53:21 | eeevil | mrpeters-isl: have you been upgrading to (at least) 1.6.0.4? |
| # | 15:53:30 | eeevil | for this test, I mean |
| # | 15:53:47 | mrpeters-isl | yes |
| # | 15:54:03 | mrpeters-isl | starting with 1.6.0.0-1.6.0.1-upgrade-db.sql and working my way up |
| # | 15:54:20 | eeevil | really, you should be going through all the scripts in the 1.6.0.10 tarball, and then moving on to 1.6.0-1.6.1, if you can |
| # | 15:55:12 | mrpeters-isl | i see...ive been stopping at 1.6.1.3-1.6.1.4-upgrade-db.sql and then moving on to 1.6.1-2.0-upgrade-db.sql |
| # | 15:55:54 | dbs | rename 1.6.1-2.0 to 1.6.1.10-2.0? |
| # | 15:55:56 | mrpeters-isl | just using what was in the latest tarball |
| # | 15:56:20 | mrpeters-isl | and then 1.6.1-2.0-upgrade-db.sql from beta4 |
| # | 15:57:31 | mrpeters-isl | so 1.6.0.10 has an upgrade for .1 thru .10 while 1.6.1.4 just has 1.6.0.4-1.6.1.0-upgrade-db.sql |
| # | 15:57:39 | mrpeters-isl | was that 1.6.0.4-1.6.1.0-upgrade-db.sql not the same as going through these 1 by one |
| # | 16:00:00 | mrpeters-isl | i'm not finding ... (16, 'subject', 'complete', oils_i18n_gettext(16, 'All Subjects', 'cmf', 'label'), ... in any of the scripts from 1.6.0.0 on up :( |
| # | 16:01:20 | Granitize1 has quit IRC |
| # | 16:01:34 | Granitize1 has joined #evergreen |
| # | 16:02:38 | eeevil | hrm... ack says it's indeed not in rel_1_6 (no instance of the string 'All Subjects') |
| # | 16:02:46 | eeevil | that's lame :( |
| # | 16:03:38 | Meliss has quit IRC |
| # | 16:03:43 | dbs | mrpeters-isl: hah, even I got confused by the jumps from 1.6.0.x to 1.6.1.x |
| # | 16:04:18 | mrpeters-isl | eeevil: also not in the beta 3 tar |
| # | 16:05:22 | mrpeters-isl | so ill try th upgrade again with that insert in there and hope for the best! |
| # | 16:08:48 | dbs | Perhaps "database upgrade script maintenance strategy" should be a dev meeting agenda item |
| # | 16:08:48 | mrpeters-isl | if anyone discovers in which upgrade that *should* have occured please let me know |
| # | 16:08:58 | djfiander | heh |
| # | 16:09:28 | eeevil | mrpeters-isl: according to svn, it should be part of the 1.6.1-2.0 upgrade |
| # | 16:09:48 | mrpeters-isl | hmm not in beta 3 or 4 tarballs |
| # | 16:10:02 | b_bonner has quit IRC |
| # | 16:10:11 | mrpeters-isl | downloading it just 1 more time to make sure.... |
| # | 16:10:29 | eeevil | mrpeters-isl: no, I'm not saying the script is correct (anywhere) |
| # | 16:10:58 | mrpeters-isl | ah ok i thought if svn showed it as being updated it would have made it into the tarball |
| # | 16:11:07 | eeevil | I'm saying that index def 16 is, in fact, new to 2.0. I thought it came in earlier because dbs invented it quite a long time ago |
| # | 16:11:15 | mrpeters-isl | ok |
| # | 16:11:22 | eeevil | (which is not pointing a finger at dbs!) |
| # | 16:11:25 | mrpeters-isl | got it |
| # | 16:11:28 | dbs | hah |
| # | 16:11:36 | mrpeters-isl | well now we know :) |
| # | 16:11:50 | mrpeters-isl | I'm rolling with |
| # | 16:11:50 | mrpeters-isl | INSERT INTO config.metabib_field ( id, field_class, name, label, format, xpath ) VALUES |
| # | 16:11:50 | mrpeters-isl | (16, 'subject', 'complete', oils_i18n_gettext(16, 'All Subjects', 'cmf', 'label'), 'mods32', $$//mods32:mods/mods32:subject//text()$$ ); |
| # | 16:12:09 | mrpeters-isl | and placing it just before INSERT INTO config.metabib_field ( id, field_class, name, label, format, xpath, facet_field ) VALUES |
| # | 16:12:09 | mrpeters-isl | (17, 'identifier', 'accession', oils_i18n_gettext(17, 'Accession Number', 'cmf', 'label'), 'marcxml', $$//marcxml:datafield[tag="001"]/text()$$, TRUE ); |
| # | 16:12:22 | eeevil | dbs: seriously ... it's likely my fault. I just recall you building that for conifer ... and I could be wrong about that too |
| # | 16:12:38 | eeevil | anyway, I'll put it in the upgrade script now |
| # | 16:12:42 | dbs | eeevil: version control by blog post |
| # | 16:12:45 | mrpeters-isl | eeevil++ |
| # | 16:12:57 | dbs | I'm pretty sure I blogged / mailed the list about it. |
| # | 16:13:09 | dbs | and we definitely talked about it in IRC :) |
| # | 16:13:42 | b_bonner has joined #evergreen |
| # | 16:13:49 | dbs | So who wants to go check out that CI webinar next week? :) |
| # | 16:13:51 | dbs whistles |
| # | 16:14:13 | eeevil | heh |
| # | 16:15:55 | b_bonner_ has joined #evergreen |
| # | 16:17:55 | b_bonner has quit IRC |
| # | 16:17:55 | b_bonner_ is now known as b_bonner |
| # | 16:20:26 | mrpeters-isl | thanks for 18919 |
| # | 16:21:57 | mrpeters-isl | eeevil: dang it that won't work |
| # | 16:22:08 | mrpeters-isl | it failed on me again ERROR: insert or update on table "metabib_search_alias" violates foreign key constraint "metabib_search_alias_field_fkey" |
| # | 16:22:08 | mrpeters-isl | DETAIL: Key (field)=(16) is not present in table "metabib_field". |
| # | 16:22:13 | mrpeters-isl | maybe it has to be there before the transaction ever runs? |
| # | 16:25:11 | eeevil | no, just before the aliases are added |
| # | 16:25:25 | mrpeters-isl | yeah...i've got it in the same place as you |
| # | 16:26:21 | eeevil | yeah... need to move it up, or the aliases down |
| # | 16:26:38 | eeevil | I'm going to move the aliases down |
| # | 16:26:52 | mrpeters-isl | cool! i'll grab the new one monday |
| # | 16:26:54 | eeevil | no I'm not ... |
| # | 16:26:56 | mrpeters-isl | have a good weekend guys - thanks |
| # | 16:29:44 | b_bonner_ has joined #evergreen |
| # | 16:33:00 | b_bonner has quit IRC |
| # | 16:33:00 | b_bonner_ is now known as b_bonner |
| # | 16:34:42 | kmlussier has quit IRC |
| # | 16:35:46 | mtcarlson has quit IRC |
| # | 16:39:11 | dbs | vandelay++ # currently simultaneously processing three batches of MARC records |
| # | 16:39:49 | eeevil | rad :) |
| # | 16:40:17 | eeevil | now to clean up the UI so it matches the backend .... (with regard to item import, at least) |
| # | 16:40:20 | eeevil out |
| # | 16:40:52 | dbs | eeevil++ # enjoy the weekend |
| # | 16:51:04 | AaronZ has quit IRC |
| # | 17:08:44 | dbs has quit IRC |
| # | 17:08:51 | djfiander has left #evergreen |
| # | 17:11:05 | mjg_ has left #evergreen |
| # | 17:36:42 | brian_f has joined #evergreen |
| # | 17:47:58 | yboston has quit IRC |
| # | 17:55:55 | jenny has left #evergreen |
| # | 18:22:30 | b_bonner has quit IRC |
| # | 18:26:27 | b_bonner has joined #evergreen |
| # | 19:01:26 | phasefx | granitize: getfirefox.com, has a screenshot with UPEI's wikipedia entry :) |
| # | 19:01:40 | phasefx | well, not upei |
| # | 19:01:43 | phasefx | but pei |
| # | 19:07:26 | b_bonner has quit IRC |
| # | 19:51:40 | moodaepo has quit IRC |
| # | 21:57:40 | brian_f has quit IRC |
April 10-13