Evergreen ILS Website

IRC log for #evergreen, 2021-12-09

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat

All times shown according to the server's local time.

Time Nick Message
04:08 JBoyer joined #evergreen
06:00 pinesol News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>
07:33 rjackson_isl_hom joined #evergreen
07:48 collum joined #evergreen
08:19 rfrasur joined #evergreen
08:21 Dyrcona joined #evergreen
09:00 mantis joined #evergreen
09:04 Dyrcona I'm working on a program to read a list of copy barcodes from a spreadsheet, look them up in Evergren, check them in, and then delete the copy.
09:09 Dyrcona I have a spreadsheet with 2,779 barcodes. Of these., 324 appear to be deleted already.. https://pastebin.com/H9A0RYwU should output "Total Barcodes: 2779\nBarcodes Found: 2459". Instead, the latter number is always 499.
09:11 Dyrcona I've also tried variations where the cstore session is created outside the loop, I've tried searching with a CStoreEditor. The result is always 499. A database query for non-deleted copies with the barcode list returns 2459.
09:11 Dyrcona The question for the OpenSRF experts is, what am I not doing/doing wrong?
09:14 Dyrcona When I run the full program to do the deletions, it reports the same 499 found, deleted, etc. Subsequent runs fail to find any copies.
09:18 Dyrcona I'll try a version no using Cronscript.pm to see if that makes a difference.
09:21 Dyrcona And, no, I don't think I will. That would be too much effort for this. :)
09:28 Dyrcona Ok. I get the correct numbers in production, so it is something in my development environment.
09:29 Dyrcona I also switched to using Apputils->simplereq rather than manually creating the session.
09:30 Dyrcona That second version still only gives 499 on my development server.
09:34 Dyrcona So, what setting should I look at?
09:35 Dyrcona Hmm... git diff says my opensrf.xml.example files are the same.
09:39 Dyrcona It's not a data difference, either. The development server uses a copy of production, and the lookup SQL gives the same number of not-deleted copies as in production: 2,455.
09:41 jvwoolf joined #evergreen
09:46 Dyrcona 2021-12-09 09:16:58.597 [info] <0.2465.0>@ejabberd_c2s:process_terminated:271 (tcp|<0.2465.0>) Closing c2s session for opensrf@jasontest.private/client_at_​jasontest.cwmars.org_20184: Connection failed: connection closed
09:47 Dyrcona Maybe I missed an ejabberd setting. Bet it's the rate limiter/whatever.
09:58 csharp_ Dyrcona: I'm seeing tons of those messages in my ejabberd.log
10:04 Dyrcona Yeah. I looked it up online, and it's a nonissue. I've determined it is not an ejabberd problem on my development vm.
10:04 Dyrcona The behavior suggests a data problem, but I've checked the database so many times, unless it's the number of connections allowed....
10:05 jeff Have the program sleep when it gets to 499 - 1 and look at the system state?
10:06 jeff If you're not interacting with anything that passes through Apache, then it's probably not the recommended "MaxConnectionsPerChild 500" config setting there.
10:07 jeff But I wonder if you're creating a new *something* (connection/request object/session/etc) for each query and not closing/destroying/reusing it.
10:09 Dyrcona jeff: I'll try the sleep if upping the Pg connections doesn't work. It's not using Apache. It's Perl. If you look at the script that I pasted, it finishes and disconnects. Apputils' simplereq disconnects when it is done. It works in production, so it's got to be a configuration or resource issue.
10:11 Dyrcona Pg max_connections made no difference.
10:13 Dyrcona Weird...It seems to skip the first couple of thousand barcodes.
10:17 Dyrcona Remove the uniq and sort, it still does only 499, but it starts finding them earlier.... The query to look up the copies by barcode still says there are 2,455.
10:18 Dyrcona I think I'll wipe out the /openils directory and the perl modules and reinstall. I did a make install yesterday, but maybe there's some junk from a different eg version hanging around.
10:18 Dyrcona Failing that, I was going to build a new VM, so what's another?
10:28 Dyrcona Now, the script can't authenticate, says no seed found.
10:29 Dyrcona services say they're running.
10:30 Dyrcona Oh, right. need to fix the memcached settings. :)
10:34 Dyrcona And, I'm back where I started....
10:34 Dyrcona "Nuke it from orbit. It's the only way to be sure."
10:35 Dyrcona @blame Excel/xlsx
10:35 pinesol Dyrcona: Excel/xlsx WILL PERISH UNDER MAXIMUM DELETION! DELETE. DELETE. DELETE!
10:36 Dyrcona Looking at the log more carefully, I see this: Dec  9 10:34:05 jasontest open-ils.cstore: [INFO:6887:osrf_application.c:1075:] CALL: open-ils.cstore open-ils.cstore.direct.asset.copy.search {"barcode":"3.0453000494612E13","deleted":"f"},{"​flesh_fields":{"acp":["circulations"]},"flesh":1}
10:37 jeff aha!
10:37 Dyrcona The barcode is a "number" according to that output. I guess only the barcodes that begin with the letter A are working. Now, why this happens on the test vm and not in production... Hm. I edited the sheet in LibreOffice this morning. Maybe that "fixed" it?
10:38 jeff (now some fun is to be had with determining if the text->number type confusion happened at log time or before.)
10:39 Dyrcona jeff: It's before. Excel does that with barcodes. I think the spreadsheet I used in production was editing in LibreOffice. I'll copy it to the test vm again and see what happens.
10:39 jeff oh, if you're using Spreadsheet::Read and the input barcode file is different between test and production, I suspect the input file is your issue now.
10:40 * jeff nods
10:42 Dyrcona yeahp. It's chugging away, now.
10:42 Dyrcona GIGO
10:44 Dyrcona So, they want the already deleted copies checked in, so I should go back to the cstore session or editor, remove the deleted=>'f', etc., etc., and so on and so forth.
10:44 Dyrcona jeff++
10:44 Dyrcona csharp++
10:54 Dyrcona OK. It is working now. I forgot to change to check in by copy id, though. I'm getting asset copy not found on the deleted ones.
12:35 collum joined #evergreen
13:06 Dyrcona joined #evergreen
14:26 rjackson_isl_hom joined #evergreen
15:19 alynn26_away joined #evergreen
15:38 bott_grpl joined #evergreen
15:48 bott_grpl Can someone point me toward where the summary.bibCallNumber is populated in the bib summary for /eg2/en-US/staff/catalog/record/ ? It's showing me LC info, where I've got Dewey showing in the volcopy editor, which is what we want.
16:07 berick bott_grpl: eg2/src/app/share/catalog/bib-record.service.ts
16:07 berick it calls open-ils.cat.biblio.record.marc_cn.retrieve
16:10 jeffdavis ... which in turn will pull the call number from the MARC record; looks like you may be able to set cat.default_classification_scheme to Dewey to make it pull a Dewey call number instead of LC
16:11 bott_grpl The scheme is set, which was part of my confusion.
16:13 bott_grpl As it works as expected in the volcopy editor, but not the record display in /eg2/
16:16 jeffdavis It should be grabbing the first value it finds from the list of fields specified in the field column of asset.call_number_class for your default classification scheme
16:17 jeffdavis at least, that's my read of the source for the open-ils.cat.biblio.record.marc_cn.retrieve call
16:18 berick srfsh# request open-ils.cat open-ils.cat.biblio.record.marc_cn.retrieve 1
16:18 berick change record ID to taste
16:19 bott_grpl Ooo, a Status: 500 from that srfsh request is probably bad!
16:21 bott_grpl Ugh. id typo
16:22 bott_grpl But exactly what I was looking for!
16:22 bott_grpl berick++ jeffdavis++
17:31 jvwoolf left #evergreen
18:00 pinesol News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>
18:45 pinesol [evergreen|Jason Stephenson] Lp 1862652: pingest.pl reingest record attributes fix - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=5ac76ef>
23:03 alynn26_away joined #evergreen

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat