Result pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139
12:48 | sandbergja | I spent an embarassing amount of time troubleshooting why every request I made to the open-ils.serials service returned an error. Fun fact: the service is actually called open-ils-serial. |
12:49 | sandbergja | ^ open-ils.serial, rather |
12:52 | mantis1 joined #evergreen | |
13:03 | jeffdavis | abneiman: thanks for the reminder about the DYM-related search slowness bug - our testing results weren't as definitive as I would like but I believe the fix works, I'll make a signoff branch and update 2038472 |
13:04 | jeffdavis | I guess there's already a signoff branch, I'll just add a signoff via LP |
13:07 | Stompro joined #evergreen | |
13:08 | abneiman | jeffdavis++ |
13:08 | abneiman | thank you! |
13:15 | jeffdavis | More specifically, our test environment remained somewhat slow after applying the fix, but I believe the remaining slowness was for other reasons. We decided not to enable DYM so we effectively stopped testing the fix at a certain point; it's possible that it introduces other issues, but no point in delaying the fix over such hypotheticals. |
13:16 | jeffdavis | "other reasons" = slower hardware + that unrelated bug that required disabling a PG "optimization" |
13:41 | Dyrcona | jeffdavis: I have not experienced the issue that "requires" disabling jit. |
13:42 | Dyrcona | I'm using Pg 15 for quite a bit of testing. Of course, the VPN to where that server is just went down, so..... |
13:42 | jeffdavis | required it on our test server, that is - we haven't seen it in production |
13:44 | pinesol | News from commits: LP#2038472: stamp DB update <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=10527ec78952a27d447f675a93d13e2846fa7ae6> |
13:44 | pinesol | News from commits: LP#2038472: Revive DYM optimizations <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=7f72e22bc23ae887fefd98074a7415e852fcb4c0> |
13:44 | Dyrcona | I knew that I should have just started my test program in a tmux session and installed updates after. |
13:45 | Dyrcona | The test was not related to JIT, but to checking in and deleting copies from a spreadsheet. |
13:51 | Dyrcona | VPN is definitely down. Wonder if the site also lost Internet? |
13:53 | Dyrcona | I suppose I could test the changes to command line options on a local vm. |
14:11 | jihpringle joined #evergreen | |
14:12 | Dyrcona | Well, that looks like it will work. Guess I can test it on the training server. |
14:16 | csharp_ | @decide open-ils.serial or open-ils.cereal |
14:16 | pinesol | csharp_: go with open-ils.cereal |
14:16 | csharp_ | a nutritious part of this complete breakfast! |
12:33 | Dyrcona joined #evergreen | |
12:50 | collum joined #evergreen | |
12:57 | Dyrcona joined #evergreen | |
13:15 | eeevil | sandbergja: I'm pointing this in your general direction because you've been AWESOME about pushing for more test automation ... do you have thoughts about where we might want to stick the following test? `(cd Open-ILS/examples && xmllint --schema fm_IDL.xsd --noout fm_IDL.xml)` |
13:19 | sandbergja | eeevil: ooh, good question! Maybe we could lightly wrap it in a perl test? Like run that command in backticks, and confirm that the exit code is 0? |
13:20 | sandbergja | or something along those lines |
13:21 | eeevil | maybe as part of `make check` directly? I don't have strong feelings about it, but would prefer it yell at us easily/quickly |
13:22 | sandbergja | yes, +1 to yelling :-) |
13:23 | sandbergja | example of the perl approach in the first subtest here: https://git.evergreen-ils.org/?p=working/Evergreen.git;a=commitdiff;h=2f264c6ad779bacef4dd910bc63ece0174f82083 (the ok($output, 'runs without error');) |
13:24 | Dyrcona | if you only want the return value, system() would be preferable to backticks, and if you care about the output, I'd prefer a pipe open, but that's just my Perl style.... |
13:25 | sandbergja | Dyrcona++ # thanks! That is good to know. |
13:27 | Dyrcona | Just running it from 'make check' would also be good. |
13:49 | terranm | Bmagic or anyone - I'm trying to set one of my test servers to have the Czech option in the staff client and I can't figure out what I'm missing. These are the set up instructions I have so far: https://wiki.evergreen-ils.org/doku.php?id=newdevs:i18n |
13:49 | terranm | Of course, now it is working |
13:50 | terranm | No it's not. Argh |
13:52 | terranm | at https://terran-testbox.gapines.org/eg/staff - Czech is working in the OPAC. On the eg login page the Czech option is appearing in the top menu, but when logged in and on an eg2 page it's not there |
13:55 | jvwoolf left #evergreen | |
14:30 | jvwoolf joined #evergreen | |
14:33 | jihpringle joined #evergreen |
17:35 | jihpringle | I think good to have it pulled out as its own bugs, since both of us initially missed it within that bug |
17:35 | Bmagic | I agree |
17:47 | terranm | Claiming 1397 |
17:51 | pinesol | News from commits: LP1965985 Empty alt text for OPAC decorative images <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=b74e76e25a0856674348bfc970858cf8b6d115df> |
17:51 | pinesol | News from commits: LP2040314 BOOPAC 'send test email' link keyboard support <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=f15cb15e87518229e386cd7c28a767dba9d906e7> |
17:51 | pinesol | News from commits: LP2040314 OPAC 'send test email' link keyboard support <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=c75831ebb0f138ea4068f1aaa985946f5e8c94c8> |
18:04 | kmlussier left #evergreen | |
18:21 | pinesol | News from commits: LP2039750 Stamp upgrade script <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=ba54e37d253058ce5bc446533d9115ffaf72361e> |
18:21 | pinesol | News from commits: LP#2039750 Save Stat Cat Grid Settings <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=5bef90979832bd4fe9250fea0a367c7ecba28d9b> |
19:59 | jihpringle joined #evergreen |
13:41 | dguarrac | You're welcome! I used to have to use it a whole bunch in a past life. |
14:24 | kmlussier left #evergreen | |
14:58 | mmorgan1 joined #evergreen | |
15:11 | pinesol | News from commits: LP#2043437: stamp DB update <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=b40431111fd7f9bb20311788b3dc32b51d0fc90d> |
15:11 | pinesol | News from commits: Revert "lp1895699: search results page lets user know that course materials filter... <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=ead4f82da75728b459d12f595bd1bab5d2da1990> |
15:11 | pinesol | News from commits: lp2043437: fix flakey test <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=00003d7aa91bdf39272891da9f7c8e80aab24fb1> |
15:11 | pinesol | News from commits: lp2043437: add missing import to CatalogModule <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=ee6cb432f8d9eb1f8cc0b87665826c74a8d14dd4> |
15:17 | mantis1 left #evergreen | |
15:36 | Dyrcona | Well, one thing this project has revealed is that we should probably remove all of the 998 fields from our records. |
15:50 | Dyrcona | Interesting how MarcStreamReader and MarcPermissiveStreamReader throw different exceptions, even if you tell the latter not to be permissive. |
12:12 | Dyrcona | That's at 11:24 EST. :) |
12:13 | Dyrcona | I've already done this for a set of records that evergreen-universe-rs doesn't like. |
12:13 | berick | oh, gotcha |
12:15 | Dyrcona | I think I'll put this down for now and work on a program to convert some marcxml to binary marc to see if CPAN's RT will let me upload that. I get a 403 when I try to upload the marcxml examples from the Rust test. |
12:15 | Dyrcona | "That should only take half an hour," he said knowing it was very likely to be a lie. |
12:16 | Dyrcona | Also, mexican_coca-cola++ It tastes so much better with cane sugar than with HFCS. |
12:22 | Dyrcona | What? libmarc-perl does not install MARC::File and friends? I thought that it did. |
12:35 | Dyrcona | I'm just full of complaints today, aren't I? |
12:44 | Bmagic | you? never! |
12:44 | Dyrcona | I'm installing MARC::File::XML with cpan set to local::lib, and there sure are a lot of prerequisites. |
12:50 | Dyrcona | Failed 11/11 test programs. 3/5 subtests failed. |
12:50 | Dyrcona | Right. I'll just run it on a server where this is already installed. |
12:51 | Dyrcona | And, I'll wipe out the stuff that CPAN installed locally. |
13:08 | Dyrcona | Looks like I may have to reboot. I just swapped monitors and the laptop doesn't see the new one. |
13:33 | Dyrcona | Apparently, all of the software in the world has chosen this week to hate me: https://rt.cpan.org/Ticket/Display.html?id=150348&results=a5d68555ff4b4354e65ce6ec51f76634 # Read to the bottom... |
13:37 | Dyrcona | gmcharlt: RT on CPAN is apparently broken for uploads at the moment. I've tried 3 times to add a file of records to that ticket above. |
13:39 | Dyrcona | Stompro++ I'll give MARC::Lint a whirl. |
13:57 | Dyrcona | Stompro: It looks like MARC::Lint may help. I'm running a test program already. |
13:59 | Stompro | I wonder if it will be too verbose, or if you can pick out the bigger issues. I'm curious how it performs also? |
13:59 | Dyrcona | And, maybe not so much: is_valid_checksum: Didn't get object! at /usr/share/perl5/Business/ISBN.pm line 481, <DATA> line 244. |
13:59 | Dyrcona | Well, it gets totally clobbered by our data after bib id 233519. |
14:58 | Dyrcona | Google sheets is 10 million cells, roughly speaking. Maximum column is zzz (or 18.278). I only have 2 columns. |
15:00 | Dyrcona | LibreOffice Calc has a limit of 1,024 columns per row. Maximum number of rows is just over 1 billion. I hope I don't have to worry about the 32,767 characters per cell limit. |
15:01 | Dyrcona | So as long as we stick with Google Sheets and LibreOffice, we should be OK. I doubt that I'll hit 10 million cells. It may go over 1 million rows. |
15:01 | Dyrcona | Right, enough blather. I'll test the version that writes the CSV files. |
15:03 | * Dyrcona | tries hot swapping monitors again. If I disappear, I had to reboot. |
15:10 | Dyrcona | Well, looks like I have to reboot. |
15:16 | Dyrcona joined #evergreen |
10:27 | pinesol | News from commits: DOCS: LP#1871211 Follow-up eg_vhost.conf <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=ded2dd7815a9d3bcf0305c1b55dd53ee3f7ae4f4> |
10:27 | Dyrcona | It probably should. |
10:29 | sandbergja joined #evergreen | |
10:30 | sandbergja | Dyrcona++ # thanks for taking a look at that live test |
10:32 | berick | sandbergja++ # branches, redis, and more! |
10:32 | Bmagic | sandbergja, abneiman, terranm, colum : Would you mind if I back ported this Docs commit to rel_3_12? ded2dd7815a9d3bcf0305c1b55dd53ee3f7ae4f4 |
10:33 | sandbergja | +1 from me |
11:18 | berick | hm, i thought we did decide to make it default for post-3.12 main. |
11:18 | berick | i could be misremembering |
11:19 | berick | in part, i think, since it would only affect new installs |
11:20 | eeevil | I thought we'd agreed that we should see what things look like after some more testing and review. |
11:20 | Dyrcona | I'm not sure what we agreed at this point. Maybe we should have that discussion on the list? |
11:20 | eeevil | where will there be more new installs than for the purpose of dev, though? |
11:21 | Dyrcona | The dev meeting is also today. |
11:37 | eeevil | since I came to complain ;) , I'm happy to pick them into main. are there objections? |
11:38 | eeevil | man... my fingers are not moving in the right order today. rel_3_12 |
11:41 | eeevil | it's close to lunch time, so I'll leave the question open for a while. but, ping Dyrcona and sandbergja (in particular) |
11:52 | csharp_ | eeevil: I'm good for what my opinion is worth |
11:55 | csharp_ | I've lightly tested it on my puny server |
11:55 | csharp_ | @band add Irregular Expressions |
11:55 | pinesol | csharp_: Band 'Irregular Expressions' added to list |
12:02 | sleary | can I still target bug fixes to 3.12 or is that bad manners at this point? |
12:03 | eeevil | csharp_: it's worth One Full Commit Bit, as it happens, sir! |
15:04 | Bmagic | no problem, carrying forward |
15:04 | Bmagic | #action mmorgan will explore moving LP stats to community site and automating same |
15:04 | Bmagic | #info sandbergja will write tutorial: "Do a database call (Galen’s cat counter)"#info sandbergja will write tutorial: "Do a database call (Galen’s cat counter)" |
15:04 | Bmagic | #info sandbergja will go over the Nightwatch test reorg with folks at the Monday at 2pm ET meeting or another time as available |
15:05 | Bmagic | go ahead sandbergja |
15:05 | sandbergja | kinda fidddling with my partial draft for the tutorial |
15:05 | sandbergja | probably need to check back with me next time :-) |
15:05 | Bmagic | no problem |
15:05 | sleary | sandbergja I looked at what you sent me a while back and it's looking great |
15:05 | sandbergja | we didn't get around to moving the nightwatch tests, but we got them working! |
15:05 | sandbergja | sleary++ # thanks for the review! |
15:05 | terranm | sandbergja++ |
15:05 | Bmagic | #action sandbergja will write tutorial: "Do a database call (Galen’s cat counter)" |
15:05 | smayo | sandbergja++ |
15:08 | Bmagic | I refreshed and I see new stuff |
15:08 | abneiman | and terranm++ sandbergja++ and mmorgan++ for many lovely merges |
15:08 | Bmagic | #link https://wiki.evergreen-ils.org/doku.php?id=dev:bug_squashing:2023-11 |
15:09 | abneiman | though there is a test conflict to talk about below in the agenda |
15:09 | Bmagic | terranm++ sandbergja++ mmorgan++ |
15:09 | terranm | 3.12 (currently) has an even 100 patches committed |
15:09 | sleary | terranm++ sandbergja++ mmorgan++ |
15:33 | Bmagic | sandbergja++ sleary++ |
15:33 | sleary | sandbergja++ # thanks for staying on top of linting! |
15:33 | Bmagic | thanks again sandbergja! next.... |
15:33 | Bmagic | #topic New Business - Test failures, including at least one critical regression (bug 2043437) - Jane |
15:33 | pinesol | Launchpad bug 2043437 in Evergreen "Three test failures on rel_3_12 and main" [Critical,New] https://launchpad.net/bugs/2043437 |
15:34 | Bmagic | #link https://bugs.launchpad.net/evergreen/+bug/2043437 |
15:34 | sandbergja | oh god, I have a bunch in a row :-D |
15:34 | Bmagic | yes, yes you do |
15:34 | sandbergja | we have 3 failing tests, one of which points to a major problem |
15:34 | sandbergja | tests++ # catching that before we released it to users! |
15:35 | sandbergja | specifically, the holdings view doesn't load (maybe just a missing import or something) |
15:35 | terranm | sandbergja++ tests++ |
15:35 | sandbergja | I feel pretty strongly we should take care of those before building a beta. |
15:35 | mmorgan | +1 |
15:35 | sandbergja | But I don't know that I'll have much time to look into them |
15:36 | sandbergja | Dyrcona already started looking at the perl one, and posted some notes |
15:37 | Dyrcona | When I said that I don't know how to fix the syntax error, I should have said that it's not obvious to me what's wrong. |
15:37 | Bmagic | the course reserves issue is fine because the test is bad, so we're looking at the holdingsView.spec.ts issue, and the query issue |
15:37 | sleary | we should fix the bad test since the problem is obvious, but yes |
15:38 | Bmagic | agreed on fixing the test. Should each of the three things be it's own bug so folks can claim them? |
15:38 | Dyrcona | Well, we could use a collab branch to avoid separate bugs, but I'll let the consensus decide. |
15:40 | Bmagic | I assume this pause is because everyone is reading the bug |
15:40 | sandbergja | :-) |
15:41 | Bmagic | Dyrcona: that query works on two of my test systems |
15:42 | Dyrcona | Bmagic: It blew up for me on a 3.12 vm with stock concerto and Pg 15. I tried the function by itself with different parameters. |
15:43 | Bmagic | Just the query? Not integrated in Evergreen? |
15:43 | Dyrcona | Different parameters, I mean interger array and string that should have matched. |
15:43 | Dyrcona | Just the function by itself, as well as the query I pasted. |
15:43 | Dyrcona | The query comes from the error output of the test. |
15:44 | JBoyer | Fun thing: it worked for me on eg3.11 / pg15 and broke on 3.11 / pg14. |
15:44 | Bmagic | ok, yes, it breaks on newer versions of the database |
15:44 | Bmagic | bugsquash machine throws the error |
15:46 | Bmagic | let's move this to post-meeting |
15:46 | Dyrcona | JBoyer: Gotcha. |
15:46 | JBoyer | Bmagic, +1 |
15:47 | Bmagic | #topic New Business - How can we get computers running our tests regularly again? - Jane |
15:47 | eeevil | I'll also look at the search one, later |
15:47 | sandbergja | #info for anybody wanting to run the tests, or try it out: https://wiki.evergreen-ils.org/doku.php?id=dev:contributing:qa#types_of_tests |
15:47 | Bmagic | sandbergja++ |
15:47 | mmorgan | sandbergja++ |
15:48 | sandbergja | I am just feeling fired up about tests, and wanted to see if there's capacity for getting buildbot running them automatically for us, or some new solution |
15:48 | Bmagic | sandbergja: where's the buildbot now? (I've never known where that lives and who's in charge of it) |
15:49 | shulabear | sandbergja++ |
15:49 | sandbergja | no clue. Was it an EOLI server? |
15:52 | sandbergja | the ng lint always passes, for reasons mentioned above hahaha |
15:52 | Bmagic | haha |
15:52 | Bmagic | not sure if we've arrived at anything I can put down as action |
15:53 | sandbergja | I can investigate getting more tests into gh actions, if there aren't concerns with tying ourselves more to that platform |
15:53 | Bmagic | #action sandbergja will investigate getting more tests into gh actions |
15:53 | JBoyer | It doesn't have to be the projects definitive home to provide a useful function, even if temporarily, |
15:54 | Bmagic | almost out of time |
15:54 | Bmagic | #topic Keep an eye out for Angular 17 / Bootstrap 5.3 upgrade blockers and note them on bug 2043490 - Stephanie |
15:54 | pinesol | Launchpad bug 2043490 in Evergreen "Angular 17 + Bootstrap 5.3 Upgrade" [Wishlist,New] https://launchpad.net/bugs/2043490 |
15:54 | kmlussier | sandbergja++ tests++ |
15:54 | JBoyer | (so long as it's still easy to run tests locally) |
15:54 | Bmagic | #link https://bugs.launchpad.net/evergreen/+bug/2043490 |
15:54 | Bmagic | sleary: go for it |
15:54 | sleary | ah! So, I went ahead and opened a bug for the next big Angular/Bootstrap upgrade, which should be less painful than the last one |
15:55 | sleary | I haven't looked too closely into what's involved on the Angular side, so I wanted to ask you all to keep an eye on that as you skim your news, and add comments to that bug if you find any potential blockers other than the ng-bootstrap accordion issue listed |
15:55 | sandbergja | sleary++ |
15:55 | smayo joined #evergreen | |
15:55 | eeevil | #info I've requested we keep XMPP as the default OpenSRF transport in EG main for the time being. There's no redis release of OpenSRF yet, so support is only in a side branch, and having redis be the default will make dev (especially backport-focused dev, like bug fixes) more painful because you can't just switch branches and test the code. Also, I'm not convinced that it's deeply understood by more folks than berick, and that puts a lot of pressure |
15:55 | eeevil | on him to Fix All The Things if Things need Fixing. I'm asking here for any strong objections to applying the 2 existing commits that will make that so, as it is now for rel_3_12. Barring any, I'll pick those commits into main and life will be a little simpler for all the not-testing-redis cases, for now. |
15:56 | eeevil | (separately, I think I know where the search test failure is coming from, and I'll poke at it early tomorrow) |
15:57 | Bmagic | eeevil: and when we've all installed and tested redis, then make it default? |
15:58 | eeevil | Bmagic: well, and when more-than-Bill can help work on it, ideally, but yes. "when it and we are ready" |
15:58 | eeevil | it's not something we should force Right Now, IMNSHO. but, hopefully, soon |
15:58 | eeevil | for a definition of soon somewhere between "months" and "geologic time" |
16:06 | jihpringle joined #evergreen | |
16:06 | shulabear | bmagic++ |
16:10 | Bmagic | Dycona: that's strange. Worked for me on 3.9.1 and 3.11.1, but not busquash main |
16:11 | Dyrcona | Bmagic: Are you talking about the function itself or the search test? I suspect the function has been broken since 0940, but doesn't look like it is used much. |
16:12 | Bmagic | copy/paste the query I mean |
16:13 | Dyrcona | I get a syntax error any time/any where I run the function. |
16:14 | JBoyer | yeah, query_int_wrapper does so little I'm surprised we even have it (syntax escaping simplicity if I had to guess) but sometimes passing '()' as the second param is accepted and sometimes it's a syntax error, which is weird. |
16:20 | Dyrcona | select query_int_wrapper(vlist, '()') from metabib.record_attr_vector_list where source =2; <- Blows up for me, even though record 2 is in metabib.record_attr_vector_list |
16:20 | Dyrcona | Can we just replace it with the code from the function? |
16:22 | Dyrcona | select vlist @@ '()'::query_int from metabib.record_attr_vector_list where source = 2; Also produces a syntax error at the text value '() |
16:23 | eeevil | I wish I weren't orange right now ... TL;DR: no, we can't (without testing that we don't need it anymore -- and there were reasons) |
16:24 | Dyrcona | eeevil: It's used in only 1 place AFAICT. |
16:24 | Dyrcona | I've tried a few variations: '{}' also fails but null works. |
16:25 | eeevil | Dyrcona: yes, and that 1 place is important. more later, though |
16:45 | eeevil | I'm transforming back into pure eeevil, leaving my pumpkin state behind! |
16:45 | eeevil | and, https://bugs.launchpad.net/evergreen/+bug/1438136/comments/12 |
16:45 | pinesol | Launchpad bug 1438136 in Evergreen 2.8 "OPAC searching significantly slowed by adding format filters" [High,Fix released] |
16:45 | eeevil | that's why we have query_int_wrapper, and what we need to test before we stop trying to use it. |
16:45 | eeevil | HOWERVER |
16:46 | eeevil | that's not the problem here, regardless. the problem here is that we should never be sending an effectively-empty query_int (that is, and int array query "thing") to the database |
16:47 | eeevil | that clause is a question of the data, but the question being asked is empty in this case. we need to elide the clause altogether |
16:48 | eeevil | which we normally do (you don't always see a query_int clause either direct or via query_int_wrapper), but something is convincing us to generate an empty query_int at the perl level |
16:49 | eeevil | I suspect an interaction with the new on_reserve() filter, but mostly because it's new and not because I can point to a problem with it |
16:54 | Dyrcona | Well, that's somewhere to start. I was thinking of resorting to git bisect. I can take a look tomorrow morning, probably. |
16:55 | Dyrcona | Anyway.... I'll turn into a pumpkin now..... |
17:11 | kmlussier left #evergreen |
13:16 | degraafk joined #evergreen | |
13:18 | sandbergja joined #evergreen | |
13:34 | sleary joined #evergreen | |
15:20 | pinesol | News from commits: Fix test failure <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=a82a28788090fb31a00b0a15f3c3474f57a2c939> |
15:50 | pinesol | News from commits: LP1754364: Show timezone combobox in library settings editor <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=687c403fcd29cbd11b891fb91a681ba9ffc8bb56> |
17:21 | pinesol | News from commits: LP192012 Localize staff catalog search placeholders <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=8481ad5e6c8b20c5158b67e964ea479040e86a9a> |
19:03 | sleary joined #evergreen |
09:24 | jblyberg7 joined #evergreen | |
11:12 | jeffdavis joined #evergreen | |
11:33 | sandbergja joined #evergreen | |
11:34 | sandbergja | One of the live tests is failing on main. Turns out, the problem was that I had committed the wrong patch to main, thereby causing the failure. |
11:34 | sandbergja | Revert commit coming up |
11:35 | sandbergja | tests++ |
11:46 | pinesol | News from commits: LP#2009891: add item author to open-ils.actor.carousel.get_contents <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=dd778b5060fff492b1cdb83f3fa91ea7d3eb61ed> |
11:47 | pinesol | News from commits: Revert "LP#2009891: add item author to open-ils.actor.carousel.get_contents" <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=301fdb85adb5078e2aaaa128e679b8e36e3717a3> |
15:28 | sleary joined #evergreen | |
18:48 | pinesol | News from commits: LP2041364: Fix a mistake in the marc_export error reporting code <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=11a4c120352bc63642edcee003b3d2816ea074db> |
18:48 | pinesol | News from commits: LP2041364: Refactor marc_export cursor code <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=b2091fa8ffeab3598a4bd16370d4f88e94ae4952> |
18:48 | pinesol | News from commits: LP#2041364 - Use a cursor to load 10000 bibs at a time. <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=ac211b0f450177be8d3b27bcc565cad0ae63ecf9> |
18:48 | pinesol | News from commits: LP2041364: Attempt to speed up 852 insertion in marc_export <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=71b7e501a305a495b374d6971b185c02aae43672> |
19:42 | jblyberg7 joined #evergreen | |
19:48 | pinesol | News from commits: LP2012402 follow-up: let tests know about PcrudService constructor's new API <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=532315227a16223624eebc333e6ebb12b0d66be2> |
19:48 | pinesol | News from commits: LP2012402 Authoritative APIs are optional; disabled by default <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=db09291462c6102c83d4fc4eb59593d8225c9cd8> |
20:28 | ejk joined #evergreen | |
21:50 | eglogbot joined #evergreen | |
21:50 | Topic for #evergreen is now Welcome to #evergreen (https://evergreen-ils.org). This channel is publicly logged. Logs for today: http://irc.evergreen-ils.org/evergreen/today |
11:38 | terranm2 joined #evergreen | |
11:43 | pinesol | News from commits: LP2040186 Survey Q&A button colors and input labels <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=4edae0c4be27d2e4adb822542696ad546b9c02e3> |
12:13 | pinesol | News from commits: LP2037685 - Staff Catalog: Default Search and Preferred Library settings are deleted... <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=bcbb33a5fa97aff70f81f27cd16b0e18144323f6> |
12:18 | abneiman | anyone have some test love to show to https://bugs.launchpad.net/evergreen/+bug/1999158 ?? It's a needed a11y fix and is currently on https://terran-main.gapines.org/ |
12:18 | pinesol | Launchpad bug 1999158 in Evergreen 3.11 "Labels for eg-org-select combobox" [Medium,Confirmed] |
12:33 | eby joined #evergreen | |
14:05 | terranm joined #evergreen | |
15:00 | sandbergja joined #evergreen | |
15:02 | JBoyer | terranm++ This week really didn't end up working out for me at all but it looks like some good stuff has been done. |
16:04 | terranm | sandbergja and I are both working on getting more of the tested things committed too, so I think 3.12 is going to have a lot of stuff in it |
16:27 | abneiman | terranm++ sandbergja++ |
16:27 | abneiman | I've also put bugs (pun absolutely intended) in some other committer ears, too |
16:38 | sleary | terranm do you know if bug 1934018 is still on one of the testing servers? I think it was on the spreadsheet but got lost in the shuffle |
16:38 | pinesol | Launchpad bug 1934018 in Evergreen "Angular Staff Catalogue: More Link Missing from Facets" [High,Confirmed] https://launchpad.net/bugs/1934018 |
16:39 | jweston joined #evergreen | |
16:41 | terranm | sleary no, it's not on one of them right now |
16:44 | pinesol | News from commits: LP2035535: Add accessibility tests to nightwatch <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=74a240914ca86fa06d169b3928e57352156b7171> |
16:44 | pinesol | News from commits: LP#2033067 The "prev" and "next" navigation buttons in carousels are not translated. <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=632aca33848e7415d5b008c000c8cc432bbd6e54> |
16:44 | pinesol | News from commits: LP1908568: double-click item to open editor <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=8fba1be55ba291c2deb44604ed0fb8dc2539875f> |
17:00 | berick | grabbing 1390 |
17:14 | pinesol | News from commits: LP#2035389: Don't use locg to determine whether course reserves links display in... <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=e4288874c5ba6bbf6004f8fba038549d4d58d511> |
17:14 | pinesol | News from commits: LP#2037128: release note <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=f5ea0711e36254d03bfab8bb49f141320192f8e4> |
10:32 | terranm | redavis++ cool beans, thx |
10:41 | redavis | mmorgan, looking at https://bugs.launchpad.net/evergreen/+bug/1956241. In the patch, it says it adds it maps the missing permission to the concerto circ administrator and circulator permission groups. I'm only seeing it in the circulator permission group. BUT I think the main goal of the patch is to add the permissions. If I'm understanding that correctly, I think it's ready for a sign off. |
10:41 | pinesol | Launchpad bug 1956241 in Evergreen "Hold Groups: Staff Users Cannot View Patron Hold Groups" [Medium,Confirmed] |
10:46 | mmorgan | redavis: Yes, the goal was to add the permission to some permission groups that made sense, esp for testing. Not sure why the circ admin didn't get them. |
10:48 | redavis | ++, sign off added |
10:49 | redavis | Though Elizabeth Davis (not to be confused with Ruth Elizabeth Davis) did testing before me. |
10:49 | mmorgan | redavis++ |
11:03 | briank joined #evergreen | |
11:03 | Dyrcona joined #evergreen | |
11:10 | redavis | woohoo! Used the Axe DevTools and was able to run an accessibility test! |
11:10 | terranm | I noted in my comment that the seed data SQL and the upgrade SQL aren't consistent. It doesn't prevent it from working properly in production, but could be a little confusing for test servers. |
11:10 | redavis | terranm+ |
11:10 | redavis | terranm++ also |
11:16 | Rogan joined #evergreen | |
12:13 | jblyberg7 joined #evergreen | |
12:21 | Dyrcona | Y'know what? I'm done with captchas. If a site asks me to solve a puzzle or some crap, I'm going away and not coming back. So, how is this relevant here? Well, it means no bug report on MARC::Records' RT. Well, let's see what happens if I sign in, first. |
12:24 | Dyrcona | It wants you to solve a puzzle to logout of "guest." I know they've had issues with spam, but I'm so sick of "proving" that I'm human. (Also, turns out that bots are better at solving some of the puzzles than humans, so they consider failure to be success, too.) |
12:31 | pinesol | News from commits: LP2039612: regression test for creating carousels <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=b418f2dfa8b9ec16d132b74c25c240480d2b5ed0> |
12:31 | pinesol | News from commits: LP2039612: Fix Carousel create / edit <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=00bd4c88d60fedaf1a39a7d34a24f0c2e2bd2c00> |
12:34 | collum joined #evergreen | |
12:39 | redavis | terranm, have signed off on several of the accessibility LP tickets on terran-main.gapines.org. Gonna peace out for while. Maybe the day. Back at it tomorrow as time allows. |
12:52 | collum joined #evergreen |
09:57 | pinesol | News from commits: LP#2009891: add carousel default name to open-ils.actor.carousel.retrieve_by_org <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=0b2561df57fbed32351e07d883370358bc3a9314> |
09:57 | pinesol | News from commits: LP#2009891: add item author to open-ils.actor.carousel.get_contents <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=79080a199753a2240767d5ff8036af2fa18445a4> |
10:01 | sleary joined #evergreen | |
10:20 | terranm | Bmagic: Do you have reports running on your test server? |
10:22 | smayo joined #evergreen | |
10:51 | smorrison joined #evergreen | |
10:55 | briank joined #evergreen | |
12:18 | jeffdavis | we've been on PG14 in production for >6 months, it's just fine |
12:19 | berick | jeffdavis++ |
12:19 | Dyrcona | jeffdavis++ |
12:19 | Dyrcona | I've been testing production data with Pg 15, but not full time. |
12:19 | Dyrcona | I still run Pg 10 most of the time. |
12:20 | jeffdavis | we had to do a bit of cleanup prior to our Postgres upgrade because our EG install is so old, but nothing too arduous and we don't have any ongoing issues related to PG versions |
12:21 | Dyrcona | I can upgrade from Pg 10 to Pg 15 simply by applying the Evergreen db upgrades and doing a pg_dump and pg_restore. We have applied most of the necessary updates already. |
12:22 | Dyrcona | Hm. Maybe I'll swap Pg 10 and Pg 15 on my test system and just use Pg 10 to do the needed updates.... |
12:23 | Dyrcona | Yeah, I'll set that up this afternoon. |
12:24 | Dyrcona | Someone should open a Lp about Pg 10 and 11, and possibly Pg 16. |
12:34 | Dyrcona | jeffdavis: Have you seen the bug about disabling JIT in Pg? Lp 2042158 |
13:36 | jihpringle | the only common thing we could find with the holds that weren't showing was that it was the older holds |
13:52 | jonadab | Item-level holds have a number of gotchas, yeah. The capability is useful for a variety of technical reasons, but it almost ought to be locked behind a "Yes I want this specific copy" confirmation or something. |
14:26 | mmorgan | mantis1: By "item" do you actually mean item hold? That is hold_type = C? We have seen visibility oddities with metarecord holds (hold_type = M). |
14:26 | mantis1 | mmorgan: yes |
14:26 | mantis1 | I did try replicating the issue on a test server but it does work as 'intended' |
14:26 | mantis1 | I let them know just to do bib level for now |
14:26 | mantis1 | just baffled by what could cause this specifically |
14:35 | jeff | if you look at the hold in question, assuming it is hold_type = 'C', does the hold's target point to a valid, non-deleted copy in asset.copy? Does the hold have an entry in reporter.hold_request_record? |
14:42 | mantis1 | jeff: ah good question I will check |
14:45 | mantis1 | jeff; yes to all |
08:32 | Dyrcona joined #evergreen | |
08:34 | mmorgan joined #evergreen | |
08:49 | collum joined #evergreen | |
08:52 | Stompro | JBoyer, morning. My testing of NCIP AcceptItem and RequestItem with ToAgencyID set to our system level OU seems like it works ok. As long as <PickupLocation> is set in the AcceptItem message at least. I'll send you my notes. |
08:54 | JBoyer | Stompro ++ |
09:03 | sandbergja joined #evergreen | |
09:04 | sleary joined #evergreen | |
15:21 | berick | that would be ideal for the 2-hackers-on-the-same-keyboard setup |
15:21 | berick | just slappin keys |
15:28 | Stompro | Can't just add "use OpenSRF::Utils::Logger;" "Bareword "OpenSRF::Utils::Logger" not allowed while "strict subs" " |
15:39 | JBoyer | Stompro, when I've wanted this I just used the raw perl syslog module, hard-code the various params, and just dropped a log line with the whole incoming and outgoing xml messages for testing. I don't use it in prod because SHAREit does send passwords even though we don't check them. |
15:39 | Dyrcona | JBoyer++ |
15:42 | Stompro | JBoyer, thanks, I just want to add some print statements to figure out why certain code isn't working correctly. I don't actually need the messages right now. |
15:43 | JBoyer | Ah, fair. But the end result is the same, sprinkle syslog() calls wherever you want to shine a light. |
12:24 | Stompro | Dyrcona, what are you seeing now? |
12:41 | Dyrcona | Stompro: It's just that one of the exports also tried to pull in consortium-wide URIs. I think everything is OK. I've installed marc_export with the experimental patches to production. |
12:42 | Dyrcona | So, I meant that maybe the grep was not causing issues, but that doesn't explain the other library, so I guess it was really the problem. |
12:43 | Dyrcona | I'm getting similar numbers in my test environment as production at this point, so that's why I'm throwing it over there. |
12:44 | berick | rusty bits for the redis changes from above https://github.com/kcls/evergreen-universe-rs/tree/working/opensrf-bus-addrs-username |
12:45 | Dyrcona | berick++ # I was going to ask if I should update my development system for the latest OpenSRF/Redis stuff. Maybe later today..... |
12:46 | Dyrcona | Stompro: I'm about to add this to Lp 2015484: https://git.evergreen-ils.org/?p=working/Evergreen.git;a=shortlog;h=refs/heads/collab/dyrcona/lp2015484_exclude_opac_visible_from_export-rebase |
12:46 | pinesol | Launchpad bug 2015484 in Evergreen "marc_export: provide way to exclude OPAC-suppressed items" [Wishlist,Confirmed] https://launchpad.net/bugs/2015484 - Assigned to Jason Stephenson (jstephenson) |
12:46 | Dyrcona | That's what I'm testing and have installed in production. |
12:48 | Stompro | Dyrcona, ahh, I thought you meant the @orgs lookups with grep, from the performance branch. |
12:52 | Dyrcona | I was talking about the extra grep in the exclude hidden items branch. |
12:52 | Dyrcona | If its any consolation, I had to sit for a minute earlier and rest my brain to focus on one thing at a time. :) |
12:58 | Stompro | Dyrcona, understood. |
12:59 | jeffdavis | eeevil++ Bmagic++ # jit_above_cost = -1 fixed the thing on our test server also - I would never have figured that out! |
13:12 | collum joined #evergreen | |
13:19 | Stompro | JBoyer++, thank you for helping me understand the possible roadblocks with NCIPServer and ReShare. Sorry about all the questions. |
13:20 | Stompro | Dyrcona has explained to me before about how varied NCIP implementations can be. |
14:39 | pinesol | News from commits: LP1965326 Hatch Printer Settings Port Release Notes; Lint. <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=2c707d83d6625137e7a89a7da5035b1bd84abd9f> |
14:39 | pinesol | News from commits: LP1965326 Move Hatch Printing to Printer Settings <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=ac97129d9adb5a36026fc25cb44ba4debc56ddc3> |
14:39 | pinesol | News from commits: LP1965326 Printer Settings Angular Port <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=b01b14fed8cee098b9911f24f4d656003f7b0f2a> |
15:07 | Stompro | JBoyer, if you haven't started replying to my last NCIP email... ignore it for now. I'll test out the AcceptItem and RequestItem requests with AgencyID set to our system shortname to see how it behaves in real life. |
15:08 | JBoyer | Stompro++ it was on the list but I'll set it aside now. |
15:32 | jihpringle joined #evergreen | |
16:02 | sandbergja joined #evergreen |
10:43 | BAMkubasa joined #evergreen | |
10:50 | mantis1 joined #evergreen | |
10:55 | briank joined #evergreen | |
11:15 | Dyrcona | OK. I have the latest Rust marc-export test running. I'll see how long this takes. |
11:27 | jihpringle joined #evergreen | |
12:14 | Christineb joined #evergreen | |
12:18 | eeevil | berick: drive by question (from immediately-post hack-a-way) incoming! meetings will probably keep me away from here for the rest of the day, but want to get it to you asap for RediSRF(?) goodness |
12:20 | eeevil | berick: I'm looking at opensrf/redis anew via the collab branches. I have two questions that jump out at me: is the --reset-message-bus documented beyond "you must do this at redis restart" somewhere? (the commit it comes in with isn't obvious to me; also, I want to help remove any instances of boilerplate stuff that we can automate or fold into other actions as early as possible -- ideally before merge -- and I know this is under discussion or at |
12:20 | eeevil | least being thought about); and, while the config and password list files don't require it themselves, it /looks/ like the usernames are pinned as "opensrf", "router", and "gateway" in the code that implements each part. is that correct and do you see any reason we /can't/ replace those strings with the value pulled from opensrf_core.xml as with ejabberd? the result being ACL-separated user sets that share a redis instance but can't talk to each |
12:20 | eeevil | other or guess each other's user and queue names. |
12:32 | Dyrcona | eeevil: berick said earlier that he modified osrf_control to do --reset-message-bus as required, so its not necessary now. I'm testing that and other updates. That's a good question about the names. (I generally don't change them.) |
12:33 | Bmagic | I'm troubleshooting a database issue. explain analyze on an insert into biblio.record_entry results in 6+ seconds. The analyze shows me that it spends the majority of the time in reporter.simple_rec_trigger(). Which is odd, because if it's an INSERT, it should just skip down to RETURN. Weird right? |
12:33 | Dyrcona | Bmagic: No. It still builds the rmsr entry. |
12:34 | Dyrcona | Wait until you update an authority with hundreds of bibs "attached." :) |
12:35 | Bmagic | this is EG 3.11.1 with queued ingest. I updated the global flag to turn off queued ingest, and tested it. and visa versa. it's 13 seconds for an insert with it turned off, and 6 seconds with it turned on. But still, 6 seconds seems insane for an insert on biblio.record_entry, especially if we're deferring all of the metabib stuff for the ingester process to come later |
12:35 | Dyrcona | Updating/inserting bibs and auths can be slow because of all the triggers. |
12:36 | Bmagic | I got here because it's breaking the course reserves interface. Taking too long |
12:37 | * Dyrcona | colour me unsurprised. |
16:38 | Bmagic | Maybe I can change PG configuration to eliminate the secondary DB to prove it has something to do with that (or not) |
16:38 | Dyrcona | Nope. None of them checked in. |
16:40 | Dyrcona | Bmagic: Could be. I often have queries fail on the secondary DB, 'cause changed rows from replication are needed. I know that's not your issue, but you shouldn't be inserting into a replicated db, so I'm not sure what you're checking. |
16:40 | Bmagic | This is my test query: explain analyze select * from reporter.old_super_simple_record where id=671579 |
16:41 | Bmagic | 6+ seconds on db1. and less than one second on db2 |
16:42 | Dyrcona | OK. Das stimmt. |
16:43 | Dyrcona | Hm... I need to find the circulations. The copies are all Lost status.... They're also deleted, and I thought using the copy_id would resolve the latter. |
16:51 | Bmagic | Repeated for all tables in metabib |
16:52 | Bmagic | retested, and was dissappointed. Restarted postgres, still dissappointed |
16:52 | Bmagic | disappointed |
16:55 | jeffdavis | Is it consistently ~6 seconds even for other IDs? Like, not 4 or 8 seconds? |
16:59 | jeffdavis | We were having an issue on a test server where calls to open-ils.pcrud.search.circ were consistently taking about 6.5s per circ. I never got to the bottom of it (fortunately we're not seeing that in production). It's a long shot but maybe there is a connection? Don't want to send you down a rabbit hole though! |
17:00 | Dyrcona | There are only 1 or 2 tables involved in that function. Everything else is 4 views that percolate up. |
17:00 | jeffdavis | But the flesh fields for the problematic call include wide_display_entry. |
17:01 | Dyrcona | Which is a view, IIRC. |
17:09 | jeff | out for now, back later! |
17:10 | Bmagic | https://explain.depesz.com/s/Df5A |
17:11 | Bmagic | so that seems to have revealed that it's sequential scanning display_field_map |
17:16 | jeffdavis | I can confirm we're seeing the same behavior on our test server with your example query (explain (analyze, buffers) select * from reporter.old_super_simple_record where id=x) |
17:16 | Bmagic | so weird, two databases, one a replicant of the other. I just synced up those two settings in postgresql.conf and restarted the primary database. The query is still sequentially scanning that table instead of using an index. I run the same query against the secondary database and it's fast |
17:17 | Bmagic | jeffdavis: yous is sequential scanning? |
17:17 | jeffdavis | test db (slow): https://explain.depesz.com/s/pkwY |
17:17 | jeffdavis | prod db (fast): https://explain.depesz.com/s/UWu5 |
17:19 | Bmagic | jeffdavis: interesting. In my case, it's prod that's slow and the secondary database is fast |
17:20 | Bmagic | what's even more interesting, is for your fast case, it's seq scanning display_field_map and apparently that's not taking much time at all |
17:24 | Bmagic | jeffdavis: PG14? |
17:24 | jeffdavis | yes |
17:25 | Bmagic | jeffdavis++ |
17:25 | jeffdavis | The servers are pretty different, the test db is the product of pg_dump/pg_restore rather than replication |
17:25 | Bmagic | yeah, I'm tempted to dump/restore as well |
17:26 | Bmagic | I out of ideas, so that one comes to mind :) |
17:27 | jeffdavis | well, we're seeing it after dump/restore, but who knows, maybe it would help! |
10:03 | Dyrcona | This could be a case where joins are slower than subqueries for .... reasons. |
10:04 | Dyrcona | It could also be completely different on a newer Pg release. |
10:04 | Stompro | I started using NYTProf now, it is much heavier weight, but the output much more detailed. https://metacpan.org/pod/Devel::NYTProf |
10:08 | Stompro | Devel::Profile is much faster for quickly seeing results. NYTProf generates a 2G output file in my testing that then has to be processed into an HTML report. |
10:08 | Dyrcona | I think I'll go for fiaster/lighter weight. I'll read the docs. |
10:09 | Dyrcona | My code to log subroutine names with timestamps as they're called produces a largeish output and is likely less accurate since it spends time generating timestamps and printing them. |
10:10 | Stompro | All I had to do was run marc_export with "perl -d:Profile ./marc_export_testing" and it generates the profile log prof.out |
10:10 | Dyrcona | I'll make a patch and throw it on the LP. |
10:11 | Dyrcona | I mean a patch for my logging code. |
10:11 | Dyrcona | So, you think we should just switch from insert_grouped_field to insert_field? |
10:13 | Stompro | I put a diff of the changes I was testing with at https://gitlab.com/LARL/evergreen-larl/-/snippets/3615366 |
10:14 | Stompro | Put all the 852s in an array, then once they are all added, call insert_grouped_field for the first one to get the same ordering, and use insert_fields_after for the rest with one call. |
10:18 | Dyrcona | Stompro: There's a simpler way to do the insert: push the fields to an array, then do the first one with shift and the rest of the array after that. |
10:19 | Stompro | I figured, my perl array skills need work. :-) |
10:20 | Dyrcona | I wonder if the first one even needs to be grouped? |
10:20 | Dyrcona | I'm going to look at MARC::Record again. |
10:21 | Dyrcona | Stompro++ # For the notes in the snippets. |
10:21 | Stompro | In my test data, the 901 tag would be placed before the 852 without using the insert_grouped_field for the first. |
10:23 | Stompro | I don't think MARC::Record re-orders the fields. |
10:24 | Stompro | If I'm understanding where you are going with that. |
10:33 | Dyrcona | OK. LoC says the records are supposed to be grouped by hundreds, they don't have to be in order. |
10:41 | briank joined #evergreen | |
10:44 | Dyrcona | Oh! That patch I threw on Lp is missing a local change to format the microsecond timestamps to %06d..... |
10:53 | Dyrcona | Heh. This branch is a mess.... |
10:54 | Dyrcona | So, I was testing with a dump of 1 library's records. It took about 1 hour 4 minutes to run. I'll make a change based on Stompro's bug description and see what happens. |
11:06 | Dyrcona | OK. Here goes.... |
11:11 | Stompro | Dyrcona, does this library have some of the bibs with large numbers of copies? |
11:20 | Dyrcona | I don't know. I doubt it. It doesn't seem to have made much difference so far. I'll try a larger library or the whole consortium next. |
11:20 | Dyrcona | It's one we do a weekly export for, so that's why I chose it to test. |
11:28 | Dyrcona | It does use slightly less (~3%) CPU |
11:30 | Stompro | In my testing, with our production data it had only a slight improvement. But it really improved the run that was stacked with bibs with lots of copies. |
11:36 | Stompro | acps_for_bre needs to be reworked to improve the --items performance in general. Maybe the first call just pulls in all call numbers and copies and caches them in a hash... |
11:37 | Stompro | Or go with the rust version that is already better :-) |
11:43 | Dyrcona | When I ran the queries through EXPLAIN ANALYZE, none of them were particularly slow. The slowest was the acp_for_bres query. On one particular record, it spent 40ms on a seq scan of copy.cp_cn_idx. I'm not sure how to improve a seq scan on an index, unless it can be coerced to a heap scan somehow. |
12:34 | collum joined #evergreen | |
12:38 | Dyrcona | Heh. Almost 1 minute longer..... |
12:50 | collum joined #evergreen | |
13:02 | Dyrcona | I am testing this now: time marc_export --all -e UTF-8 --items > all.mrc 2>all.err |
13:14 | Dyrcona | The Rust marc export does batching of the queries by adding limit and offset. I wonder if we should do the same? I've noticed that the CPU usage goes up over time, which implies that something is struggling through the records. The memory use stays mostly constant once all of the records are retrieved from the database. |
13:20 | Stompro | Dyrcona, if you use gnu time, it gets you max memory usage also. /usr/bin/time -v... so you don't have to check that separately. |
13:25 | Stompro | Dyrcona, I'm surprised the execution time increased for you... hmm. |
14:00 | csharp_ | berick: after installing the default concerto set, notes work - everything is speedy under redis - no errors yet |
14:07 | sleary joined #evergreen | |
14:16 | berick | csharp_: woohoo |
14:35 | Dyrcona | I've been testing Redis with production data, but not much lately. I need to write an email to ask the relevant staff here if they'd like me to update the dev/training server to use Redis on the backend. |
14:37 | Dyrcona | Current calculation puts it at only 20% faster, i.e. -1 day. |
14:38 | Dyrcona | I'm going to add a --batch-size option. If it is specified the main query will retrieve that number of records per request. I don't know if I'll get that implemented today. |
14:51 | Dyrcona | Looks like adding tags isn't my bottleneck My current estimate is minimal difference in performance. I'm going to let this run over the weekend to see if I'm wrong. On Monday, I'll add a batching/paging option to the main query and see if that helps. |
09:26 | redavis joined #evergreen | |
09:27 | Dyrcona | berick: I'm going to do some more research on Rust and look into adding a debug option to the Rust marc-esport.rs. I'd like to log the queries it runs, because it looks like the main query ran for 5 secs each time it was called on my produciton db the other day. I want to get the whole thing to run through explain analyze. It gets truncated in the postgres logs. |
09:37 | Dyrcona | Rogan left. I wanted to ask him about some queries he wrote, i.e. why he used subqueries instead of join in some places. Is it because the subqueries were faster or was that just how it came to him? I have seen subqueries be faster than joins in some cases, but not always. |
09:38 | Dyrcona | I suppose I could test it on my data and adjust if appropriate. It often depends on indexes and number of rows. |
09:49 | terranm67 joined #evergreen | |
09:51 | csharp_ | Dyrcona: I'm going to make a separate bug for my SFTP reimplementation so we can at least have functional SFTP (which currently doesn't work without my change or something like it) |
09:51 | Rogan joined #evergreen |
17:17 | berick | sleary_: working/user/berick/lpxxx-marc-edit-rec-type-change |
17:23 | mmorgan left #evergreen | |
17:26 | sleary joined #evergreen | |
17:30 | sleary | berick: I think https://bugs.launchpad.net/evergreen/+bug/2015163 is the bug filed on that issue (although it manifests in several ways). Seems to work with the tag table check commented out! We'll test it some more. |
17:30 | pinesol | Launchpad bug 2015163 in Evergreen "Fixed Fields Grid in Enhanced MARC Editor Not Updated on Save" [Undecided,Confirmed] |
17:51 | sandbergja joined #evergreen | |
18:26 | sandbergja joined #evergreen |
09:08 | Dyrcona | berick: I did `cargo build --release --package evergreen` then copied eg-marc-export to /openils/bin/. I missed the password on of the two lines for eg-marc-export in my script, so I don't know if it is faster, but the binary is certainly smaller without the debugging symbols, etc. |
09:14 | redavis joined #evergreen | |
09:18 | terranm joined #evergreen | |
09:18 | Dyrcona | FWIW, I haven't used --release on my test system. I did that for the production server. |
09:40 | csharp_ | one of the Supy/Limnoria bots in another IRC channel I was in had a @botsnack command and the bot would reply "YUM!" |
09:41 | csharp_ | actually in the Ubuntu channel, it would say "YUM!.. I mean, er, APT!" |
09:41 | bshum | Put that on the wishlist for bot development tasks... got it :) |
15:18 | kmlussier1 | I don't know what the harlequin/silhouette subscription thing was, but I was just talking fondly this week about my Columbia House mail order subscription. Those were good times! |
15:19 | kmlussier1 | Oh, my nick is messed up. How sad. Anyway, have a good weekend #evergreen! Safe travels to all who will be going to Indianapolis! |
15:19 | kmlussier1 left #evergreen | |
15:21 | terranm | The true Gen Xer test is the exposure to the Flowers in the Attic books |
15:21 | terranm | Those were passed around like contraband in my junior high |
15:22 | redavis | Oh, full exposure here. I was just thinking about those the other day and how wrong it all was. |
15:23 | redavis | hmm, I'm not sure how I got ahold of them. Kinda thing my mom might have accidentally bought them at a garage sale or something like that. The paperbacks swirled. Same with Stephen King. |
15:24 | terranm | So much Stephen King! |
10:28 | mantis1 | was missing one of those thank yo! |
10:32 | jeff | recall holds are a thing that requires some additional A/T setup. I don't know if the defaults are suitable / functional out of the box. Org unit settings likely also. |
10:33 | jeff | I was actually just wondering if it made sense to have an option to hide recall as a hold option, for those reasons and more. :-) |
10:39 | Dyrcona | berick: My test of the Rust marc export finish in 9 hours 23 minutes, and exported a 3.4GB binary file with 1,737,349 records. There are 411 records in the error output. That looks good to me, compared to what I'm getting from Perl. |
10:46 | csharp_ | rs++ |
10:57 | Dyrcona | csharp_: Do you think the marc_export is slow with the --items option? |
10:57 | briank joined #evergreen | |
10:59 | berick | Dyrcona: cool, good to know. |
11:00 | berick | re: errors, I saw a record or 2 in my tests that had subfield codes of "" (zero bytes) |
11:00 | berick | could add a flag or something to massage those into " " or some such |
11:01 | csharp_ | Dyrcona: yes it is |
11:01 | Dyrcona | I see some of those. I'll look through it later. I know we have bad records, because the Perl also complains. |
11:02 | Dyrcona | csharp_: Thanks. I felt like I'm going crazy because others have said it's not that slow. It was taking 5 days to the same export as I mentioned above. |
11:19 | kmlussier | Dyrcona: Wow, that's great that you were able to see such a big improvement! |
11:21 | Dyrcona | kmlussier: Yeah, I think Rust is more efficient than Perl, at least in my situation. I'm doing a test in produciton this evening. |
11:27 | Dyrcona | Y'know....It would be cool in psql if you could use \i for a subquery or in a CTE. That would save having two copies of the same code. (I know.... "Find another way to do it.") |
11:56 | Dyrcona | @decide Redgum or Rimu |
11:56 | pinesol | Dyrcona: go with Redgum |
15:28 | sharpsie | Burn It Down and Rebuild™ |
15:32 | book` joined #evergreen | |
15:40 | JBoyer | Sometimes a forest fire is responsible forest management! :p |
15:43 | jeff | okay, about to bug this unless it rings a bell for someone and they point out it's already got a bug: Angular MARC editor stashes field data in current_tag_table_marc21_biblio in a format that's just different enough to break context menus for the AngularJS MARC editor (which you may encounter in at least the Z39.50 Edit then Import interface). |
15:44 | jeff | for the Angular MARC editor, the way it creates the local data structure results in an array with element 100 being tag 100, etc. The older AngularJS MARC editor created an object with key 100, not array index 100, etc. |
15:45 | jeff | and of course, it's technically the tagtable service (old and new) that are doing this, not the "MARC editor"... |
15:47 | jeff | looks like it affects 3.10 and 3.11 at least, haven't tested main or looked too closely to see if the Z39.50 Edit then Import has moved to Angular yet. |
15:52 | jeff | steps to reproduce: clear (at least) current_tag_table_marc21_biblio from localStorage, then open the Angular MARC editor followed by Edit then Create from a Z39.50 search (which uses the older AngularJS MARC editor), and try to open the context menu on a tag value (other than the leader). |
15:52 | jeff | If you do the reverse, it works okay (the newer editor can tolerate the data in both formats, possibly by design, possibly by happenstance). |
15:53 | jeff | (or happy accident?) :shrug: :-) |
16:16 | sleary | Z39.50 in Angular and a revised MARC editor that relies less on context menus are both under construction now |
16:19 | jeff | the latter is bug 2006969, I think. |
16:19 | pinesol | Launchpad bug 2006969 in Evergreen "wishlist: enhanced MARC editor rewrite" [Wishlist,Confirmed] https://launchpad.net/bugs/2006969 - Assigned to Stephanie Leary (stephanieleary) |
17:01 | jeff | sleary++ jihpringle++ |
17:02 | jeff | (I'll still create a bug so we can fix 3.10/3.11) |
17:05 | mmorgan left #evergreen | |
17:07 | abneiman | jeff: I am not finding an LP for sprint A (which is on me, and I will rectify shortly) - but the spec you linked is the one we're working from. It's opening testing this week, and MARC Editor just opened testing so look for both on the spring release roadmap! |
17:08 | abneiman | also, apologies in advance for the flood of commits about to ensue ... next time I will learn to squash |
17:11 | pinesol | News from commits: Docs: LP2022100 updates to Item Status docs <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=e36ddd543aba55fd6ce4aeb5dd968b6f26eec4a4> |
17:11 | pinesol | News from commits: Docs: LP2022100 updates to Item Status docs <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=483d2f78eda53393c02a05eb54f5b91baf503c26> |
17:11 | pinesol | News from commits: Docs: LP2022100 updates to Item Status docs <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=c2ce3c25973f9a2128970cef0e1b2e0a85f1a022> |
10:41 | redavis | Thank you for your many hours of coffee roasting sacrifice. |
10:47 | smayo joined #evergreen | |
10:57 | briank joined #evergreen | |
11:00 | Dyrcona | berick: I don't know what the size of the file was at 143,000 records. The final file is 15GB with 2,251,864 bib records and 7,775,264 items. It took 18.3 hours to complete on my test system. |
11:00 | Dyrcona | berick++ |
11:04 | Dyrcona | Hmm. I guess that export running on production explains the "processor load too high on utility server emails." It keeps 1 CPU busy all the time, and there are only 8 cpus. Load was 11.5 when I just checked. |
11:04 | Dyrcona | Bmagic: I'm installing Rust on the CW MARS utility server today. |
14:53 | Dyrcona | Currently, it's actually returning acn.record, bre.marc, and the count on acp.id. |
14:56 | Dyrcona | Dude..... I just noticed the file ends with two semicolons.....I'll bet that's it. |
14:56 | Dyrcona | Still, I think I'll do the CTE. |
14:57 | berick | my test file contain: SELECT id, marc FROM biblio.record_entry WHERE NOT deleted (no semicolons needed) |
14:58 | berick | afk for a bit |
14:59 | Dyrcona | Yeah, ; is a habit from writing stuff for psql. |
15:07 | pinesol | News from commits: LP2035287: Update selectors in e2e tests <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=f562b3ac30a3753d63d565c2d7be4d3a7121a2fb> |
15:11 | Dyrcona | Now, I'm cooking with gas. I got the "large" bibs file (85 records with 87,296 items) dumped to XML in under a minute. That took almost 2 hours with the Perl program the other day. |
15:18 | Dyrcona | Using query-file to feed the eg-marc-export, it is using a lot more RAM than before, about the same as the Perl export was using. It still uses less CPU. We'll see if that changes over time. |
15:28 | jeff | you have 87,296 items that are spread across only 85 bibs? |
15:55 | brianmk joined #evergreen | |
15:56 | berick | Dyrcona: --pipe option pushed. i'm also happy to see any errors you encounter. |
15:57 | Dyrcona | berick++ |
15:57 | Stompro | Do the test email / sms features require that the app servers be able to send email? |
15:57 | Dyrcona | I might work up the nerve to submit some PRs later. |
15:58 | Dyrcona | Stompro: I don't remember, but there is something that does. |
15:59 | Stompro | I'm so used to only our utility server sending out email... |
16:01 | Dyrcona | Stompro: I think that's bug worthy if that's the case. |
16:02 | Stompro | I think I've been under the incorrect impression that action_trigger_runner had to be called to send email.... |
16:07 | Dyrcona | I could be wrong, but I swear I remember there being something that required the app servers to send email. Test emails and SMS go through action trigger. |
16:08 | berick | Stompro: open-ils.actor.event.test_notification ? looks like that fires the A/T event in real time from whatever server the API is called at. |
16:08 | berick | it doesn't wait for a_t_runner |
16:09 | berick | a variation on that API that creates but does not fire the event would be useful |
16:10 | Dyrcona | berick++ |
16:13 | Stompro | Hmm, my flowroute sms reactor is only setup on our utility server... I'll have to figure out how to get those test messages working. |
16:14 | Stompro | There probably is no reason it couldn't be on the app servers also. |
16:16 | Dyrcona | Stompro: I think a setting could be added to make the test notification API NOT fire the event, just leave it pending, then the pending a/t runner could pick it up, or it could be given the same granularity and the password reset emails. |
16:16 | Dyrcona | It would require development, of course. |
16:16 | Stompro | I was just stumped when I looked at the event def 155 entries and saw that they were processed 1 second after they were created. |
16:17 | Dyrcona | Yeah, in my case they're 523 and 524... |
Result pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139