Evergreen ILS Website

IRC log for #evergreen, 2021-02-09

| 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
06:00 pinesol News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>
06:11 kip joined #evergreen
07:00 Dyrcona joined #evergreen
07:10 kip joined #evergreen
07:16 rjackson_isl_hom joined #evergreen
08:01 Christineb joined #evergreen
08:23 mantis1 joined #evergreen
08:27 rfrasur joined #evergreen
08:41 mmorgan joined #evergreen
08:45 Dyrcona I'm going to work on bug 1904737, and I wonder if that is a bug or a new feature. I set it to Importance: Wishlist as a new feature, but something csharp said recently makes me think it is part of an incomplete feature, and therefore a bug. What do others think?
08:45 pinesol Launchpad bug 1904737 in Evergreen "Flag/setting for Items with Specific Statuses to Display on Holds Pull List" [Wishlist,Confirmed] https://launchpad.net/bugs/1904737 - Assigned to Jason Stephenson (jstephenson)
08:50 Cocopuff2018 joined #evergreen
08:51 terranm joined #evergreen
09:11 jvwoolf1 joined #evergreen
09:33 collum joined #evergreen
09:44 collum joined #evergreen
09:45 Dyrcona Ugh: So, what's the point of is_avaialable if ccs already has holdable?
09:57 terranm can't something be holdable but not yet available, like on-order items?
10:00 csharp yes - that's the purpose of is_available (as I remember)
10:01 csharp I don't remember what I said about the statuses thing, but nice to know others are paying attention :-D
10:03 Dyrcona OMFG! Why do we do this to ourselves?
10:04 Dyrcona Yeah, we just had that conversation in internal emails.
10:04 Dyrcona So, the pull list should look at holdable and is_available.
10:05 Dyrcona The targeter should look at holdable, but prevent is_available: false from becoming the current_copy.
10:05 Dyrcona On the plus side, I think the latter will reduce or eliminate the need for "retarget all statuses" on checkin of new copies.
10:06 * Dyrcona also went back and read the release notes, which do not explain the purpose all that well, though they're pretty good for the mechanics of it.
10:08 Dyrcona Also, and I know I have mentioned this before, but there's too much "hold targeter" code and some of it should be deleted because I know it is dead, but I can never remember which of the 3 implementations that is.
10:14 terranm +1
10:15 Dyrcona It was also mentioned in our internal emails that you can have a status that is available and not holdable for "on the shelf" or walk in items. You might call it a ready reserve.
10:17 mmorgan Dyrcona: Not all of that needs to be controlled by status, though. We use holdable flags in items and locations for the "ready reserve" situation.
10:18 mmorgan Re: pull list, is that currently looking at the is_available flag on the status?
10:19 Dyrcona mmorgan: No, it is hardcoded for statuses 0 and 7.
10:20 mmorgan Also, not sure what you suggest will eliminate teh "retarget all statuses" at checkin.
10:20 mmorgan Dyrcona: Ok, thought that was still true.
10:22 Dyrcona mmorgan: I think it will reduce the need for retarget all statuses if things like on-order copies appear in a hold's copy list and don't have to be forced onto it.
10:22 Dyrcona The latter is the point of retarget all statuses....
10:23 Dyrcona Also, after looking into this, plus this and other discussion, I think I've answered my first question. I'm going to leave the bug as a wishlist. Even if it is a bug fix, it's big enough not to warrant a backport.
10:25 mmorgan Re: on order, definitely true, we have On order as a holdable status. We still have to use "retarget all statuses" as not all newly added items get a chance to have their holds targeted completely
10:30 malexander joined #evergreen
10:30 Dyrcona Yes, and I am probably wrong about what I said. The holds code is ridiculously convoluted, and it is nearly impossible to reason about it, reasonably. I think I'm going to give up on targeting changes and just work on the pull list code. I'm relatively confident what needs changing there.
10:30 jeff What are you working on / trying to do?
10:30 jeff oh, scrolled back half a screen more and i see, bug 1904737?
10:31 pinesol Launchpad bug 1904737 in Evergreen "Flag/setting for Items with Specific Statuses to Display on Holds Pull List" [Wishlist,Confirmed] https://launchpad.net/bugs/1904737 - Assigned to Jason Stephenson (jstephenson)
10:33 Dyrcona Conclusion: Instead of the currently hardcoded statuses of 0 and 7 for the pull list, I'm going to change it to status->{holdable:true, is_available:true}.
10:33 Dyrcona And, that's it.
10:34 Dyrcona If it turns out that hold targeting needs other changes later (not likely), then that can be a different bug.
10:38 Dyrcona terranm++ chsarp++ mmorgan++
10:38 Dyrcona csharp++
10:38 Dyrcona @karma chsarp
10:38 pinesol Dyrcona: Karma for "chsarp" has been increased 1 time and decreased 0 times for a total karma of 1.
10:38 Bmagic :)
10:40 Bmagic Back on my invoice-matched-ancient-PO issue - I'm starting to think that Evergreen matched that PO from 2013 because it didn't find a match at all. And it needed to match something*. Vendcode needed to be blank for the EDI account but wasn't And vendcode/vendacct were backwards. This is an EDI account that has existed for years
10:40 Dyrcona There is line 798 in OpeniILS::Utils::HoldTargeter.pm.
10:41 jeff i'd omit the requirement that the status be holdable=true, though I can't come up with a specific use case for you right now. it just seems like a new thing that is in the wrong place.
10:41 Dyrcona Ah, well if vendacct/vendcode were backwards, I can see that leading to problems.
10:42 Bmagic It's not had any trouble until now. The difference is that now it's using edi attrs. I bet, that the PO > vendor is now containing strange identifiers (because edi attrs respects the vendcode/vendacct differently). Then later, vendor invoice > Evergreen. Garbage in garbage out.
10:42 Dyrcona jeff: That was the original thing I was going to do, just is_available, and then I got lost in the code.
10:42 * Dyrcona is tempted to use a print out of Holds.pm to start a fire as a symbolic gesture. :)
10:43 Bmagic Dyrcona++
10:43 jeff bah, I looked at the bug again.
10:43 mmorgan jeff: Dyrcona: I'm leaning toward agreeing with jeff about omitting the holdable=true requirement. This could further reduce the need for the "retarget all statuses" checkin modifier.
10:44 Bmagic speaking of printing the Evergreen source, ehem, miker! Sirsi....
10:44 Dyrcona Well, its only code, and it isn't committed, yet, so......
10:45 jeff I'm not convinced that using statuses of Display or Storage are good ideas, for some of the reasons mentioned already in that bug and more. :-)
10:46 miker Bmagic: a few reams of dead tree with v0.9 on it is probably in a safe somewhere in AL or UT
10:47 Dyrcona TIMTOWTDI is not always a good thing. In fact, it usually isn't.
10:47 Bmagic miker: Oh good. I was worried. LOL
10:48 Bmagic miker: I often recall your laughing fit and start giggling all over again
10:48 Bmagic kmlussier suggested at the time that we start recording. I wish someone had!
10:49 * csharp keeps his printed source code in a file cabinet and updates it daily
10:49 csharp so much whiteout
10:49 Bmagic csharp: I thought pinesol was programmed to immediately re-print any source file that gets changed in master along with the post in IRC
10:49 miker csharp: all my diffs from 2020 are just marginalia
10:50 csharp Bmagic++ miker++
10:50 sandbergja joined #evergreen
10:51 Bmagic The dot matrix printer has a never-ending supply of track paper. We use dot matrix to keep our ink costs low and easy to splice new paper onto the supply on the floor
10:51 jeff Bmagic: no, you're thinking about our DR copy of the catalog.
10:51 jeff (on cards, of course)
10:52 Bmagic yeah, that's why it takes us a little longer to re-compile the source on feedback fest machines. We have to re-feed the new source (with patches) into the feeder arm
10:52 miker Dyrcona: 1) hard agree on changing every single instance of "status in (0,7[,and sometimes 12])" to "is_available" (in living code) 2) you're probably right about holdable being the only requirement for getting on the potentials list re retarget-all-statuses. of course, the checkin flag almost certainly predates the holdable status flag
10:53 Bmagic once Evergreen hit 2 million lines, we had to invest in a taller printer stand to make room for the paper
10:54 Bmagic We have a "compile room' where Tom and Jerry are working tirelessly. They hate it when a new patch comes out. They have to start all over again. The machine looks like the machine that Touring used to crack enigma.
10:55 Bmagic Sometimes I write patches just to watch their faces. Priceless.
10:57 terranm joined #evergreen
11:01 jeff "System throughput improved significantly in January of last year, when staff eliminated one of two places where MARC data was being transmitted via semaphore. A plan to eliminate the current round trip to morse code and back is contingent upon identification of a vendor capable of dealing with more modern inputs, such as magnetic tape."
11:08 Dyrcona And, with an internal email, and miker's comment, I'm back to "complicated." Think I'll shut up for a bit and look through the code again. Too bad I can't debug the server code any more. kivilahtio's steps haven't worked since OpenSRF 3.0 or so.
11:11 miker Dyrcona: to be clear, I'm agreeing with you... :)
11:12 pinesol [evergreen|Tiffany Little] Updates the lineitem catalog link to point to the new Angular catalog. - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=2f8a477>
11:13 Dyrcona miker: Yes, and I think that means more work, because I've been told contradictory things, by others, that don't necessarily coincide with what I see in the code.
11:16 Dyrcona Heh.  "White Rabbit" by Jefferson Airplane just started playing.....
11:16 Dyrcona It's the mono single version remixed to stereo.
11:16 pinesol [evergreen|Bill Erickson] LP1914630 Create MARC record goes to AngstCat - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=169c4b1>
11:28 gmcharlt so, since there's a big spate of wishlist bugs that don't have release notes, but which individually are easy to describe in a simple sentence, I think I'm going to start a docs/RELEASE_NOTES_NEXT/minor_improvements.adoc for committers to add those one-liners
11:28 gmcharlt any immediate objections?
11:31 terranm gmcharlt++
11:34 gmcharlt well, miscellaneous.adoc
11:34 terranm berick: On this one, should I create a separate wish list for a select/deselect all box? https://bugs.launchpad.net/evergreen/+bug/1899406
11:34 berick +1
11:34 pinesol Launchpad bug 1899406 in Evergreen 3.5 "Angular catalog: missing "View Basket" functionality" [Undecided,Confirmed]
11:35 * berick looks
11:36 terranm sandbergja: Happy birthday!
11:36 malexander joined #evergreen
11:36 pinesol [evergreen|Jane Sandberg] LP 1816655: Add a patron search button to create booking screen - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=ccb9cbe>
11:36 pinesol [evergreen|Galen Charlton] LP#1816655: add brief release notes entry - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=9618d2c>
11:38 berick terranm: i don't think a new bug is needed.  i'll try to get this one this week.
11:38 terranm berick++
11:40 pinesol [evergreen|Jane Sandberg] LP 1907921: Add a patron search button to course admin screen - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=a478844>
11:40 pinesol [evergreen|Galen Charlton] LP#1907921: add release notes entry - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=4346c20>
11:42 pinesol [evergreen|Terran McCanna] LP1900159 Staff Catalog Typo - No Results - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=d242d6a>
11:44 * gmcharlt claims 1244
11:49 pinesol [evergreen|Galen Charlton] LP#1786100: fix upgrade issue that can prevent unmapping item tags - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=bad899b>
11:49 pinesol [evergreen|Galen Charlton] LP#1786100: stamp schema update - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=834c617>
11:51 pinesol [evergreen|Jane Sandberg] LP1907923: Move course materials button to more suitable location - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=f797b13>
11:53 pinesol [evergreen|Garry Collum] LP1906858 Carousels Ignore Deleted Flag - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=bb2e6dd>
11:57 pinesol [evergreen|Terran McCanna] LP1908724 Staff Catalog Sort by PubDate Wording - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=893a5f9>
11:59 pinesol [evergreen|Bill Erickson] LP1907286 Staff catalog sets last retrieved record - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=449df59>
12:12 jihpringle joined #evergreen
12:28 Dyrcona Speaking of release notes, the branch on  bug 1861319 has release notes AND tests!
12:28 pinesol Launchpad bug 1861319 in Evergreen "Allow Item Renewal on Expired Patron Record" [Wishlist,Confirmed] https://launchpad.net/bugs/1861319
12:28 sandbergja joined #evergreen
12:56 mantis1 joined #evergreen
12:58 alynn26 joined #evergreen
12:59 khuckins joined #evergreen
13:01 pinesol [evergreen|Mike Risher] lp1847519 Port of Circulation Limit Set UI - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=db7271e>
13:01 mantis1 left #evergreen
13:05 mantis1 joined #evergreen
13:09 Bmagic JBoyer: our test server does* have the RemoteIP settings: RemoteIPInternalProxy 127.0.0.1/24 RemoteIPInternalProxy ::1  RemoteIPHeader X-Real-IP
13:10 Bmagic I'm not sure based on the comments above whether those are or are not suppose to be uncommented? Ours are uncommented
13:10 JBoyer Is nginx (or whatever) running on the same VM?
13:10 Bmagic yes
13:13 JBoyer That is odd. Is nginx listening to the public IP address or is it behind something else?
13:14 Bmagic nginx is attached to an internal IP which is DNAT/SNAT to/from the external firewall
13:16 Bmagic well, technically nginx is attached to a docker IP, which docker networking makes seemless from the outside host, which is DNAT/SNAT
13:18 JBoyer Huh. That should probably be enough to get nginx the correct client ip to pass along to apache. If you're not seeing any of the same "Uncaught Transaction begin error" errors that Chris was seeing yesterday I may be grasping at straws.
13:18 Dyrcona Bmagic: You may need to add that internal IP address as a RemoteInternalProxy. Also, are you using X-Real-IP in the nginx configuration?
13:18 Bmagic I'm ruling out the external firewall by hitting the box from inside the network. here is the log from nginx
13:19 Bmagic "POST /osrf-http-translator HTTP/2.0" 400 404 "https://bugsquash.mobiusconsortium.org/eg/c​onify/global/config/circ_matrix_matchpoint" "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:85.0) Gecko/20100101 Firefox/85.0"
13:20 Bmagic Dyrcona: you might be onto something. nginx has this for the translator location: proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for; proxy_http_version 1.1;       proxy_set_header Upgrade $http_upgrade;        proxy_set_header Connection "upgrade";
13:21 Dyrcona So, you want X-Forwarded-For in your Apache configuration, not X-Real-IP, or the other way 'round. :)
13:21 Bmagic right, that's what I was thinking
13:21 Dyrcona They have to agree on the header that they're using.
13:24 Bmagic didn't help :(
13:25 Bmagic I'll restart all services just to be sure
13:25 pinesol [evergreen|Chris Sharp] LP#1901900: Update EDI Webrick installer for Ubuntu 18.04 - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=4c5a7d5>
13:25 Dyrcona You should only have to restart Apache or nginx, depending on which one you changed.
13:26 Bmagic same problem still
13:27 Dyrcona I'd also add the internal IP address with a /24 in its own RemoteIPInternalProxy entry.
13:27 pinesol [evergreen|Garry Collum] LP1908763 Survey column sorting broken - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=8d7edf3>
13:29 csharp yes, Firefox, I *would* like to close the 63 open tabs to the right
13:29 Bmagic Dyrcona: I've done all that, no change.
13:30 Bmagic I guess I can delete the container and go with a straight VM, but it's odd that it works in a container everywhere else
13:31 Dyrcona Well, I've had fun with a VM where things didn't work right, than I blew it out and built it over, and everything was fine. Dunno if I messed something up or things were just wonky with the first image.
13:32 Dyrcona If it's quicker, I'd build a new docker image, first.
13:34 Bmagic yeah, ok, I'll start from scratch. And if this is any clue, the browser includes this message in the error: "no-referrer-when-downgrade"
13:34 Bmagic the server is on 16.04 if that matters. I'll go up to 18.04 this time
13:34 Dyrcona Well, it could be something new with Firefox 85.
13:35 gmcharlt csharp++ # TIL that Firefox has "close tabs to the right"
13:36 csharp heh
13:37 csharp built a new PC over the holidays - my old one would've been sweating with that many opn
13:37 csharp open, also
13:37 Dyrcona Bmagic: I don't think no-referrer-when-downgrade has much to do with it. Also, I'm not 100% what "it" is. I realize that I seem to  have only part of a conversation.
13:49 pinesol [opensrf|Bill Erickson] LP1827055 Remove Python libs, install bits, and docs - <http://git.evergreen-ils.org/?p​=OpenSRF.git;a=commit;h=264958b>
13:51 pinesol [evergreen|Bill Erickson] LP1827055 Remove Python libs, install bits, and docs - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=4613155>
13:53 pinesol [evergreen|Bill Erickson] LP1881650 Staff catalog TCN includes delete records - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=730f0f4>
13:53 pinesol [evergreen|Galen Charlton] LP#1881650: add release notes entry - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=d3f5f94>
14:22 terranm joined #evergreen
14:57 sandbergja joined #evergreen
15:19 malexander_ joined #evergreen
15:31 mantis1 left #evergreen
15:41 malexander__ joined #evergreen
15:44 nfBurton joined #evergreen
15:45 nfBurton joined #evergreen
15:55 JBoyer_ joined #evergreen
15:56 Dyrcona Anyone ever had drone servers on a brick run out of RAM?
15:57 Dyrcona And, what's crazier is the swap is basically full with 13GB of free RAM!
15:58 JBoyer if something forced ram into swap and then died from an OOM or something the stuff in swap still sits there until re-paged in. If that process is largely idle it may take ages.
15:59 JBoyer And I've seen some pretty ugly things re: ram use on occasion, yeah.
16:02 Cocopuff2018 joined #evergreen
16:02 Dyrcona OK, I'll check for OOM killer. Also, both drones servers on this brick have a lot of prcrud drones running according to osrf_control --diagnostic.
16:03 malexander_ joined #evergreen
16:05 Dyrcona Yeah, oom-killer ran on both drones; oom-reaper on only 1 of the two, so far.
16:08 malexander joined #evergreen
16:23 jeffdavis I think it's been awhile since we've seen something like that, but "oom" is an annoying string to grep for in chat logs
16:26 malexander joined #evergreen
16:29 jihpringle joined #evergreen
16:35 jeffdavis Last drone-related OOM that I can confirm here was Nov 2019, then a few Apache-related incidents the following March. Increasing RAM made the biggest difference, although EG upgrades/security fixes may also have helped a lot. I've also been babysitting the servers more lately so may have caught some stuff before oom-killer did.
16:39 Dyrcona I'll take a look tomorrow to see if I can figure it out. Hasn't happened here in a long time.
16:39 mrisher jihpringle: I put in a fix on that bug you reported yesterday.  Thanks for spotting it!  1857911
16:40 mrisher (bug #1857911)
16:40 pinesol Launchpad bug 1857911 in Evergreen "Angular: Local Admin Statistical Categories Port" [Wishlist,New] https://launchpad.net/bugs/1857911
16:42 jihpringle mrisher: awesome, I'll keep an eye out and give it another test if it gets updated on one of the test servers this week
16:52 terranm Err... so if I have a carousel that is manually created from a bucket, then I change the contents of that bucket, how does the carousel update? (Mine is still showing the initial bucket records and it won't let me manually refresh it)
16:55 terranm Oh never mind, I was looking at the wrong bucket. X)
17:15 mmorgan left #evergreen
17:30 Bmagic alright - JBoyer Dyrcona, I got the server to save circ policies once more!
17:30 Bmagic replying to the email
17:35 jihpringle thanks Bmagic++
17:36 Bmagic no prob!
17:36 Bmagic classic unplug it and plug it back in
18:00 pinesol News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>
18:04 jihpringle joined #evergreen
18:51 JBoyer Any day you have to enable trace logging in apache is probably not a good day. Incidentally, I've learned you generally don't have to go any higher than trace4 if you're diagnosing named virtualhost problems.
18:52 JBoyer ALSO, if you include :443 or :80 in a servername / serveralias you're committing a crime.
18:52 * JBoyer stumbles away
21:07 Cocopuff2018 joined #evergreen
23:42 akilsdonk joined #evergreen

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