Evergreen ILS Website

IRC log for #evergreen, 2017-03-29

| 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
03:00 Jillianne joined #evergreen
04:23 hawahaniz_ left #evergreen
05:01 pinesol_green News from qatests: Test Success <http://testing.evergreen-ils.org/~live>
06:55 agoben joined #evergreen
07:08 rjackson_isl joined #evergreen
07:08 rjackson_isl_ joined #evergreen
07:10 Dyrcona joined #evergreen
07:28 rjackson_isl joined #evergreen
07:34 graced good morning #evergreen... see you in a week!
08:06 collum joined #evergreen
08:12 kmlussier joined #evergreen
08:15 kmlussier Good morning #evergreen!
08:15 JBoyer 🎉
08:30 _bott_ joined #evergreen
08:31 gmcharlt a little game for folks, arising from prep for one of my presentations next week
08:31 gmcharlt guess the name, or at least nature, of the file that has been in the Evergreen source code the longest without any modification
08:32 kmlussier No modification?
08:32 gmcharlt more like file that hasn't been modified the longest time
08:33 gmcharlt modifications before it "froze" don't count
08:34 gmcharlt @coffee me
08:34 * pinesol_green brews and pours a cup of Nicaragua Maragogipe Organic, and sends it sliding down the bar to me
08:37 kmlussier I'm up against gmcharlt's program. I'm always scheduled against programs I want to go to.
08:38 kmlussier Is there a schedule on the web site that has the program descriptions?
08:38 gmcharlt hmm, just titles, as far as I know, although I assume that the link to the propsal spreadhseet may still be floating aroudn
08:43 mmorgan joined #evergreen
08:46 bos20k joined #evergreen
08:55 JBoyer terran++ # our first round of SMS courtesy notices is currently running.
08:55 JBoyer 11K texts, glad we're not charged on this end. D:
08:56 JBoyer gmcharlt, I don't suppose LICENSE.txt is near the oldest?
08:56 gmcharlt JBoyer: bzzzzzzzt!
08:57 JBoyer Ah, so we're talking real back catalog here, heh.
08:57 gmcharlt JBoyer: yeah, there was a textual update to the GPL2  that we applied back in 2010
08:59 JBoyer Ah, I'm out of the loop on licenses.
09:06 Dyrcona joined #evergreen
09:14 kmlussier gmcharlt: Is it an image file?
09:14 gmcharlt kmlussier: close, but there are two files that whose last modification time predates the oldest image file
09:15 * kmlussier considers this to be a game of 20 questions rather than a guessing game.
09:15 gmcharlt fair enough
09:16 kmlussier Is it bigger than a breadbox?
09:17 gmcharlt it's fairly large as files of its type go
09:19 JBoyer Animal, vegetable, mineral, or the model of a modern major general?
09:19 JBoyer ;)
09:19 * gmcharlt checks the Evergreen source code base for cats
09:20 gmcharlt *nothing meows back*
09:20 yboston joined #evergreen
09:28 jvwoolf joined #evergreen
09:35 bos20k joined #evergreen
09:40 collum joined #evergreen
09:47 collum kmlussier:  The proposals are here: https://docs.google.com/spreadsheets/d/1-n2-_yp​SvQbxHilv5SmkUaYcjl-a3yFhpmrIeqxVTJg/edit#gid=0
09:48 collum The schedule with the descriptions has not yet been uploaded to the web site.
09:48 kmlussier collum: Are you planning to upload the descriptions to the web site at some point?
09:49 kmlussier collum: Also, I'm looking forward to the conference next week. It looks like you all have put together a great program!
09:49 collum There's a pdf that was sent to the printer.  I'll try to get it on the website.
09:50 kmlussier Great, thanks! collum++
09:50 * kmlussier always loses her printed program part way through the conference, so it's always nice to have a backup on the web.
10:02 terran joined #evergreen
10:03 collum kmlussier:  Well that was easy, I thought I would run into a file size restriction.  It's the first link on this page under the conference logo.  https://evergreen-ils.org/conference/20​17-evergreen-international-conference/2​017-conference-programs-and-schedule/
10:04 kmlussier Yay! collum++
10:04 kmlussier I can confirm there is a file size restriction, because I've run into it before with the annual report, but I don't know what it is.
10:06 jeff Have I imagined a config / policy setting that doesn't exist? I was fairly certain that someone was doing "don't allow holds to be placed on items that are on the shelf".
10:07 kmlussier I think it's a YAOUS
10:09 * mmorgan agrees with kmlussier
10:09 kmlussier jeff: Has local copy block
10:09 jeff hah
10:09 jeff i found that one with a search on "block", but skimmed over it.
10:09 jeff thanks!
10:09 jeff kmlussier++ mmorgan++
10:22 mmorgan jeff: We have one library that uses that org setting. A problem they've run into is that there's no way to override the behavior.
10:24 dbwells gmcharlt: MARC21slim2MODS.xsl?
10:24 gmcharlt dbwells: ding-ding-ding we have a winner!
10:25 kmlussier gmcharlt: Is there a prize?
10:28 dbwells Well, several of us have been wanting to update that thing for at least the last 7 years.
10:28 * dbwells blushes
10:30 gmcharlt there is!
10:30 gmcharlt and here it is: https://librarypolice.com/winning_cat.jpg
10:31 JBoyer Ah, nuts. I was thinking that but was afraid it would come across as snark because of the aforementioned "let's update it." :)
10:31 JBoyer dbwells++
10:31 gmcharlt by the way, the oldest file that didn't ultimately originate from another project is Open-ILS/src/templates/marc/book.xml
10:33 * dbwells sends a job to the campus poster printer
10:37 bshum collum++ # pretty, thanks for sharing!
10:44 rhamby bshum++ : for helping out with wordpress
10:45 kmlussier bshum++
10:45 collum terran gets the credit for the program. terran++
10:45 * collum can make a link. :-)
10:50 * kmlussier is also in the "can make a link" camp.
10:50 terran collum++ thanks for posting! I kept thinking about it while I was away from my desk and then getting sidetracked by the time I got back
10:51 terran I can't take credit for the pretty cover - that was done by Samantha Tribble
10:53 kmlussier The cover is very pretty, isn't it?
10:53 kmlussier And I like the theme: Open and Unbridled
10:54 abneiman terran++ : program looks great!
10:57 kmlussier I'm also very happy that I'm presenting on Thursday this year. It looks like I'll be able to relax for most of the conference this year!
11:04 terran Thanks!
11:07 jvwoolf joined #evergreen
11:14 rlefaive joined #evergreen
11:21 Bmagic Anyone else struggle with the staff client freezing up when replacing cards? I know it has to do with autogen but even after I run autogen, we have this issue. We identified an issue a year ago with the hash result being different on each app server
11:21 Bmagic but concluded that the app servers would show a different hash because of the order of JSON strings getting parsed is not always the same.
11:22 Bmagic Right now I wonder if it matters which app server the client is assigned vs. which app server the client is assigned when replacing the card? Just a theory, how does that sit with you?
11:36 csharp Bmagic: yes, we've seen that - pretty sure it's related to https://bugs.launchpad.net/evergreen/+bug/1007380
11:36 pinesol_green Launchpad bug 1007380 in Evergreen "XACT_COLLISION Error in Patron Edit" [Medium,Confirmed]
11:37 Bmagic Not sure if this is our issue. We don't get an error
11:37 Bmagic The staff client just locks up, and you have to force kill the process
11:41 mmorgan Bmagic: We have not had reports of this issue from our libraries.
11:41 Bmagic mmorgan: Do you get differing autogen hashes?
11:43 terran Bmagic: We've seen that too, but I don't think we've gotten the complaint since our January upgrade. (Of course, they might have just stopped telling us about it.)
11:48 mmorgan Bmagic: is this in /openils/var/web/eg_cache_hash?
11:48 Bmagic mmorgan: Yep
11:49 mmorgan ok, it's different on our four bricks, but we no longer have app servers.
11:58 csharp Bmagic: we don't get an error either, and I don't remember how I tracked it to that bug, but I did at some point
12:05 csharp Bmagic: what our libraries reported was that they open the edit screen to see the cards, then go to the bills screen to create a billing/accept payment for the replacement card, then the hang happens when they try to replace and save
12:07 khuckins joined #evergreen
12:09 Bmagic csharp: roger that
12:19 Christineb joined #evergreen
12:24 jihpringle joined #evergreen
12:25 kmlussier @coffee
12:25 * pinesol_green brews and pours a cup of Decaf Organic Caffe Volcan, and sends it sliding down the bar to kmlussier
12:25 kmlussier Decaf!!!
12:25 kmlussier @blame pinesol_green
12:25 pinesol_green kmlussier: It really IS itself's fault!
12:25 JBoyer Bmagic, We get that when new libs migrate in and they're replacing cards non-stop for a few days. Evergreen.exe pegs a processor core and deadlocks on something until you kill it. I always assumed it was related to the memory leak issues of old.
12:26 JBoyer It's one reason we're jumping on 2.12, we want the web client at circ asap.
12:26 Bmagic JBoyer: I bet it's autogen for you especially since you mention new migrations
12:27 JBoyer Our hashes never match since all 10 are generated independently. It's more the frequency of user updates and possibly the process used.
12:28 Bmagic Yeah, there has to be something lurking in there somewhere
12:28 Bmagic I just reran autogen, and we shall see
12:28 JBoyer Better debugging on windows would be helpful, but since you can't really trigger it on demand (and I don't know if debugging symbols are still available for that vintage xulrunner...) we've just ridden it out.
12:30 kmlussier JBoyer: When are you planning to upgrade to 2.12?
12:30 JBoyer Easter weekend.
12:31 JBoyer Hoping to squeeze in a vacuum full analyze too, takes 8 hours. :/
12:31 kmlussier Wow! Just a few weeks away. Awesome!
12:31 * JBoyer remembers that he needs to actually tell people in Indiana this...
12:31 kmlussier Be sure to file lots of bugs so that we can make the web client even better for 3.0
12:31 JBoyer Thats the hope!
12:32 kmlussier Or, even, 2.12.2. No reason to wait for 3.0. :)
12:32 JBoyer Well, the hope is that there aren't that many to *be* filed, but you know.
12:32 kmlussier JBoyer: I always feel very confident that my thorough testing revealed everything there is to find. And then people start using it in production, and I find out I'm wrong.
12:33 JBoyer "But WHY would anyone do THAT?! That's not how any of this works!"
12:33 berick JBoyer: vacuum full analyze the whole db?  nice
12:35 * berick has a few vacuum-full's lined up, expects to recover several hundred GB of disk space
12:35 berick after cleaning up copy auditor and anonymizing holds
12:35 JBoyer berick, Yeah, that's where all the time will go. I think the db upgrade scripts (minus reingest, of course) only takes a few minutes. I was about to do another test run on our dev server to double check.
12:36 JBoyer we keep all of our auditors trimmed (we delete 2 hogsheads a fortnight), and we run circ/hold purges daily. I'm hoping it all goes smoothly.
12:36 berick heh
12:37 JBoyer I'm reminded though that slony will be the hassle here. Perhaps for the next large upgrade we move to Pg 9.5 and back to the built in sync.
14:05 kmlussier joined #evergreen
14:12 jvwoolf joined #evergreen
14:21 kaffenkj joined #evergreen
15:03 Dyrcona So, if I configure the hold targeter to run in parallel, should I see more than 1 running with ps?
15:03 Dyrcona I basically see the bash -c that started it and 1 hold targeter.
15:05 berick Dyrcona: there will be multiple storage (old-style) or open-ils.targeter (new-style) processes actively working.
15:05 berick still just one hold_targeter*pl process
15:06 Dyrcona berick: Thanks! That makes it tougher to figure out, so I'll assume it is working. :)
15:06 Dyrcona I'm using old style. I don't recall new style making it into master, yet.
15:07 berick it's in 2.12
15:07 Dyrcona OK.
15:08 Dyrcona So, another question: has anyone noticed the same copy being targeted at different holds with parallel targeting?
15:08 Dyrcona I had an example sent to me and the prev_check_time for the two holds is more or less identical.
15:10 Dyrcona Different users, different pickup locations, etc. Same bib, same current copy.
15:11 berick parallel targeting segregates holds by metarecord to prevent that exact scenario.
15:11 berick i can't say I've ever seen it happen
15:11 Dyrcona The holds were placed about six minutes apart and the previous checck times are 30 seconds apart.
15:11 Dyrcona Mabye this is more memcached "weirdness."
15:12 berick are the holds old enough that they have not yet been re-targeted by the batch targeter?
15:12 berick or young enough, rather
15:12 Dyrcona I don't thinks the request time is on 3/13/17 and the previous check time was yesterday afternoon in this example.
15:13 Dyrcona They would be retargeting about now, actually.
15:13 Dyrcona ugh.. "I don't think so..."
15:14 Dyrcona Think I'll check them in the database.
15:16 Dyrcona Prev check time is still yesterday at 15:02. Probably need to wait another hour or so.
15:16 berick still have same current_copy ?
15:16 Dyrcona I'll see how many copies there are on this bib.
15:16 Dyrcona Yes, they do.
15:17 Dyrcona Wait a minute, the database disagrees with the spreadsheet on prev_check_time or I read the spreadsheet wrong.
15:17 Dyrcona The prev_check_time is only 9ms apart.
15:19 jvwoolf joined #evergreen
15:19 Dyrcona Neither hold is frozen.
15:20 Dyrcona Wonder if only 1 copy is currently eligible to fill the holds?
15:21 Dyrcona That's more difficult to figure out.
15:21 Dyrcona I read the spreadsheet wrong, of course.
15:21 Dyrcona It doesn't have nanosecond resolution, though.
15:22 berick i would still only expect the copy to be targeted once.
15:33 Dyrcona Yes, that what we expect, but that is not what is happening.
15:33 Dyrcona If only I could type..... ;)
15:33 Dyrcona I think it is related to a hardware crash on 2/20/17.
15:34 Dyrcona We were doing some o/s package updates and the util2 server that runs the hold targeter did not reboot because of a bad drive.
15:34 Dyrcona I built a vm to serve as a replacement.
15:35 Dyrcona Then, we got the a replacement hard drive and put the original hardware back in service, but with a freshly installed o/s because the system would not automatically recover with the new drive in place like it was supposed to.
15:35 Dyrcona I believe that had to do with a failure to load the RAID driver.
15:36 Dyrcona So, anyway, it is now basically a brand new machine and no guarantees that the configuration was the same.
15:36 Dyrcona This is also the machine I suspect of being better off running its own memcached.
15:37 Dyrcona So, we'll restart services, etc., tomorrow morning to get it to use localhost for memcached.
15:37 Dyrcona And the reports of this started after the replacement VM went into service.
15:38 * Dyrcona mumbles something about moving parts.
15:41 Dyrcona In the meantime, the holds retargeted about 13ms apart and neither has a current copy.
15:47 * berick launches VM on laptop, sees a snapshot of where things ended at the Fall hackfest
15:48 * Dyrcona doesn't have enough available RAM to start a VM, but wouldn't have on that old still hanging around anyway. :)
15:58 Dyrcona Did the open-ils.targeter service make it into the opensrf.xml.example file?
15:58 Dyrcona I just compared my custom config with master, and I don't see it in the diff.
16:00 berick yep
16:00 Dyrcona I don't see it in a checkout of master or rel_2_12. Is it called open-ils.targeter?
16:01 Dyrcona Oh. its' hold-targeter
16:01 Dyrcona it's... I can apostrophe, I can.
16:03 Dyrcona Heh. I've been installing config with open-ils.hold-targeter in it on my 2.10 servers.... ;)
16:04 Dyrcona Or, actually, no. The cherry-picking has done what it should.
16:05 Dyrcona There's also a new ebook_api service!
16:05 Dyrcona Cool stuff.
16:06 Dyrcona I've been too bogged down in the day to day to keep up with everything.
16:28 mmorgan joined #evergreen
16:33 khuckins_ joined #evergreen
17:01 pinesol_green News from qatests: Test Success <http://testing.evergreen-ils.org/~live>
17:10 mmorgan left #evergreen
17:38 khuckins joined #evergreen
18:56 kenstir joined #evergreen
19:00 kenstir Is there anybody that hangs out here in the evenings, or is this a 9-to-5 crew?
19:03 jihpringle kenstir: the channel tends to be active 9-5 Eastern Time
19:03 jihpringle there are usually a few of us from the west coast around until at least 5pm Pacific
19:05 kenstir Sadly I have a day job, so I have most of my OSRF questions in the evening.  Maybe with a TARDIS IRC client I could send my questions back to when folks are around.
19:06 jihpringle kenstir: you may want to try the mailing lists - https://evergreen-ils.org/​communicate/mailing-lists/
19:06 kenstir That's a very good idea
19:07 jihpringle depending on your questions the general or the dev (technical) lists are probably the best ones
19:08 kenstir Definitely the dev list.  I totally forgot about the list!  Thanks for snapping me out of it.
19:08 jihpringle no problem, glad I could help :)
19:14 frank_guel joined #evergreen
19:17 frank_guel Hi all, Is there any kind of bug or how can I fix it? When I create a new marc record into the 2.12v using a marc template used in 2.8.4, it created it but looks like it is encoded with other format, I mean If for example I have an © in the template, when I load it in staff client it looks well, but when I save it ir shows weird symbols, and if I open and replace the record with a © manually, save it, and it looks well
19:22 jihpringle frank_guel: that sounds like a bug to me, I checked our templates which came from 2.10 to 2.12 and we don't have any symbols in our templates so I can't replicate
19:23 frank_guel could be an accent or a symbol.
19:25 jihpringle our templates are pretty much just lists of blank fields per item type
19:27 frank_guel ah ok ok, I did another test, I open the k_book.xml default template, add a word with an accent, and it save it as a weird character, if I open it again and replace the character with the correct word, save the changes and it looks well
19:28 jihpringle frank_guel: I'll ask one of our techs to add some accents/symbols to one of our templates and see if we can replicate
19:30 frank_guel that sounds great, thanks
19:30 jihpringle I'll let you know whether or not we can replicate
19:37 frank_guel ok, that's good thanks
22:11 khuckins joined #evergreen
22:20 genpaku joined #evergreen

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