Evergreen ILS Website

IRC log for #evergreen, 2019-07-12

| 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
05:00 pinesol News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>
07:06 rjackson_isl joined #evergreen
07:26 collum joined #evergreen
08:17 bos20k joined #evergreen
08:38 mmorgan joined #evergreen
08:50 Dyrcona joined #evergreen
08:58 bos20k_ joined #evergreen
09:08 sandbergja joined #evergreen
09:09 aabbee joined #evergreen
09:32 jvwoolf joined #evergreen
10:00 sandbergja joined #evergreen
10:44 dbs Circulation policies, hmm. I want to be able to block users from borrowing once they've accumulated a certain threshold of fines within our consortium
10:45 dbs Is that where standing penalties come into play?
10:47 Dyrcona Indeed it is.
10:48 dbs Having a tough time finding a writeup about how system-generated penalties get applied in the docs. Lots about staff-generated penalties though
10:49 mmorgan Group Penalty Thresholds is where to set thresholds
10:50 mmorgan ... under Local Administration
10:52 dbs mmorgan++ # thanks
10:52 dbs now to figure out why the "Penalty" dropdown in the "New penalty threshold" dialogue is empty :(
10:56 dbs Ah well, back to plain old SQL. Much better.
10:58 mmorgan Well, if you want to take the easy way out... ;-)
11:01 * mmorgan has not tried getting to Group Penalty Thresholds in the web client before. Appears broken.
11:09 mmorgan Ah. Ok. bug 1812389
11:09 pinesol Launchpad bug 1812389 in Evergreen "Group Penalty Threshold link does not work" [Medium,Fix released] https://launchpad.net/bugs/1812389
11:11 dbs weirdly we're on 3.1.2 and it is broken for us, but... we're weird
11:11 dbs thank you!
11:12 dbs wait, different bug, we get to the right page but the "Penalty" dropdown when you go to add a penalty is empty
11:12 * dbs shrughs
11:12 dbs (that's a shrug and a sigh I guess)
11:13 Dyrcona :)
11:13 * Dyrcona goes back to poking logs for information about segfaults.
11:16 Dyrcona Hmm... Someone used operator change just before midnight last night?
11:16 Dyrcona I guess some of the community colleges were open.
11:28 dbs We don't have 3.3 docs yet eh?
11:29 dbs (well, we have them, just not generated on docs.evergreen-ils.org...)
11:29 rhamby looks like it
11:30 rhamby docs are overrated anyway .... who wants to actually know the outcome of their actions ....
11:34 Dyrcona JBoyer: I have found a segfault that looks like it happened after a regular staff login timed out. Same pgt search that triggered the segfault.
11:34 Dyrcona The authtoken appears in the logs for several hours before it was deleted and then the crash.
11:36 Dyrcona Yeahp: Session retrieve at 21:44, then nothing until 00:44 when the session was deleted, followed by crash.
11:42 dbs So just restart the server every 10 minutes or so to avoid that segfault?
11:43 dbs (Seriously though - that's a great find between you and JBoyer)
11:51 dbs Oh yeah and none of the existing Group Penalty Thresholds are being displayed in that interface, which is also weird. Weird on weird
11:51 dbs I guess I'll revisit it once we upgrade from 3.1.2 to 3.3.x
11:59 BAMkubasa joined #evergreen
12:03 jihpringle joined #evergreen
12:10 jihpringle gmcharlt: for the second of the new patches on LP 1777207 is the "previous sort order" in bullet 3 the original check in order or the order after the patron sorted?
12:10 pinesol Launchpad bug 1777207 in Evergreen "Web Client - Check In List Refreshes After Every Scan and Gets Progressively Slower" [Medium,Confirmed] https://launchpad.net/bugs/1777207 - Assigned to Galen Charlton (gmc)
12:11 jihpringle just want to make sure I know what I should be seeing when I test
12:15 Christineb joined #evergreen
12:22 miker jihpringle: donning gmcharlt mask, the user-applied sort
12:23 miker IIUC
12:23 Dyrcona joined #evergreen
12:24 jihpringle miker: thanks :)
12:26 sandbergja joined #evergreen
12:48 Dyrcona And, now, I think I've seen multiple of the pgt searches with null authtoken from the same session delete.
12:49 Dyrcona Either that, or log messages are being duplicated.
12:51 Dyrcona Well, it definitely came in twice, because I have two different traces, with four total messages from the gateway. All with the same source IP and only 1 session deleted from that IP at that time.
12:53 Dyrcona OK... I see why the duplication: 1 is ACT and the other INFO
12:53 collum joined #evergreen
12:54 collum_ joined #evergreen
12:55 Dyrcona So, I've got 4 from 1 session delete, 2 from another, and 1 from a third.  The third does not coincide with a segmentation fault.
13:04 collum joined #evergreen
13:08 Dyrcona Nope. Miscounted, the numbers are 5, 1, and 1.
13:13 Dyrcona Oh! Other drone crashes correspond on the second drone server.
13:14 Dyrcona So, that last one did cause a drone segfault.
13:14 Bmagic I love it
13:15 Bmagic (finding the answers)
13:42 csharp dbs: not sure if it's the same problem, but some of our dojo weirdness ended up being caused by our nginx proxy not being properly configured to forward the origin IP - might be useful :-)
13:42 csharp "kinda working" is always hard to diagnose
13:51 * Dyrcona is reminded that there are branches and commits that he should review but has been bogged down in log files....
13:52 dbs csharp: always worth keeping in mind!
13:53 * dbs currently wondering how to turn a yaz-client query for pubdate that works ("find @attr 1=31 @attr 4=4 2010") into something that works in Z39.50 search
13:54 HomerPublic joined #evergreen
13:54 dbs ah there we go, set config.z3950_attr code=31, format=4, truncation=100
13:55 dbs kinda wish we had aligned the column names in that table to the Bath Profile names at http://www.ukoln.ac.uk/interop-focus/activiti​es/z3950/int_profile/bath/draft/stable1.html but ah well
14:06 gmcharlt jihpringle: I confirm that the person wearing my mask earlier is correct ;)
14:06 jeffdavis I still want to know why miker had that mask handy.
14:07 gmcharlt dbs: that strikes me as something we could fix up almost completely transparently (assuming that few if any folks have been writing clark kent reports on the config.z3950_source table)
14:07 csharp jeffdavis: we all have gmcharlt masks - do you not?
14:07 jeffdavis csharp: I guess I need to start attending the conferences.
14:07 csharp jeffdavis: or the hackaway!
14:08 csharp fwiw, I've never made a report with that source
14:08 jihpringle thanks gmcharlt
14:09 jihpringle I've tested both new patches on our test server and so far it looks good
14:11 gmcharlt cool
14:17 dbs gmcharlt: those would be some weird reports indeed!
14:18 gmcharlt dbs: "but whatever is wrong with me using Evergreen as my platform for hosting an IRSpy clone?" ;)
14:23 dbs csharp: are you using "proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;" or "proxy_set_header X-Forwarded-For $remote_addr;"
14:23 Moggers1986 joined #evergreen
14:35 Dyrcona So, now, I've found a segfault that does not appear to correspond to an auth timeout, unless the session delete happened on another brick... Is that possible?
14:35 * Dyrcona double checks.
14:37 Dyrcona Indeed. My first cross-brick crash.
14:38 Dyrcona This was a different search, too, asc, not pgt.
14:38 nfBurton joined #evergreen
15:06 csharp dbs: proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for
15:06 Dyrcona So, I have found two so far that don't seem to correspond to open-ils.auth.session.delete calls, but 1 has two session retrieves right before it: 1 with an authtoken and 1 with null for the authtoken.
15:07 Dyrcona Two segfaults, that is.
15:07 dbs csharp: okay, that matches what we're doing. thanks though!
15:07 Dyrcona dbs | csharp: I'm using the same on my test systems with nginx. I will likely use haproxy in production to replace ldirectord, IF I ever get the time.
15:08 dbs Some of the Dojo UIs were replaced with Angular UIs in the last couple of releases, or am I dreaming?
15:09 berick dbs: they were
15:09 Dyrcona They have to be enabled, though, don't they?
15:09 Dyrcona I mean, explicitly, by the admin.
15:10 berick no, there's no enabling
15:10 berick they're just swapped
15:11 dbs berick: sweet!
15:12 dbs Was there a noticeable speed (or at least perception of speed) increase?
15:12 berick oh no doubt
15:12 dbs awwwwwesome
15:20 berick dbs: just did an exaggerated experiment on evdemo.kcls.org (cross country) -- Angular page DOMContentLoaded at 646ms.  Dojo page at 21 seconds.
15:20 berick the latency really impacts the dojo stuff a lot more, of course.  disparity wouldn't be that bad on a nearby server
15:33 dbs oh that's going to make so many people happy here
15:33 * dbs eyes acquisitions
16:03 jeffdavis Our most distant library is over 2000 miles from the data centre.
16:04 jeffdavis (Speaking of cross-country latency experiments.)
16:08 Dyrcona Chasing NULL pointers is worse than chasing rabbits.
16:09 Dyrcona And, speaking of Canadia [sic], my wife and daughter are in Quebec province at the moment.
16:13 Dyrcona @later tell JBoyer I believe I have found the crashing bit in oils_pcrud, though its actually in oils_sql.c.
16:13 pinesol Dyrcona: The operation succeeded.
16:17 jvwoolf left #evergreen
16:38 csharp @quote add <Dyrcona> Chasing NULL pointers is worse than chasing rabbits.
16:38 pinesol csharp: The operation succeeded.  Quote #200 added.
16:40 gmcharlt those wascally nullits
16:41 Dyrcona "But, if you go...chasing pointers..."
16:46 Dyrcona "And you know it's gonna null...."
16:46 Dyrcona :)
16:50 Dyrcona "Tell 'em a bearded, older programmer has given you a patch..."
16:50 Dyrcona "poor pcrud...when it has a segfault..."
16:51 * Dyrcona hopes he never runs into Grace Slick.
16:52 Dyrcona Although, that would be cool...
17:00 finnx joined #evergreen
17:00 pinesol News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>
17:02 finnx left #evergreen
17:08 mmorgan left #evergreen
17:12 Dyrcona So, I'm partly wrong and partly right about where the segfault occurs, and the partly wrong could make a better fix easier....
17:12 Dyrcona Even though I'm technically done for the day, I think I'll play around with this some more.
17:23 Dyrcona libosrf_json could probably stand to be its own library.
18:10 Dyrcona So, i'm trying to write a client in C and I'm having trouble with the contextnode argument to the bootstrap functions. Nothing I've tried so far works.
18:38 Dyrcona OIC.... took me long enough.
18:40 Dyrcona Eh.. no. It still fails to bootstrap when I claim to srfsh.
18:48 * Dyrcona shakes his head in shame....
18:49 Dyrcona Returns 1 on success, zero on failure...
18:51 Dyrcona THAT took WAY too long...
18:51 * Dyrcona signs out and gives it a rest for tonight.
19:36 RyanMcg joined #evergreen
19:37 RyanMcg Quick query for the Great brain!  Is there a way to force the evergreen webclient (Server says 3-1-11) to default to the maximum number of horizontal rows on data displays such as bills, items out, items checked out?  Our current default is only 25 rows, and we don't want to have to remember to reset that every time we start the web client.
19:38 RyanMcg I'm looking for a workstation/local administration method, not a server-side solution for the issue
19:43 RyanMcg Or is it not possible aside from as a server side change, or not at all?
19:47 sandbergja joined #evergreen
20:37 book` joined #evergreen
21:54 sandbergja joined #evergreen

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