Evergreen ILS Website

IRC log for #evergreen, 2017-09-19

| 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
00:33 roycroft joined #evergreen
04:08 Jillianne joined #evergreen
05:00 pinesol_green News from qatests: Test Success <http://testing.evergreen-ils.org/~live>
05:35 rlefaive joined #evergreen
07:14 rjackson_isl joined #evergreen
07:15 rlefaive joined #evergreen
08:25 Jillianne2 joined #evergreen
08:26 kmlussier joined #evergreen
08:37 mmorgan joined #evergreen
09:00 bos20k joined #evergreen
09:06 jvwoolf joined #evergreen
09:08 _adb joined #evergreen
09:09 _bott_ joined #evergreen
09:10 Dyrcona joined #evergreen
09:13 Bmagic @coffee
09:13 * pinesol_green brews and pours a cup of Kenya Gethembwini, and sends it sliding down the bar to Bmagic
09:18 Bmagic Soooo delicious
09:19 * Dyrcona listens to morning maniac music.
09:21 _bott_ joined #evergreen
09:22 _adb joined #evergreen
09:26 rlefaive_ joined #evergreen
09:35 jvwoolf1 joined #evergreen
09:39 _adb joined #evergreen
09:40 _bott_ joined #evergreen
09:47 Dyrcona @tea
09:47 * pinesol_green brews and pours a pot of Golden Orchid, and sends it sliding down the bar to Dyrcona (http://ratetea.com/tea/whisper​ing-pines/golden-orchid/7244/)
10:01 rlefaive joined #evergreen
10:03 mmorgan1 joined #evergreen
10:11 collum joined #evergreen
10:23 Dyrcona emerge-files++
10:29 yboston joined #evergreen
10:35 Dyrcona @coffee everyone for National ${SOMETHING} Day.
10:35 * pinesol_green brews and pours a cup of Finca La Tacita, and sends it sliding down the bar to everyone for National ${SOMETHING} Day.
10:36 berick @praise Nation [someone] Day!
10:36 * pinesol_green Nation fbeaudry_ Day! is the hardest working person in #evergreen.
10:38 Dyrcona A little bird tweeting in a meadow told me that today is National IT Professionals Day, and I had a cynical reaction.
10:43 kmlussier It's also Talk Like A Pirate Day.
10:43 * fbeaudry_ tips her hat to pinesol_green
10:46 fbeaudry_ Aye, me hearties - tis indeed Talk Like a Pirate Day.
10:46 fbeaudry_ http://www.pirateglossary.com/
10:47 gmcharlt QUAARRK!
10:47 gmcharlt </quack_like_a_pirate_day>
10:49 fbeaudry_ gmcharlt++
10:49 abneiman gmcharlt++
10:49 abneiman actual LOL
10:50 Dyrcona @swill everyone for National Talk Like a Pirate Day
10:50 * pinesol_green grabs a can of Sparks and sends it sliding down the bar to everyone for National Talk Like a Pirate Day
10:55 csharp @quote add < gmcharlt> QUAARRK! </quack_like_a_pirate_day>
10:55 pinesol_green csharp: The operation succeeded.  Quote #177 added.
10:57 csharp @quote random
10:57 pinesol_green csharp: Quote #16: "<berick> why's it broken? just bugcause." (added by gmcharlt at 01:37 PM, October 12, 2011)
11:22 mmorgan joined #evergreen
11:33 Bmagic I am looking hard at some osrfsys.log logs - trying to detect why a certain action trigger is setting it to "error"
11:33 Bmagic I have a question, is it possible for an issue with a different event definition to cause another one to be set to "error"
11:33 Bmagic The logs seem to indicate that
11:35 Bmagic I have the line that updates the row with "state=reacting"
11:38 Bmagic Followed by some fleshing fields lines for another action trigger
11:40 Bmagic followed by a perm check for setting it to lost
11:40 Bmagic Then this line: Event reacting failed with No field by the name call_nuarcode in Fieldmapper::asset::copy! at /usr/local/share/perl/5.22.1/​OpenILS/Utils/Fieldmapper.pm line 273.
11:42 Bmagic searching the ated, I found the definition with the munched column name, but that's not related to the at that was marking something else lost, it's like the Thread crashed and affected several other ATE's
11:44 Dyrcona Well, if memory gets garbled somehow, no telling what can happen.
11:45 Dyrcona But, I've never seen that phenomenon before, so can't say.
11:45 Dyrcona You checked all copies of the fm_IDL?
11:46 Dyrcona Though that looks more like a code bug than IDL bug from the message.
11:46 Bmagic If it's possible for AT to error other AT's when one of them has an issue, then it would all make sense
11:47 Bmagic The error message is referencing something located in the configured template - I found the line in the ATED template and corrected it. It was perplexing because that particular ATED is not the ATED in question
11:50 Dyrcona Did you check the template for the ATED in question? It may be borked, too.
11:50 berick Bmagic: are the 2 related in any way -- e.g. checkout.due and lost.auto -- where one directly affects the creation of another?
11:50 Dyrcona And, there's that.
11:50 Bmagic berick: no, two different libraries in different system trees
11:51 Bmagic One was overdue notice (munched column name) and the other is marking lost
11:51 Bmagic The marking lost trigger doesn't have a template
11:51 gsams joined #evergreen
11:53 _adb joined #evergreen
11:54 berick Bmagic: what are the hooks?
11:56 Bmagic checkout.due on both
12:02 _adb joined #evergreen
12:07 berick just to be totally clear, none of this is threaded.  (i realize you may have been using the term loosely).  each event def is handled within its own process. if you're using parallel, it offloads work to drones that might be used by multiple event-defs over time, but they would never be working on 2 things at once.
12:08 berick i've never seen a template error in one process affect another process
12:09 Dyrcona Perl5 is not thread safe, and "threads" in Perl5 are a major kludge that blow up on all kinds of libraries, such as Encode.pm.
12:09 Dyrcona Threads are the only reason to switch to Perl6, but in that case, I'd switch to Python 3.
12:10 Dyrcona In short, we can't use Perl threads, so we don't.
12:10 Dyrcona The only way that I could see one process affecting another is bad RAM, but then you got bigger problems.
12:12 berick I have seen an oddity with the MarkItemLost reactor.  I have yet to collect meaningful data, though.  in short, (in parallel anyway) it can sometimes leave events stranded in the 'collected' state.  commenting out the call to generate events for 'lost.auto' from within the MarkItemLost reactor fixed it.  something about the combo of recursive A/T and parallel handling -- something goes haywire.
12:15 Dyrcona Race condition on the reactors and the db?
12:17 Dyrcona DB transaction gets committed before 1 process is really done with it and the other picks it up? (For berick's case, shouldn't happen with unrelated reactors.)
12:18 Bmagic The logs that I am looking at sure makes me feel like one ATED is affected the other
12:19 khuckins joined #evergreen
12:19 pastebot "Bmagic" at 64.57.241.14 pasted "ATED's eating eachother" (74 lines) at http://paste.evergreen-ils.org/816
12:27 Dyrcona What in the log makes you think that?
12:29 Dyrcona Oh....
12:30 berick Bmagic: you confirmed there's no message_template for this event-def?
12:30 Bmagic yes
12:30 berick and no call_nuarcode in the event environment?
12:30 Bmagic do you have any marklost triggers with templates?
12:30 Bmagic berick: right
12:31 Bmagic call_nuarcode is only found in another template
12:31 berick what about the event-def delay_field or usr_field?
12:31 Bmagic id 535
12:32 berick ?
12:33 Bmagic the ATED 535 contained the template with the munched column name
12:33 Bmagic not this one from the pastebin (id 579)
12:34 Bmagic looking at the ATED for 579, nothing unusual. delay=14 days, usr_field=null
12:34 berick and message_usr_path and message_library_path
12:35 Bmagic select * from action_trigger.event_definition where id=579 and template is null and usr_field is null and message_usr_path is null and message_template is null and message_library_path is null
12:35 Bmagic 1 row returned
12:36 berick k
12:36 Bmagic furthermore, I deleted all of the error rows from ATE, and I am running the same definition again, all completing with no errors
12:36 Bmagic no changes to the definition
12:37 Dyrcona Mabye you found a memory/process bug in docker? :P
12:37 Bmagic the utility machine is not running in docker
12:37 Dyrcona The kernel, then, but I doubt it.
12:38 Dyrcona I never do like wading though the mess of the logs, and I honestly can't say I'm convinced. I'm not dissuaded, either. Without seeing the templates, before, I can only speculate, so I'll bow out.
12:39 Dyrcona The message about call_nuarcode is definitely coming from process 13876, though, and that was processing an event with definition id 579.
12:40 Dyrcona Unless syslog messed up, but that's not likely, either.
12:41 Bmagic right, I noticed that the PID was the same, but there is no way that the event definition that it was working on had call_nuarcode anywhere
12:42 dwgreen joined #evergreen
12:43 berick Bmagic: event def 535 is also checkout.due?
12:43 Bmagic yes
12:43 Bmagic they both have the same hook in common
12:51 berick Bmagic: can you check your /openils/var/log/open-ils.trigger_stderr.log too?
12:51 berick starting to think the logged error is a red herring
12:51 rlefaive joined #evergreen
12:51 Bmagic log is empty
12:51 berick the call to set_item_lost did not complete
12:52 berick one notable oddity is
12:52 berick [2017-09-15 12:02:57] open-ils.trigger [INFO:13876:CStoreEditor.pm:139:] editor[1|1] request en-US open-ils.cstore.json_query.atomic {"from":["actor.org_unit_an​cestor_setting",null,273]}
12:52 berick an org setting lookup with no setting name
12:52 Bmagic interesting
12:53 berick I believe that is happening in AppUtils::get_copy_price()
12:53 Bmagic makes sense
12:54 berick there should be 3 AOUS lookup calls back to back there.  there's only 1 and it looks weird
12:55 berick suggesting it dies there
12:58 berick hm, there should also be a copy update call after void_overdue_on_lost lookup and before the first AOUS lookup in get_copy_price, but there's not one in the logs
12:59 * Dyrcona often finds things to be missing from the logs, particularly with syslog to a remote server.
12:59 * berick is confuzzled but unfortunately has to step away for a bit
12:59 Dyrcona Even after applying JBoyer's suggested configuration.
13:02 mmorgan Bmagic: Are you using a granularity or filter in your command to run the trigger?
13:03 Bmagic mmorgan: no
13:03 * mmorgan thought that without one of those, all the active triggers with the same hook would be looked at.
13:04 Bmagic mmorgan: It seems reasonable to me that these two triggers would be executing along side one another
13:04 Bmagic and if one crashed the perl code, it could* affect the other
13:07 Dyrcona No, it shouldn't. Each is a separate process on the O/S. It's not like there's one copy of Perl running both. It's a separate Perl process for each one.
13:07 Dyrcona I think you have data missing from the logs.
13:43 acautley joined #evergreen
14:12 Dyrcona joined #evergreen
14:15 kmlussier Bmagic / dbwells: We have a maintenance release scheduled for tomorrow. Are you both available to build? I should have release notes done first thing in the morning.
14:15 Bmagic sure
14:15 dbwells kmlussier: sounds fine.
14:16 kmlussier For the 2.11 release, it looks like we have just one bug fix. This would also be the last point release for 2.11, unless we have any security fixes over the next three months.
14:19 * Dyrcona needs a new laptop.... :(
14:20 kmlussier Would anyone have time to review the fix for bug 1706124 before tomorrow? It's small, but would be a nice addition to the other web client fixes going out with 2.12.6.
14:20 pinesol_green Launchpad bug 1706124 in Evergreen 2.12 "web client: include inactive option in patron search is not sticky" [Medium,Confirmed] https://launchpad.net/bugs/1706124
14:25 Dyrcona I can probably take a look if my laptop doesn't just shut off again.
14:25 gmcharlt upon some discussion - release cutting for 3.0-beta2 is going to be pushed back to Thursday, so tomorrow will be a full day for last minute stuff for the beta
14:26 gmcharlt and at this point, I'm satifisfied that we'll make the beta2 goal of having Debian Stretch support
14:26 gmcharlt as well as likely resolution of some more i18n issues
14:26 Dyrcona Yay!
14:26 kmlussier That's this week too? Time flies.
14:27 Dyrcona kmlussier: Are you looking for just a sign off or do you want it pushed?
14:28 kmlussier Dyrcona: If it looks good to you, go ahead and push it.
14:29 gmcharlt dbwells++ # flexibility
14:55 csharp @band add Call Nuarcode
14:55 pinesol_green csharp: Band 'Call Nuarcode' added to list
15:16 dbwells Did a little research, found this: "According to Evergreen legend, the call of the Nuarcode bug is a devastating shriek able to penetrate nearby but otherwise unrelated processes.  It has never been successfully logged."
15:18 Dyrcona Anyone know why we use Class:DBI::Frozen, and why we can't just use Class::DBI?
15:18 jvwoolf joined #evergreen
15:19 csharp dbwells++
15:19 csharp @quote add < dbwells> Did a little research, found this: "According to Evergreen legend, the call of the Nuarcode bug is a devastating shriek able to penetrate nearby but otherwise unrelated processes.  It has never been successfully logged."
15:19 pinesol_green csharp: The operation succeeded.  Quote #178 added.
15:20 csharp Dyrcona: I've wondered the same thing - was told by berick recently that frozen is still required
15:21 * csharp hasn't dug into the differences/why at all
15:21 Dyrcona Maybe I'll take a look in my not-so-copious free time.
15:21 Dyrcona It has been nagging at me every now and then.
15:22 csharp it nags at me every time we build a custom deb for Class::DBI::Frozen::301 ;-)
15:22 * Dyrcona waits a bit on npm install....
15:22 Dyrcona :)
15:23 dbwells Dyrcona: I think that might have come up in our dev docs research this past year, I'll check.
15:23 berick Dyrcona: to use regular Class::DBI requires EG code changes.  I don't know what those code changes are, though.
15:23 berick dbwells++
15:23 Dyrcona berick: Yes, I figured as much.
15:23 berick Dyrcona: glad I could be of service :)
15:24 Dyrcona :)
15:24 * csharp wonders if that's hack-a-way-able, then looks at all the other more pressing things
15:24 Dyrcona berick++ dbwells++ csharp++
15:24 Dyrcona I might take it on at the hack-away. I also want to make changes to our use of autotools in Evergreen.
15:25 Dyrcona There are things I think we could more canonically and that would help with portability.
15:25 csharp also way down on my TODO list is to ferret out all the C warnings during 'make' on Evergreen
15:25 berick i'd be up for some of that, csharp
15:25 csharp berick: awesome
15:25 Dyrcona Well, that's easy: CFLAGS+=-werror. :)
15:25 berick :)
15:26 csharp yeah, dh_make fails on those unless we loosen the rules to not care about warnings
15:27 Dyrcona Our technical debt has been on my mind lately.
15:28 * csharp googles "technical debt"; agrees with Dyrcona
15:29 jihpringle joined #evergreen
15:29 dbwells d252b2d019d00068efc8277447120704a5a4fa3e is when the change happened, but not a lot of details which couldn't be inferred.
15:29 pinesol_green dbwells: [evergreen|dbs] Provide support for Class::DBI::Frozen::301 via UNIVERSAL::require - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=d252b2d>
15:35 Dyrcona Well, that was inconsiderate of the Class::DBI maintainers.
15:36 Dyrcona Incompatible API changes should have warranted a 4.0 release.
15:36 Dyrcona Version numbers are meaningless. :)
15:55 dbwells Especially in a project which renumbered 0.96 to 3.0.0 :)
15:58 Dyrcona heh
15:58 berick hm, browser client console showing lovefield error.  error url takes me to..
15:58 berick Constraint error: (201) Duplicate keys are not allowed, index: Setting.pkSetting, key: "ui.staff.max_recent_patrons"
15:59 berick anyone else getting this?
15:59 Dyrcona So, I tested Lp 1708048 on trusty. make check succeeds for OpenSRF and Evergreen. All Evergreen live tests pass.
15:59 pinesol_green Launchpad bug 1708048 in OpenSRF "Add support for Debian 9 Stretch" [Wishlist,Confirmed] https://launchpad.net/bugs/1708048 - Assigned to Galen Charlton (gmc)
15:59 * berick will just try clearing localStorage
15:59 Dyrcona I also commented on the bug, but thought this might be quicker to reach the intended audience.
16:01 berick hm, had to delete the offline indexdb
16:02 * berick makes mental note to look at the DB next time before clearing
16:04 gmcharlt Dyrcona++
16:08 kmlussier berick: I've seen something similar, but not with max_recent_patrons.
16:09 berick kmlussier: now i'm seeing it w/ webstaf.format.dates.  guessing it has little to do with the actual setting.
16:11 acautley joined #evergreen
16:11 kmlussier berick: Yeah, I mentioned in #5 here - https://bugs.launchpad.net/ever​green/+bug/1706107/comments/25
16:11 pinesol_green Launchpad bug 1706107 in Evergreen "Offline mode for the Web Client" [Wishlist,Fix released]
16:13 kmlussier miker mentioned further in the bug that it would be worth looking into, but I think it fell off of my radar after that.
16:13 berick thanks, kmlussier
16:13 berick that's the one
16:14 berick ugh, it's preventing pages from rendering for me
16:14 berick once I delete the DB, it clears up
16:14 berick clearly that's not happening for everyone, though
16:25 Dyrcona Installing rel_2_12 on xenial, grunt all gives: PhantomJS 2.1.1 (Linux 0.0.0) ERROR  ReferenceError: Can't find variable: angular  at /home/opensrf/Evergreen/Open-ILS/web/j​s/ui/default/staff/services/core.js:6
16:27 Dyrcona I know I can --force but I'm trying to remember what broke it last time and what the fix is.
16:38 Dyrcona I get this from npm install, didn't see it before:make: Entering directory '/home/opensrf/Evergreen/Open-ILS/web/js​/ui/default/staff/node_modules/ws/build'  CXX(target) Release/obj.target/bufferutil/src/bufferutil.o bufferutil.target.mk:92: recipe for target 'Release/obj.target/bufferutil/src/bufferutil.o' failed
16:40 Dyrcona From googling a bit, looks like an out of date dependency caught up with us.
16:44 Dyrcona Right. I forgot bower install.
16:45 * Dyrcona walks away whistling.... :)
16:47 * kmlussier was about to say we don't install bower anymore, but then read up.
16:47 bshum Yeah I didn't notice till I re-traced the steps more carefully too
16:48 bshum More reasons that 3.0 will be the bestest Evergreen release soon enough
16:49 Dyrcona Yeah, it's still needed on 2.12....
17:01 pinesol_green News from qatests: Test Success <http://testing.evergreen-ils.org/~live>
17:04 csharp dammit! A/T is totally broken
17:05 csharp I don't have a good bead on what's happening yet, but something breaks and it's a wall of "No children available"
17:09 mmorgan left #evergreen
17:17 csharp we need better default (non-debug) logging for what happens when a drone is NOT CONNECTED TO THE NETWORK!!!
17:17 csharp all I see is lalala everything's fine, then boom, off the network
17:18 Dyrcona csharp: I've been seeing that from time to time on 2.10, still.
17:18 Dyrcona But, I'm signing out now, to go to the gym.
17:20 csharp even better, "hey, this drone is NOT CONNECTED TO THE NETWORK!!!, let's kill it and respawn instead of continuing to barf about it"
17:21 csharp <ralph_wiggum>Software is hard!</ralph_wiggum>
17:24 csharp hmm open-ils.trigger.event.gather: Deep recursion on subroutine "OpenSRF::AppSession::connect" at /
17:24 csharp usr/local/share/perl/5.18.2/Open​SRF/DomainObject/oilsMessage.pm line 329.
17:32 * csharp realizes that the rebuilt utility box post upgrade has half the RAM it did before
18:14 _adb joined #evergreen
18:16 _adb joined #evergreen
18:26 _adb joined #evergreen
18:28 _adb joined #evergreen
18:32 _adb joined #evergreen
18:44 acautley joined #evergreen
18:46 khuckins joined #evergreen
19:21 jvwoolf joined #evergreen
20:32 pinesol_green [evergreen|Kathy Lussier] LP#1706124: Make include inactive patrons checkbox sticky - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=3195aac>
20:32 pinesol_green [evergreen|Galen Charlton] LP#1706124: minor tweaks - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=4d202fc>
20:47 eady joined #evergreen
21:00 acautley joined #evergreen
21:49 acautley joined #evergreen
22:50 acautley joined #evergreen
23:45 acautley joined #evergreen
23:48 pinesol_green [evergreen|Ben Shum] LP#1629078: Add missing strings in web client for i18n - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=f92437b>

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