Evergreen ILS Website

IRC log for #evergreen, 2017-10-25

| 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
01:01 tsbere_ joined #evergreen
03:06 tsbere__ joined #evergreen
04:52 jeff__ joined #evergreen
06:01 pinesol_green News from qatests: Test Success <http://testing.evergreen-ils.org/~live>
06:56 dwgreen joined #evergreen
07:19 kmlussier joined #evergreen
08:04 collum joined #evergreen
08:21 bos20k joined #evergreen
08:25 collum_ joined #evergreen
08:38 mmorgan joined #evergreen
08:41 rlefaive joined #evergreen
09:09 mdriscoll joined #evergreen
09:14 Dyrcona joined #evergreen
09:15 _adb joined #evergreen
09:31 jeff_ joined #evergreen
09:32 yboston joined #evergreen
09:41 Dyrcona berick: Just for grins, I changed my ejabberd shaper setting to 5 million from 500,000. Didn't help with the disconnects on that action_trigger.
09:43 csharp Dyrcona: I had an A/T template that was killing OpenSRF drones because the reltype was "might_have" in the fieldmapper when it needed to be "has_a" - sounds like you're deeper down than I got into the OpenSRF problem, but thought it's worth a mention
09:44 Dyrcona csharp: What action trigger?
09:44 Dyrcona If you recall...
09:44 csharp in my case it was PO JEDI
09:45 csharp bug 1702489 describes the problem I was trying to fix
09:45 pinesol_green Launchpad bug 1702489 in Evergreen "Wrong Join Type In Acq Lineitem Detail Causes Inaccurate Reports" [Low,New] https://launchpad.net/bugs/1702489
09:46 csharp but fixing the report broke JEDI message creation
09:46 Dyrcona Oh, right. I remember you talking about that.
09:47 csharp not sure if it has any bearing on your current problem, but I see some similarities ;-)
09:48 Dyrcona I see your change caused other issues for you.
09:48 csharp but I've seen other problems in A/T where the perl doesn't check that there's a value and flies off a cliff anyway
09:48 csharp the "other issues" were the broken PO JEDI template
09:53 rlefaive joined #evergreen
09:57 Dyrcona I'll have to see if I can run open-ils.trigger with the Perl debugger, though I've not had any luck with that since OpenSRF 2.4 or so.
09:58 mmorgan Dyrcona: Are you having trouble with any other action triggers, or just the pre-expiration courtesy notice?
09:59 Dyrcona mmorgan: Just this one. I'm leaning toward csharp's explanation or something like it.
09:59 Dyrcona Which also points to a serious bug somewhere in the Perl code, I would think.
10:02 berick in csharp's case, the Perl code was referencing a null value.  in Dyrcona's case, the jabber socket sysread() is returning 0 bytes, meaning the socket was closed from the remote end.
10:03 berick something is making ejabberd mad
10:03 Dyrcona berick: Or that is what it looks like....I suspect crappy Perl code.
10:03 berick that's what causes that exception.
10:04 Dyrcona I also have a simpilar instance loading a report template where the client closes the connection saying a timeout was reached....and that after about 2 seconds of waiting.
10:04 berick 'Disconnected from Jabber server' == ejabberd closed our socket
10:04 berick maybe there's a way to increase ejabberd logging
10:05 berick not saying Perl's not to blame, just that all we know for sure at this point is jabber is mad, maybe at some xml the perl code is generating that's not valid.
10:06 berick another option is to set loglevel to 5 -- that will show the XML we're sending to the jabber server
10:06 berick (and it will slow everything down, beware)
10:06 Dyrcona It's a test vm. No one else uses it.
10:06 berick you could also up the log level, but only restart trigger
10:07 berick so you're not slowing /everything/ down
10:07 csharp berick: ooh - that's a good idea - I never thought of handling debug that way
10:07 Dyrcona Meh. I restart services, truncate the logs, then just run the action trigger runner.
10:08 csharp I've been wondering if there's a better way to debug in a non-all-or-nothing way
10:08 mmorgan berick++ #Neat trick!
10:08 csharp also haven't done enough poking to know which logs to set to debug and which I don't need
10:18 Dyrcona berick: I've got about 5.8MB of log, where should I paste/post it?
10:19 Dyrcona I do see a rather odd (to me) message. It's a mass of JSON that looks like multiple objects thrown together, but of course it is cutoff.
10:20 Dyrcona It's just a giant array, but it's not the last message sent before the drone is kicked out.
10:21 berick i think that would be the "bundling"
10:21 pastebot "Dyrcona" at 64.57.241.14 pasted "Looks like the culprit?" (14 lines) at http://paste.evergreen-ils.org/906
10:23 Dyrcona Still no smoking gun. Just unceremoniously disconnected.
10:23 Dyrcona you want the bundled message, too? It's just before what I pasted.
10:24 berick hm, i really thought it was going to log the xml
10:24 berick looking at the osrf code, apparently it doesn't.  arg.
10:25 berick (probably for the best under normal circumstances)
10:27 jvwoolf joined #evergreen
10:27 berick Dyrcona: would you be up for adding a log line and restarting?
10:27 pastebot "berick" at 64.57.241.14 pasted "diff --git a/src/perl/lib/Open" (12 lines) at http://paste.evergreen-ils.org/907
10:28 berick basically ^--
10:28 berick (only needed for open-ils.trigger)
10:28 rlefaive_ joined #evergreen
10:28 berick .. to be restarted
10:29 Dyrcona Later. I have two meetings starting now.
10:30 berick re: cut-off, may have to modify logging to write to a file instead of syslog to avoid message truncation
10:30 berick Dyrcona: *nod*
10:37 Dyrcona berick: this is loggin to osrfsys.log not syslog.
10:37 berick Dyrcona: oh, great.  if it's still cutting off, i think there's an opensrf_core.xml variable to modify that
10:38 berick yeah, defaults to 1536 bytes.  can be modifed via <loglength>...</> in opensrf.xml
10:38 berick re, opensrf_core.xml
10:38 berick typing--
10:39 Dyrcona ok. I'll try that.
10:39 Dyrcona I'll try your pasted change after my meetings/lunch.
10:39 Dyrcona berick++
10:40 berick the Slack effect: being careless w/ typing in IRC forgetting you can't edit your messages.
10:40 berick Dyrcona++
11:08 pinesol_green [evergreen|Remington Steed] Docs: Remove duplicate lines from root.adoc - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=94787a3>
12:09 khuckins joined #evergreen
12:26 Dyrcona berick: I can increase the ejabberd loglevel from 4 to 5.
12:27 Dyrcona So, I'm doing that, too.
12:32 berick Dyrcona: cool
12:33 Dyrcona With other changes to opensrf_core.xml, just starting services and restarting apache2 puts 99MB in onsrfsyslog. :)
12:34 berick heh, not too surprised
12:34 Dyrcona I increased loglength to 10k and patched xmppreader.pm
12:34 * berick gets the popcorn
12:34 Dyrcona I've made different versions of opensrf_core.xml.<loglevel> and link them to opensrf_core.xml
12:34 Dyrcona So, I can switch settings easily.
12:35 Dyrcona I'm also looking at another issue with report templates and I was told in my meeting that is more important, so I'm going to gather log data for that issue first.
12:39 jihpringle joined #evergreen
12:54 Dyrcona 2017-10-25 12:46:36.834 [debug] <0.822.0>@ejabberd_c2s:send_text:1847 Send XML o
12:54 Dyrcona n stream = <<"<stream:error><policy-violation xmlns='urn:ietf:params:xml:ns:xmpp
12:54 Dyrcona -streams'></policy-violation><text xml:lang='' xmlns='urn:ietf:params:xml:ns:xmp
12:54 Dyrcona p-streams'>XML stanza is too big</text></stream:error>">>
13:01 khuckins_ joined #evergreen
13:05 berick heh
13:06 Dyrcona So, that's why open-ils.trigger is being kicked out, but I know what I have to do for the time being on the utility server.
13:06 berick i wonder if we're at the point we need to revert to our old stance on xml stanza sizes to be safe in the docs.
13:07 berick though I guess that would just allow the issues to linger
13:07 Dyrcona Yes and yes. :)
13:07 Dyrcona I won't change it on my development system so I can hunt things down there.
13:08 Dyrcona I did change the bricks last week, but I didn't change the utility server because I didn't see the stanza is too big message there.
13:09 berick yeah, i'm confused by when/where/why it displays sometimes but not others
13:12 Dyrcona Well, when you see all of the \\\ in the ejabberd logs.....
13:14 csharp I've seen the behavior, but not the message in our logs
13:14 Dyrcona The number of \ in front of " can be abitrary: 0 for the outer wrapper, then 1, then 3 and sometimes only 2.
13:14 csharp ("stanza is too big")
13:14 berick Dyrcona:  my favorite is unapi.bre xml embedded on osrf json message embedded in a containing json array embedded in xmpp xml.
13:14 Dyrcona csharp: It doesn't show up in ejabberd now unless you set the ejabberd loglevel to 5.
13:15 Dyrcona Note the [debug] in the line I pasted.
13:16 sandbergja joined #evergreen
13:16 Dyrcona Oh, that's nice. I don't recommend doing that for very long.
13:16 Dyrcona <username>opensrf</username><​password>password</password>
13:17 Dyrcona There's my 'super secret' password in the clear.
13:17 bshum Fun
13:17 Dyrcona For the logs, I'm lazy and just use password on my test vms because they're behind a firewall.
13:18 Dyrcona There is a lot of "unnecessary" chatter going on.
13:19 csharp Dyrcona: ah - good to know
13:19 Dyrcona <<"orcester Tatnuck Magnet School\\\",
13:19 Dyrcona That looks like an example of chunking working correctly.
13:20 Dyrcona The missing W is at the end of the previous message.
13:20 Dyrcona Or, if that's not how it's suppose to work.... ;)
13:34 Dyrcona So, I think I'll upload this 6MB of osrfsys.log to the Lp bug. I think miker had asked me to share it if I could.
13:35 Dyrcona It is surprisingly small consider the log settings.
13:35 Dyrcona Or, should i set loglength to the same as my ejabberd max_stana_size so the message that is too big will be clearly truncated?
13:41 kmlussier I am testing something with floating, but I'm unable to get the basic floating behavior to work. I think I may be missing a step.
13:41 kmlussier I edited a copy to use the default 'Everywhere' rule. If I check that copy in at another library, it should just stay there, right? Or do I need to do something else?
13:42 kmlussier It keeps putting the copy into transit when I check it in at another location. :-/
13:43 jwoodard joined #evergreen
13:47 kmlussier Hmmm...I can get it to work on my 2.12 server. Maybe there's a bug.
13:48 * Dyrcona has never really looked into the floating implementation.
13:54 kmlussier Oh, never mind. There was a hold on the title that the system was trying to fill.
13:55 kmlussier This is what happens when I try substituting tea for coffee.
13:56 kmlussier I should have known better than to try this on a server where I've been placing a gazillion holds at one time.
14:02 Dyrcona :)
14:18 kmlussier OK, now that the copy alert field has been added, I'm seeing another instance of what I reported here. https://bugs.launchpad.net/eve​rgreen/+bug/1720878/comments/2
14:18 pinesol_green Launchpad bug 1720878 in Evergreen "Button to apply copy tags is always disabled on some systems" [Undecided,New]
14:19 kmlussier Basically, the field is disabled because it didn't exist back when I created my workstation defaults for what fields should be enabled in the copy editor.
14:51 gmcharlt reminder - meeting regarding the 3.1 release manager in 10minutes
15:00 gmcharlt #startmeeting Evergreen 3.1 release manager selection
15:00 pinesol_green Meeting started Wed Oct 25 15:00:57 2017 US/Eastern.  The chair is gmcharlt. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00 pinesol_green Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00 pinesol_green The meeting name has been set to 'evergreen_3_1_release_manager_selection'
15:01 gmcharlt #topic Introductions
15:01 gmcharlt please introduce yourself using #info
15:01 bshum #info bshum, Ben Shum, free agent of chaos
15:01 gmcharlt #info gmcharlt = Galen Charlton, Equinox
15:01 jeff #info jeff = Jeff Godin, Traverse Area District Library (TADL)
15:01 remingtron #info remingtron = Remington Steed, Hekman Library (Calvin College)
15:01 dbwells #info dbwells = Dan Wells, Hekman Library (Calvin College)
15:01 abneiman #info abneiman is Andrea Buntz Neiman, EOLI
15:04 gmcharlt ok, great
15:04 gmcharlt #topic Release manager proposal
15:05 gmcharlt #info We have received a proposal from Dan Wells to be release manager for 3.1
15:05 gmcharlt #link http://libmail.georgialibraries.org/piperm​ail/open-ils-dev/2017-October/010457.html
15:05 berick #info berick Bill Erickson, KCLS
15:06 gmcharlt so, do folks have questions for dbwells regarding his proposal?
15:06 kmlussier #info kmlussier is Kathy Lussier, MassLNC
15:06 * dbwells braces for the deluge
15:10 berick not a question, just a comment that I appreciate the points in your RM proposal.  had there been many other candidates, you'd probably still get my vote :)
15:10 gmcharlt second call for questions, as dbwells is suspiciously dry
15:11 dbwells berick: thank you for that :)
15:12 gmcharlt OK, I"ll call the question
15:12 gmcharlt #vote Shall Dan Wells be given the Official Duck and be release manager for Evergreen 3.1? Vote Yes, No, Abstain
15:12 gmcharlt #startvote Shall Dan Wells be given the Official Duck and be release manager for Evergreen 3.1? Vote Yes, No, Abstain
15:12 pinesol_green Begin voting on: Shall Dan Wells be given the Official Duck and be release manager for Evergreen 3.1? Valid vote options are Vote, Yes, No, Abstain.
15:12 pinesol_green Vote using '#vote OPTION'. Only your last vote counts.
15:12 kmlussier #vote yes
15:12 remingtron #vote Yes
15:13 jeff #vote Yes
15:13 bshum #vote yes
15:13 abneiman #vote yes
15:13 berick #vote yes
15:13 gmcharlt #vote Yes
15:13 jeffdavis #vote yes
15:15 gmcharlt #endvote
15:15 pinesol_green Voted on "Shall Dan Wells be given the Official Duck and be release manager for Evergreen 3.1?" Results are
15:15 pinesol_green Yes (8): kmlussier, abneiman, jeff, berick, gmcharlt, jeffdavis, remingtron, bshum
15:15 gmcharlt #agreed Dan Wells shall be the release manager for Evergreen 3.1
15:15 gmcharlt dbwells++
15:15 jeff dbwells++
15:15 remingtron dbwells++
15:15 jeffdavis yay!
15:15 berick dbwells++
15:15 berick gmcharlt++
15:16 gmcharlt #endmeeting
15:16 pinesol_green Meeting ended Wed Oct 25 15:16:28 2017 US/Eastern.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)
15:16 pinesol_green Minutes:        http://evergreen-ils.org/meetings/evergr​een/2017/evergreen.2017-10-25-15.00.html
15:16 pinesol_green Minutes (text): http://evergreen-ils.org/meetings/evergr​een/2017/evergreen.2017-10-25-15.00.txt
15:16 pinesol_green Log:            http://evergreen-ils.org/meetings/evergree​n/2017/evergreen.2017-10-25-15.00.log.html
15:16 gmcharlt thanks, folks
15:17 dbwells Thank you, all.  My first order of business will be send out my proposed timeline dates by the end of this week, then we'll go from there.
15:17 kmlussier dbwells++
15:17 jeffdavis I'm excited to see where we get with the Angular move. It's a big undertaking, glad you're pushing for it.
15:21 dbwells jeffdavis: Yeah, after having passed through this stage without any wild howls from others, I am going to add it now to the proposed Hack-a-way discussion list.
15:57 csharp dbwells++ # sorry I missed the vote
16:04 yboston left #evergreen
16:10 dbs dbwells++ # also sorry for having missed the vote
16:42 Dyrcona When you've got code like this in the staff client: var req = new Request(OILS_RPT_FETCH_FOLDER_DATA,  SESSION, this.type, this.folderNode.folder.id(), oilsRptOutputLimit);
16:42 Dyrcona Where does Request come from? I can't remember and nothing I've turned up in my digging looks right.
16:44 berick Dyrcona: maybe web/opac/common/js/opac_utils.js
16:45 Jillianne joined #evergreen
16:45 Dyrcona That's it. Thanks!
16:45 Dyrcona berick++
16:45 Dyrcona I was looking in OpenSRF and in staff_client and web/js/
16:47 Dyrcona What's odd is that call is giving me "Error: TypeError: res is null" when everything looks good on the backend.
16:55 khuckins__ joined #evergreen
16:57 Dyrcona It's like it is timing out but no time out is set.
17:00 Dyrcona Oh well. Maybe I'll figure it out tomorrow.
17:01 jvwoolf left #evergreen
17:09 mmorgan left #evergreen
17:39 khuckins_ joined #evergreen
18:01 pinesol_green News from qatests: Test Success <http://testing.evergreen-ils.org/~live>

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