Evergreen ILS Website

IRC log for #evergreen, 2016-06-15

| 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
07:02 agoben joined #evergreen
07:09 graced good morning #evergreen
07:10 graced @coffee
07:10 * pinesol_green brews and pours a cup of Ethiopia Konga Co-op Yirgacheffe, and sends it sliding down the bar to graced
07:18 rjackson_isl joined #evergreen
07:35 rlefaive joined #evergreen
07:45 JBoyer joined #evergreen
07:51 gsams_ joined #evergreen
08:01 collum joined #evergreen
08:02 rlefaive joined #evergreen
08:18 ericar joined #evergreen
08:23 mrpeters joined #evergreen
08:29 kmlussier joined #evergreen
08:50 mmorgan joined #evergreen
08:52 Dyrcona joined #evergreen
09:07 collum joined #evergreen
09:09 kmlussier Is the info in bug 1297967 still valid? We still need to use berick's custom branch to install EDI?
09:09 pinesol_green Launchpad bug 1297967 in Evergreen "document openils-mapper code for enriched EDI" [Undecided,New] https://launchpad.net/bugs/1297967
09:11 sam_l joined #evergreen
09:12 bos20k joined #evergreen
09:23 Dyrcona kmlussier: I believe so, if you can get EDI to install at all.
09:24 kmlussier Dyrcona: Thanks! Why wouldn't you be able to get EDI to install at all?
09:25 Dyrcona kmlussier: Incompatible ruby versions, particularly on Ubuntu 16.04.
09:25 kmlussier I thought we fixed that with bug 1342227
09:25 pinesol_green Launchpad bug 1342227 in Evergreen 2.8 "Setting up EDI Fails with Ruby version > 1.8" [Undecided,Fix released] https://launchpad.net/bugs/1342227
09:26 Dyrcona Doesn't work on Ubuntu 16.04, 'cause no backports were available when I tried in April.
09:26 maryj joined #evergreen
09:27 kmlussier :(
09:28 Dyrcona berick is working on making the ruby unnecessary, and I've been meaning to look at that branch, but....time.
09:28 kmlussier Yeah, I was just looking at that. I wish I knew how to test it.
09:29 Dyrcona Right now, you just run the new code on an order and compare the output to what's in the database.
09:30 Dyrcona And, monthly maintenance releases are on the calendar for today.
09:31 kmlussier I can take care of release notes as soon as I send a couple of e-mails off.
09:32 * jeff reviews bug 1592565 with an eye toward having it in 2.10.today
09:32 pinesol_green Launchpad bug 1592565 in Evergreen "2.10 Login code lost key activity log message" [Undecided,New] https://launchpad.net/bugs/1592565
09:32 jeff oh.
09:32 yboston joined #evergreen
09:32 jeff i have another one that should go in.
09:32 jeff oof.
09:32 Dyrcona jeff: better hurry. :)
09:32 jeff exactly.
09:34 Dyrcona Installing all of that ruby stuff is so clunky.
09:35 Dyrcona Sometimes, it is a good idea to reinvent the wheel. :)
09:40 kmlussier Once berick's branch is done, will there be a need to install EDI separately or will it just come with a stock Evergreen install?
09:40 * kmlussier really hopes it's the latter.
09:40 Dyrcona Should be the latter.
09:41 kmlussier Yay!
09:41 Dyrcona Dunno if the plan is to enable it by default or not.
09:41 Dyrcona I would think you'd only want to enable the services if you're using EDI.
09:42 kmlussier Yeah, that makes sense
09:53 Callender joined #evergreen
09:54 Dyrcona So, for those following along, I think we fixed our public OPAC woes by adding another server just to handle the Z39.50 supercat requests.
09:55 Dyrcona We had a spare vm host hanging around because whatever it was doing got moved or was no longer necessary.
09:55 Dyrcona I set that up as a stand alone Evergreen installation talking to our production database.
09:56 Dyrcona Once I had that working, I changed the oils_z3950.xml zurl parameter to use the new server.
09:56 Dyrcona Since I did that, load has been more stable on our public OPAC an no loss of service.
09:58 Dyrcona What I noticed when we had real problems, such as load going over 100, is that we were already busy and then would get several z39.50 searches a few seconds aparat.
09:58 Dyrcona s/aparat/apart/
10:00 berick kmlussier: Dyrcona: yes, my branch removes the need to install ruby or run the webrick service.  edi is generated directly via perl.  as it stands, it's enabled by telling CRON to run a different script.
10:00 berick so people can ease into it
10:00 kmlussier Excellent!
10:01 * jeff stops watching ansible work and instead goes to get coffee && water
10:01 kmlussier @praise berick
10:01 * pinesol_green berick is kind and patient to newbies
10:02 JBoyer berick++
10:02 JBoyer All of the delays in getting Acq going here are actually a /feature/. We're helping people, they just don't know it.
10:03 Dyrcona Ha!
10:04 mmorgan1 joined #evergreen
10:09 mmorgan joined #evergreen
10:11 csharp is webby.evergreencatalog.com relatively up-to-date?
10:13 * csharp works under the assumption that it is
10:13 dbs Dyrcona++ # good finding! maybe share that on a performance tips and tricks page or something?
10:14 csharp berick: awesome - looking to test your EDI branch in the coming weeks
10:14 kmlussier csharp: My understanding is that it has the latest web client code that's in master plus the webstaff-sprint3 branch
10:14 eady joined #evergreen
10:24 Dyrcona dbs: I may just do that if time permits.
10:26 csharp kmlussier: awesome - thanks
10:26 berick csharp: great.  if you (and anyone else who tests) find local tweaks that are not represented in the code, let me know.  i'll add edi attrs for them.
10:35 collum joined #evergreen
10:46 collum joined #evergreen
11:04 collum joined #evergreen
11:05 jeff @rss
11:05 pinesol_green jeff: (rss <name|url> [<number of headlines>]) -- Gets the title components of the given RSS feed. If <number of headlines> is given, return only that many headlines.
11:06 jeff @rss list
11:06 pinesol_green jeff: Error: Couldn't get RSS feed.
11:06 jeff @rss tests
11:06 pinesol_green jeff: Error: Couldn't get RSS feed.
11:06 * jeff takes it to msg
11:06 jeff rss qatests
11:06 jeff @rss qatests
11:06 pinesol_green jeff: 11:06 AM, June 15, 2016: Test Failure - http://testing.evergreen-ils.org/~live/test.html <http://testing.evergreen-ils.org/~live/test.html>
11:08 maryj joined #evergreen
11:12 Dyrcona Not much to do about "database "evergreen" is being accessed by other users" in the code.
11:12 rlefaive joined #evergreen
11:15 jeff on Debian Jessie, "make check" for Evergreen is failing for me unless I manually manipulate LDFLAGS in Open-ILS/src/c-apps/tests/Makefile
11:15 jeff I needed to add -lpthread -lm -lrt
11:17 jeff (which is obviously not the way to correct the problem long-term)
11:17 kmlussier It looks like the perl live tests started failing on May 31?
11:18 Dyrcona jeff: I would think configure would pick those up.
11:19 Dyrcona Do the tests actually relink/recompile code?
11:20 Dyrcona kmlussier: Those live test failures could be because the database was not rebuilt.
11:20 kmlussier Yeah, I was just coming around to that.
11:20 Dyrcona Many of the tests fail after being run once.
11:21 * Dyrcona tries to write his so that they can be run more than once on the same database, but may not always succeed.
11:24 bmills joined #evergreen
11:27 jeff Dyrcona: i think "make check" compiles and links things that are not normally built -- somewhat similar for perl tests would be if Test::More could not be found, etc.
11:28 jeff it compiles and links some small test binaries.
11:29 Dyrcona OK. I've not looked at it very much. I often just run the pgtap or perl tests individually.
11:29 Dyrcona I haven't run make check in a while.
11:32 dbs Yeah, 'make check' is for C tests
11:37 jeff okay, i'm done testing bug 1592565 and i'm ready to write release notes and pick to master and rel_2_10. Am I correct in thinking that I only need to put release notes in master and they'll be backported to rel_2_10 just prior to the next point release being cut?
11:37 pinesol_green Launchpad bug 1592565 in Evergreen "2.10 Login code lost key activity log message" [High,In progress] https://launchpad.net/bugs/1592565 - Assigned to Jeff Godin (jgodin)
11:40 Dyrcona jeff: Point release is today, but I haven't heard anything from gmcharlt today, so don't know what he thinks about pushing something in right now.
11:41 jeff or am I off and bugfixes get zero release notes, and are just summarized manually... looks like that's the case.
11:41 gmcharlt jeff: correct re summary
11:41 jeff Dyrcona: thanks -- i was going to verify with gmcharlt before pushing anything. :-)
11:41 gmcharlt jeff: and re that change, I've no objection to including it now
11:42 kmlussier jeff: I'm working on point release notes now and will incorporate it.
11:42 kmlussier Or, really, I'm working on release notes when other things don't pull my attention away.
11:43 jeff There is also a rather annoying SIP2 related bug in 2.10 that I'm writing up and throwing a working branch together for, FYI. I'd neglected to notice that today was point release day, or I'd have made more of a point to get it in before today.
11:44 pinesol_green [evergreen|Bill Erickson] LP#1592565 Log successful authtoken and workstation - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=edaee08>
11:56 jeff and i lost my bug draft because i failed to compose it in something other than my browser.
11:56 * jeff corrects that mistake
12:11 ericar joined #evergreen
12:12 sam_l EG Configuration question - I'm still working on setting up a development environment, and finally have an actual error explaining my 500 server error when accessing the site - Apache is unable to parse XML config file /openils/conf/opensrf_core.xml when bootstrapping the gateway for requests. Halp?
12:13 jeff sam_l: i've often found the xmllint tool useful there.
12:14 sam_l jeff: thanks.
12:15 jeff xmllint --noout /openils/conf/opensrf_core.xml
12:15 jeff you can install it via the libxml2-utils package on most debian distros
12:15 brahmina joined #evergreen
12:15 jeff you'll either get zero output from xmllint --noout, or you'll see an error like:
12:15 jeff /openils/conf/opensrf_core.xml:2: parser error : Start tag expected, '<' not found
12:16 jeff hello<!--
12:16 jeff ^
12:16 jeff the "2" there is the line number.
12:19 dteston joined #evergreen
12:22 sam_l Nope, no errors in the XML.
12:23 bshum Permission maybe
12:23 bshum Is the run user for apache set to opensrf?
12:24 bshum Or is the file owned by opensrf?
12:26 sam_l Just double checked - run user is opensrf, everything in /openils is owned by opensrf
12:31 Dyrcona sam_l: Did you configure opensrf_core.xml after installing OpenSRF and again after installing Evergreen?
12:34 sam_l Dyrcona: I skipped the initial configuration and didn't configure opensrf_core.xml until Evergreen was installed. That doesn't seem like it should break things, but I won't be surprised if it does.
12:34 Dyrcona sam_l: That won't break things unless the configuration is wrong.
12:34 sam_l Dyrcona: Configuration checks out - I can run srfsh commands.
12:35 bshum sam_l: What versions are we talking about?  For OpenSRF and Evergreen, and Linux distro?
12:36 sam_l Opensrf 2.4.1, EG 2.10.4, Ubuntu Trusty for this iteration.
12:39 bshum Looks like rlefaive had a similar issue on http://irc.evergreen-ils.org/​evergreen/2016-04-26#i_243991 ; I wonder if it's something based on the Ubuntu image
12:39 bshum sam_l: What 14.04 image did you use?  The 64-bit server edition?
12:40 * bshum checked the logs but didn't see a resolution noted by rlefaive or others on later days
12:45 jeff gmcharlt: bug 1592891 is also a candidate for 2.10.5. i can write tests also, if you'd like.
12:45 pinesol_green Launchpad bug 1592891 in Evergreen "SIP2 failures with certain kinds of user standing penalties" [High,New] https://launchpad.net/bugs/1592891 - Assigned to Jeff Godin (jgodin)
12:45 jihpringle joined #evergreen
12:46 kmlussier Release notes are done, but I'll hold off on merging the 2.10 notes until I hear back on whether the SIP2 fix makes it in.
12:46 jeff kmlussier++
12:46 kmlussier Dyrcona: You're not expecting to merge anything else for 2.9, are you?
12:47 Dyrcona No, I'm not.
12:47 kmlussier OK, then, I'll get those off my plate now.
12:48 sam_l bshum: Yep, 64-Bit Server.
12:48 Dyrcona jeff: Good catch on that penalties bug.
12:49 jeff Dyrcona: thanks! sorry i didn't bug it until now -- sorta' fell into a SIP2 rabbit hole.
12:59 gmcharlt jeff: I agree it's a candidate. go forth and write tests!
12:59 jeff doing so! :-)
13:06 sam_l bshum: rebooted the server instance, and it's now running properly.
13:07 bshum sam_l: Oh cool, I was just about to suggest restarting apache
13:07 bshum If you hadn't yet.
13:07 * bshum wanders away for a bit
13:09 mmorgan1 joined #evergreen
13:10 Dyrcona kmlussier++
13:18 tarac joined #evergreen
13:33 JBoyer sam_l: In case you've not seen it elsewhere, apache always has to be restarted after opensrf is started/restarted or it will only give ISEs until you do.
13:33 * Dyrcona starts prep work for the 2.9.6 release.
13:33 Dyrcona JBoyer++
13:33 mmorgan joined #evergreen
13:39 jeffdavis I am finding that searches of our Evergreen system via Z39.50 are extremely slow, e.g. >13s for an author search at a specific org unit when the same search takes <2s in the OPAC.
13:40 jeffdavis Does anyone else experience this?
13:42 Dyrcona jeffdavis: Yes. Also, check the scroll back for my findings regarding the impact of z39.50 on OPAC performance.
13:43 Dyrcona I can't speak to your specific numbers, but I can attest that z39.50 searches are slower in general.
13:45 Dyrcona Though, it only took 4 secons for the following search on my new configuration: find @attr 1=1003 "paulo coelho"
13:45 berick i think it kinda depends on what Z server your talking
13:45 berick to
13:45 eady joined #evergreen
13:45 berick EG is just a pass-thru for Z searches
13:46 Dyrcona I understood jeffdavis to me searching his own evergreen system via z39.50 with the standard simple2zoom setup.
13:46 Dyrcona s/me/mean/
13:46 jeffdavis yes what Dyrcona said
13:46 berick oohhh
13:46 * berick butts out
13:47 Dyrcona jeffdavis: Have you setup up different "databases" in the Z config for the different scopes? We only search the whole consortium here.
13:47 jeffdavis Yeah, we have a number of different databases. The whole-consortium one is a bit faster, but still slow.
13:48 jeffdavis I have a use case that requires a site-specific one right now though.
13:48 Dyrcona If you have some spare hardware, you could try what I mentioned earlier and set up a self-contained Evergreen instance just for z39.50.
13:49 jeffdavis Our Z39.50 service is hosted on a separate utility server, not the main ones that are hit when people use the OPAC or staff client.
13:49 jeff gmcharlt, Dyrcona, anyone else interested: tests pushed and pullrequest tag added to bug 1592891
13:49 pinesol_green Launchpad bug 1592891 in Evergreen "SIP2 failures with certain kinds of user standing penalties" [High,New] https://launchpad.net/bugs/1592891 - Assigned to Jeff Godin (jgodin)
13:49 Dyrcona OK. That's the best way to set it up.
13:50 jeff and... removing myself from assigned.
13:50 Dyrcona jeff: Thanks. I will not have time to test it today.
13:50 jeff Dyrcona: no problem. didn't mean to imply that expectation by mentioning you. :-)
13:51 Dyrcona jeffdavis: Z searches go through multiple layers, so I expect it to be a little slower than regular search.
13:51 Dyrcona jeff: NP. Just saying....
13:52 jeff Dyrcona++ for writing tests worth copying, also.
13:52 Dyrcona jeffdavis: Is your zurl pointing at the utility server or back to your OPAC?
13:53 JBoyer I wondered about that, because we also run the Z stuff on a separate machine, but it currently points right back to the OPAC, which I'm planning to change soon.
13:53 Dyrcona The simple2zoom service is very lightweight. I found the supercat searches to be the real resource hogs.
13:54 jeffdavis Dyrcona: hm, points at the main OPAC, not the utility server (although I should double check to see how the util server resolves the domain name)
13:55 rlefaive joined #evergreen
13:55 Dyrcona If you can, just configure apache on the utility box and point your zurl at it, assuming the utility box runs its own services, too.
13:56 jeffdavis It does indeed. I'll give that a try. Thank you for the suggestion!
13:58 Dyrcona For more detail on what I saw when we'd get several z searches in the span of a minute or so, there would be something like 80 supercat, 40 to 60 cstore, and 40+ search drones running.
13:59 Dyrcona Now, when that happens during a period of otherwise high use, the load on the OPAC and sometimes also the drone server would skyrocket.
13:59 Dyrcona Since I've moved the target for the zurl to its own machine, that has not happened.
14:00 Dyrcona In fact, I've configured the drones the same, and we've not gone over 10 supercat drones on the zurl box.
14:00 Dyrcona At least, not when I was looking. :)
14:00 kmlussier @marc 534
14:00 pinesol_green kmlussier: Descriptive data for an original item when the main portion of the bibliographic record describes a reproduction of that item and the data differ. Details relevant to the original are given in field 534. (Repeatable) [a,b,c,e,f,k,l,m,n,p,t,x,z,6,8]
14:01 kmlussier @marc 775
14:01 pinesol_green kmlussier: The entry for another available edition of the target item (horizontal relationship). When a note is generated from this field, the introductory phrase Other editions available: may be generated based on the field tag for display. (Repeatable) [b,c,d,e,f,g,h,i,k,m,n,o,r,s,t,u,w,x,y,z,6,7,8]
14:01 Dyrcona I have a script to monitor the number of drones on the drone server, but haven't set it up on the zurl target.
14:01 Dyrcona And, look at that: I have a 2.9.6 tarball to test.
14:03 * Dyrcona should really document the z stuff
14:35 krvmga joined #evergreen
14:48 collum joined #evergreen
14:54 jeff force pushed a fixup commit to the tests on bug 1592891 (noticed a function was unnecessary, edited comments a bit)
14:54 pinesol_green Launchpad bug 1592891 in Evergreen "SIP2 failures with certain kinds of user standing penalties" [High,New] https://launchpad.net/bugs/1592891
15:15 pinesol_green [evergreen|Jason Stephenson] Forward port 2.9.5 to 2.9.6 upgrade script. - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=e1853b2>
15:17 Dyrcona gmcharlt: FYI: 2.9.6 tarball is uploaded to the server, but not in place, yet.
15:17 gmcharlt Dyrcona++
15:18 Dyrcona Whenever you're ready, we can release them together, or I can go ahead and update the download page for 2.9.6, whichever you prefer.
15:18 gmcharlt I will be doing the 2.10.5 cutting this evening; once I finish, I can update the downloads page for both 2.9.6 and 2.10.5 and do the blog post
15:19 gmcharlt (though I certainly no objection if you want to update the downloads page now)
15:19 gmcharlt *have no
15:19 Dyrcona Well, when the files are copied into place, I think the README gets autoupdated, at least the way that I do it.
15:20 gmcharlt yeah
15:26 jlitrell joined #evergreen
15:27 Dyrcona I've put the downloads files in place, except the readme and release notes.
15:39 jeff gmcharlt: are you intending / able to test and merge the sip bugfix, or would you like me to solicit a willing co-conspirator?
15:40 gmcharlt jeff: I am willing and able, but even happier if you find a willing co-conspirator to test and signoff first
15:43 jeff also, since i remember we had at least one time that we duplicated perl live test numbers, should i leave tests unnumbered?
15:43 jeff or rely on the merging party to re-number if needed at that time?
15:46 jeff in this case i numbered the test. just wondered if we had a general concensus on that for next time.
15:52 kmlussier joined #evergreen
15:53 kmlussier I don't think we have a general consensus. I would go with numbering for now.
15:56 Dyrcona Yeah, I agree.
15:57 Dyrcona The numbers are not that important, but it's nice to avoid duplicates.
16:05 maryj joined #evergreen
16:32 mllewellyn joined #evergreen
16:44 kmlussier gmcharlt: I'm going to merge what I have for 2.10 release notes now. Could you add whatever is needed for the SIP2 work if it gets merged?
16:45 gmcharlt kmlussier: sure
16:45 kmlussier gmcharlt: Thanks!
16:49 kmlussier hmmm...actually, it looks like I already added the SIP2 changes to the release notes and was just waiting for them to be merged. In that case, maybe I'll just push them to the working repo and let gmcharlt merge them when he's ready to cut the release.
16:50 gmcharlt +1
16:51 gmcharlt memcached and routers started for owwl210p/u
16:51 gmcharlt which means entirely nothing in this channel :)
16:51 gmcharlt *cough*
16:54 Dyrcona :)
16:56 kmlussier gmcharlt: working/user/kmlussier/2-10-5-release-notes
16:58 gmcharlt kmlussier++ # and that statements means something in all the channels!
17:00 jeff looks like we have "organizations whose employees contributed to this release" for 2.10.0 and "organizations whose employees contributed patches" for 2.10.3 but not the other point releases. Is there rhyme or reason there?
17:08 mmorgan left #evergreen
17:12 jeff HAH!
17:13 jeff just tripped over another fun bug in webstaff. catalog view breaks if (for example) your hostname starts with "staff" :-)
17:13 jeff found the problem, can fix. haven't even confirmed if it hasn't already been fixed.
17:13 * jeff -> out for a bit
17:38 jonadab joined #evergreen
17:50 eady joined #evergreen
18:09 jonadab joined #evergreen
18:19 gsams joined #evergreen
19:05 pinesol_green` joined #evergreen
19:07 phasefx_ joined #evergreen
19:10 jeffdavis I have run out of time to finish testing and signing off on bug 1592891; I'll try to get to it in a couple hours unless someone else gets to it, but not sure I'll make gmcharlt's cutoff for 2.10.5 :(
19:10 pinesol_green` Launchpad bug 1592891 in Evergreen "SIP2 failures with certain kinds of user standing penalties" [High,New] https://launchpad.net/bugs/1592891
19:11 jeffdavis biab
19:19 egbuilder joined #evergreen
19:32 jeff jeffdavis++ thanks for trying -- i know it was last minute!
19:35 gmcharlt I'll start work on the release shortly, and will give that patch a fair look
20:22 gmcharlt jeff: jeffdavis: patch suits me; it will be in the release
20:24 pinesol_green` [evergreen|Jeff Godin] LP#1592891: Tests - SIP2 standing penalty failures - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=56cc12e>
20:24 pinesol_green` [evergreen|Galen Charlton] LP#1592891: tweak tests - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=22536d1>
20:24 pinesol_green` [evergreen|Jeff Godin] LP#1592891: Fix SIP2 standing penalty failures - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=5e0c9b3>
20:27 jeffdavis gmcharlt++
20:27 jonadab joined #evergreen
20:30 pinesol_green` [evergreen|Kathy Lussier] Docs: Adding 2.10 Release Notes - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=67c6b2a>
20:38 pinesol_green` [evergreen|Galen Charlton] 2.10.4-2.10.5 schema upgrade script - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=9bfd8c8>
20:42 rlefaive joined #evergreen
21:09 rlefaive joined #evergreen
21:19 jeff gmcharlt++
21:22 gmcharlt 2.9.6 and 2.10.5  downloads now linked
22:30 rlefaive joined #evergreen

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