Evergreen ILS Website

IRC log for #evergreen, 2018-03-07

| 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
06:38 csharp gmcharlt++ # will take a look!
07:05 rjackson_isl joined #evergreen
07:26 agoben joined #evergreen
07:29 Dyrcona joined #evergreen
07:36 rlefaive joined #evergreen
08:17 * Dyrcona needs to edit better when posting stuff in web text boxes.
08:18 Dyrcona But, I maintain that the "web" is a lousy platform.
08:18 Dyrcona Yeah, I know....."Then implement a better one...."
08:20 Dyrcona And, how many parts should a single copy have?
08:26 collum joined #evergreen
08:27 Dyrcona I'll say it again, parts--
08:37 mmorgan joined #evergreen
08:39 kmlussier joined #evergreen
08:47 ngf42 joined #evergreen
08:48 Dyrcona kmlussier: Are you actively looking into Lp 1411422 ?
08:48 pinesol_green Launchpad bug 1411422 in Evergreen 2.12 "Copy details repeated in search results when item/volume moved with parts attached" [Medium,Confirmed] https://launchpad.net/bugs/1411422
08:48 kmlussier Dyrcona: No. You can remove me.
08:48 Dyrcona OK. I will and I'll assign myself.
08:49 Dyrcona I think fixing that will fix Lp 1753835.
08:49 pinesol_green Launchpad bug 1753835 in Evergreen "Merge Record Assets Can Fail Because of Duplicate Copy Part Map Entries" [Undecided,New] https://launchpad.net/bugs/1753835 - Assigned to Jason Stephenson (jstephenson)
08:49 Dyrcona I plan to test it on a 2.12.8 system today.
08:54 kmlussier Dyrcona++
08:54 jvwoolf joined #evergreen
08:58 Dyrcona Also, how many parts should a copy have?
08:58 Dyrcona I asked earlier, but I don't think anyone was paying attention.
09:00 mmorgan Dyrcona: a copy should be mapped to only one part.
09:01 Dyrcona mmorgan: Then, we have bigger problems. The copy that I used as an example has 9 parts in asset.copy_part_map.
09:02 Dyrcona And a 1:1 relationship should not be mapped in a separate table, IMHO. It should be a nullable field on the copy.
09:02 Dyrcona parts--
09:04 kmlussier Well, a part can have multiple copies.
09:05 Dyrcona Right, but I have the same target_copy in that table at least 9 times.
09:05 mmorgan Yes, what kmlussier said, but a target_copy should appear in asset.copy_part_map once
09:06 kmlussier Dyrcona: Sure, I get that. I was just explaining why there is a separate table.
09:06 mmorgan Just checked our production system and found six instances where a target_copy appears twice in the copy_part_map. Not too terrible...
09:07 Dyrcona Nope, that's not a reason for a separate table.
09:08 Dyrcona If A copy has a relationship with 1 and only 1 part, then part should be a field in copy that can be nullable in the case of there not being a part.
09:08 Dyrcona The above does not preclude different copies having the same part ID.
09:10 Dyrcona A reason for a separate table would be a 1 to many or many to many relationship.
09:10 Dyrcona If a copy should have more than 1 part, for example.
09:14 jvwoolf1 joined #evergreen
09:16 * Dyrcona has 203 copies with more than 1 part. Looks like they're all 2. Guess I misspoke about that copy being in there 9 time.
09:17 Dyrcona I'll have to work out a query later to see if the parts are both on the same bib as the copy.
09:17 Dyrcona If one is on the copy's bib and the other on a different bib, it's easy to fix.
09:18 Dyrcona If they're on the same bib, then a cataloger will have to fix it, if they even can.
09:19 Dyrcona The part must have been in the table 9 times. I haven't looked at that one specifically since yesterday.
09:20 * mmorgan would be very surprised if any of the multiple parts on the same copy are on the same bib.
09:22 kmlussier Yeah, they probably are there because the copies were transferred from one record to another, which is the problem bug 1753835 should fix.
09:22 pinesol_green Launchpad bug 1753835 in Evergreen "Merge Record Assets Can Fail Because of Duplicate Copy Part Map Entries" [Undecided,New] https://launchpad.net/bugs/1753835
09:22 kmlussier Sorry. Wrong bug. The other one.
09:22 * kmlussier is only half paying attention and should just keep quiet. :)
09:22 csharp @band add Wrong Bug
09:22 pinesol_green csharp: Band 'Wrong Bug' added to list
09:23 Dyrcona @band add Broken as Designed
09:23 pinesol_green Dyrcona: Band 'Broken as Designed' added to list
09:24 csharp @band add Monographic Parts
09:24 pinesol_green csharp: Band 'Monographic Parts' added to list
09:27 mmorgan csharp++
09:29 mmorgan1 joined #evergreen
09:30 Dyrcona IDK: Monograph and the Parts has more of a '60s vibe.
09:31 mmorgan2 joined #evergreen
09:31 csharp Dyrcona: true
09:40 Dyrcona @band add Monograph and the Parts
09:40 pinesol_green Dyrcona: Band 'Monograph and the Parts' added to list
09:40 Dyrcona @blame [band] for parts
09:40 pinesol_green Dyrcona: I know it was you, Violent Scratchy Amputees. You broke Dyrcona's heart. You broke Dyrcona's heart. for parts
09:43 * mmorgan just mentions that in our consortium that over 660 part level holds have been filled since the beginning of the year.
09:43 mmorgan parts++
09:46 Dyrcona I'm not opposed to the idea of parts. I think the implementation leaves a bit to be desired.
09:46 Dyrcona Maybe that can be fixed for 3.2.
09:50 csharp gmcharlt: fix for bug 1721807 worked for one of my failed templates - going to try another
09:50 pinesol_green Launchpad bug 1721807 in Evergreen "Web client report template errors when using sources with virtual fields" [High,Confirmed] https://launchpad.net/bugs/1721807 - Assigned to Galen Charlton (gmc)
09:50 kmlussier I think getting that bug fix in will help. I feel bad I didn't get around to testing it once the test was added. But lots of our features have bugs or could have been implemented better. There's always room for improvement.
09:50 kmlussier If there wasn't, I wouldn't have a job.
09:50 kmlussier Or, I should say, I might not have this particular job.
09:51 Dyrcona :)
09:51 gmcharlt of course, all cats likely prefer that we switch to full-time cat-petting
09:51 gmcharlt *our cats
09:52 Dyrcona Well, I think dropping asset.copy_part_map and moving part to the copy table would solve a number of issues with parts.
09:52 Dyrcona It solves this one, at least.
09:52 * mmorgan would definitely consider full time cat petting...
09:52 Dyrcona gmcharlt: You were probably right about all cats, domestic ones anyway.
09:53 kmlussier My cat prefers that we leave him alone 90% of the time.
09:53 kmlussier But during that other 10% of the time, if we're not engaged in full-time cat-petting, we're in trouble.
09:54 dbwells Dyrcona: I recall copies potentially having more than one part being discussed during the design of the feature.  Use case being, for example, a CD with art book, you might have a 'Disc' and a 'Book' part, with one library assigning both parts to one copy (to circulate together), and another library assigning each part to a separate copy.  AFAIK, there is no interface support for this (yet), though.
09:55 Dyrcona dbwells: I recall that, too, but if it hasn't happened yet, then I don't think it ever will.
09:55 Dyrcona Plus, I don't think parts are actually used in that manner. They seem to circulate as individual things.
09:57 Dyrcona And, how would that work with holds? You place a hold for the book and get the CD, too?
09:57 Dyrcona I don't see the practical use in that case.
09:57 Dyrcona But, we can hash that out when (if) I ever open a Lp bug about it. :)
09:58 dbwells I think that is the implication, and I think some might be happy to get the "part" they want even if it includes other parts.
09:58 kmlussier Dyrcona: Yeah, I think that's how it would work.
09:58 Dyrcona Yeah, but what purpose do the two parts serve if they circulate as 1 unit?
09:59 Dyrcona Can't you just catalog the book, and say "includes CD?"
09:59 gmcharlt csharp: cool
09:59 miker re copy part mapping, a copy may be labeled with a different part on a peer bib. or, I recall that being a thing
10:00 Dyrcona Parts are typically used for series and multi-volume sets, and I think that was the original intention.
10:00 dbwells Dyrcona: Sure, a single library can do whatever they want, it's working together that's the trick :)
10:00 Dyrcona peer bibs.... hmm....
10:00 Dyrcona That's something to think about. miker++
10:01 Dyrcona dbwells: That's the biggest problem with parts and software won't solve it. Seems a lot of our members do their own thing.
10:01 Dyrcona Different labels and different ways of splitting things up.
10:03 kmlussier We've kicked around an idea for years of making the system smarter about parts. For example, knowing that an item packaged with disc 1 to 3 together could fill a hold for disc 1, 2, or 3. It's something I still would like to see happen.
10:05 kmlussier It might also be a step towards a feature of allowing the system to automatically fill holds in disc order. When disc 3 is returned, the system automatically sends out the next disc in a series.
10:05 dbwells Fancy! :)
10:26 bshum Hmm, on the downloads page for 3.0 instructions, the README we get there now shows the new webpack instruction of "npm run build-prod / npm run test" and not the original "grunt all" step
10:26 bshum Did we replace the 3.0 instruction file incorrectly?  Or was that actually backported?
10:26 bshum The README in rel_3_0 doesn't show that change
10:27 Dyrcona bshum: That's supposed to be fore 3.1, not 3.0.
10:27 Dyrcona I think the wrong README is on the downloads page.
10:27 bshum So possibly the generated HTML is probably linked
10:28 bshum I'll generate a new one and put it into place...
10:28 Dyrcona bshum++
10:28 bshum Using the rel_3_0 version and updated for 3.0.4
10:29 bshum Oh yeah... it's the stock rel_name in the readme linked
10:29 bshum THat's got to be generated from master, not 3.0
10:30 Dyrcona Good catch.
10:31 bshum Just happened to be looking at it, since a little bird asked me a question about 3.0 vs. 2.12
10:34 Dyrcona Tweet! Tweet!
10:34 Dyrcona That little bird should have just used git to answer its question in the first place.
10:37 kmlussier bshum++
10:40 bshum Okay, that's better now on the README for 3.0
10:41 bshum Plus side, we got dbs change for that link back to the developer additions now too, whee
10:42 * bshum goes back to his snow day
11:19 Christineb joined #evergreen
11:26 * dbs reassures his weird parts that nobody's really judging them
11:28 kmlussier :)
11:28 berick dbs++ # heh
11:29 kmlussier People who weren't in IRC yesterday are probably really confused right now.
11:29 berick @who hired [band] to play dbs' house party?
11:29 pinesol_green kmlussier hired Nuclear Football to play dbs' house party.
11:32 csharp dbs++
11:34 csharp @who is going to the [band] vs. [band] extravaganza? ([band] will also be opening)
11:34 pinesol_green esoterica is going to the Dang Old Banjo vs. National Donut Day extravaganza? ( Classic Honey Black will also be opening.
11:37 esoterica My money's on the old banjo
11:40 csharp "Dang Old Banjo" reminds me of the old SNL ad for "Big Fat Bean" (Why eat a lot of little beans when you can eat one big bean?)
11:46 * berick prepares a giant rice cylinder for his evening repast
11:55 Jaswinder joined #evergreen
11:57 Jaswinder Guys, I am new to Evergreen library. I need help with the installation. I have setup Evergreen and its prerequisites on my Vagrant box. However, I am unable to access it using the web interface.
11:58 Jaswinder Can anyone help me troubleshoot this issue?
11:59 rlefaive joined #evergreen
11:59 Bmagic Jaswinder: sure, what do the logs say?
12:03 Jaswinder Here is the thing, I do not see anything in the logs
12:03 Bmagic Can anyone give some insight on thsi explain analyze? https://explain.depesz.com/s/ctqw
12:03 Bmagic Jaswinder: can you ssh into the machine?
12:04 Jaswinder What is the default URL that I can use to access via web?
12:04 Jaswinder Yes, I can
12:04 Bmagic Jaswinder: you should be able to open the machine using it's IP address in the URL on a web browser
12:05 Jaswinder Also, know that I have Apache, Apache websocket, and nginx as proxy servers are runnin
12:05 Jaswinder I tried that but it did not work
12:05 khuckins joined #evergreen
12:05 Jaswinder The request times out.
12:05 Bmagic I've not used the vagrant box for Evergreen but if it's using the default install paths, you should have some logs in /openils/var/log/osrfsys.log
12:06 Bmagic but, if you are getting timed out on the browser then it might be a firewall issue rather than an Evergreen issue. Can you confirm that Apache is getting the request when you hit it with a browser?
12:07 jihpringle joined #evergreen
12:07 Jaswinder Firewall is not an issue since i have other websites running fine with nginx
12:07 Jaswinder Also, no logs in /openils/var/log/osrfsys.log
12:08 Bmagic You have other websites running on the same box?
12:08 Bmagic Does that directory exist? (/openils/var/log) ?
12:08 Jaswinder Yes, I do
12:08 Jaswinder Yes, it does exists
12:09 Bmagic I think for troubleshooting, we might want to get all of the Evergreen services on an isolated machine. Evergreen runs a full stack of services and I can imagine it wouldn't play well with others on the same machine.
12:10 Jaswinder It shouldn't matter as other setup is simply a PHP website with MySQL backend
12:11 Bmagic jeff: are you here? jeff__ - I know he runs some machines with PHP in tandum but I don't think it's on the same machine
12:11 Jaswinder I can try setting this up on a separate box if needed but I want to avoid that.
12:12 Bmagic Jaswinder: If it were me, I would first try to get it running on it's own machine.
12:12 Jaswinder Can you share the recommended apache, websocket and nginx port settings/
12:12 Jaswinder I need all three. Correct?
12:13 Bmagic The minimum recommended settings are in the installation walkthough
12:13 Bmagic https://evergreen-ils.org/docume​ntation/install/README_3_1.html
12:13 Jaswinder I am using 3.0 version
12:13 Bmagic search that page for "MaxConnectionsPerChild"
12:14 Bmagic https://evergreen-ils.org/docume​ntation/install/README_3_0.html
12:14 Jaswinder That is 500
12:14 Bmagic the other apache settings are there next to it
12:14 Bmagic but that isn't your problem if you are getting timed out
12:15 jeff Bmagic: hmm?
12:15 Bmagic jeff: wasssup
12:15 jeff Bmagic: greetings.
12:15 Bmagic so, php
12:15 Jaswinder I see.
12:16 jeff Bmagic: were you saying that you thought we/I run php and Evergreen within the same Apache instance?
12:17 Bmagic jeff: Jaswinder said he is running a machine with Evergreen and he is having timeout issues getting the OPAC to load. He said that the same machine has other websites as well. PHP.
12:17 Bmagic jeff: yeah, I thought I remember you having something running PHP but didn't think it was on the same Evergreen machine. Do we have anyone that has LAMP running along side Evergreen?
12:18 Jaswinder Just to clarify my setup before the Evergreen install was Ubuntu 16.04, Nginx, MySQL, and PHP 7.1.
12:19 Jaswinder Nginx is listening for port 80. Does this pose any issues with Evergreen setup?
12:19 jeff Bmagic: Sorry, doesn't apply to us, and I would personally avoid running non-Evergreen sites/services under the same Apache instance as  Evergreen. I can't guarantee that it would have problems, but I wouldn't be surprised if it did.
12:19 Bmagic Jaswinder: what jeff said
12:20 Jaswinder I see. Alright thanks for responding. I will spend sometime on this and if it does not work, I will fire up another vagrant box. Thanks again!
12:20 Bmagic sure
12:21 Bmagic I have some long running queries that I am trying to work out. One of them is a patron search. Why is this taking 24 seconds? https://explain.depesz.com/s/ctqw
12:22 Jaswinder Aside from this issue, could you please tell me the role of Nginx with Evergreen setup?
12:23 dbwells Jaswinder: Nginx isn't necessary for running Evergreen.
12:24 dbwells I think if folks use it, they are using it as a proxy.
12:24 Jaswinder Okay, So the main webserver is the Apache?
12:24 jeff But it is one of the recommended options for having your websockets connections share port 443 with your normal https traffic, rather than needing a distinct websockets port.
12:24 jeff (it in my last message being nginx)
12:25 dbwells Jaswinder: Yes, Apache is required.
12:25 Jaswinder I see. I may now know where to look to resolve this issue. Thanks!
12:27 dbwells Yeah, I would certainly keep Nginx on the sideline for an initial install, it possible.
12:28 dbwells s/it/if/
12:28 Bmagic Jaswinder: also try the docker container https://hub.docker.com/r/m​obiusoffice/evergreen-ils/ if you are just trying to get your feet wet
12:52 beanjammin joined #evergreen
12:59 csharp we hit a formatting issue in our lib_ips.txt (for in-library OPAC/staff client IP redirection to the correct location), and I'm pondering what it would take to get that into the DB
12:59 * csharp is reading https://www.postgresql.org/docs/9​.5/static/datatype-net-types.html
12:59 csharp seems like a simple mapping table would suffice
12:59 csharp miker: or berick: or phasefx: are there reasons lib_ips.txt is a file outside the DB?
13:01 miker csharp: main reason is that it's super static, so there's no need to put the DB under more load for an IP-to-id mapping
13:01 csharp (for those unfamilar, this file contains lines with shortname, start IP, and end IP for ranges and is parsed by a perl module as apache starts up)
13:02 csharp miker: ok - makes sense, and yes, it's not changed often
13:03 csharp it would definitely add DB load given how often the data is consulted
13:04 miker csharp: even if that were in the db as configuration, I'd lobby for it to be made static somehow. a file is easier to update than, say, a memcache key (today), so have autogen.sh rewrite it, maybe
13:06 miker though there's an opportunity here for building an "invalidate known memcache keys" thing, and get rid of several files
13:06 miker depends on tuits available
13:08 Jaswinder Bmagic - will checkout docker
13:13 Jaswinder One more thing that I need your suggestion on. I am tasked to extend the Evergreen library by create a module/plugin that will call external APIs to return data. Is there a documentation on writing plugin for Evergreen that I should follow?
13:14 jeff Jaswinder: are you able to give a less-general example of what you are trying to build?
13:14 jeff Jaswinder: i.e., something more specific / detailed?
13:16 Jaswinder Sure
13:18 yboston joined #evergreen
13:18 Jaswinder I want to make another library database available with Evergreen. There will API calls that will retrieve the books data based on search
13:19 Dyrcona joined #evergreen
13:21 kmlussier Jaswinder: Are you saying that you will have a database of materials in Evergreen and another library database of materials that should be searchable?
13:21 kmlussier If so, you might want to take a look at FulfILLment.
13:22 Jaswinder Yes
13:22 Jaswinder The idea is to extend Evergreen to call both existing db and external API to search and display books
13:24 Jaswinder Is there a documentation to develop Evergreen plugin? Which language it is written?
13:26 Dyrcona Evergreen doesn't have "plugins" in the way that term is normally used.
13:26 Dyrcona You could add new features via new Perl modules rather easily.
13:28 Dyrcona Well, I say easily....but....
13:28 Jaswinder Is there any documentation that I use? Also, should I setup Evergreen using binary or using the source code for development?
13:29 Dyrcona Jaswinder: To get Evergreen setup you pretty much have to start with source code, unless you're using csharp's deb packages.
13:29 Dyrcona Most of it is Perl, so most of it is source anyway.
13:30 Dyrcona As for documentation...There is some but the best things are kind of scattered about.
13:30 Dyrcona Let me look...
13:31 kmlussier Do we have plans for a dev meeting today?
13:32 Dyrcona Well, it's on the calendar but I don't think anyone started an agenda.
13:32 berick heh, just came to ask about that
13:32 miker kmlussier: I won't be available for the entire 3pm hour (not that that should stop everyone else...)
13:33 Dyrcona And, http://docs.evergreen-ils.org/3.0/ is giving me a 403
13:34 Dyrcona Jaswinder: Here's a link to the official documentation for release 2.12 with information about developing for/with Evergreen: http://docs.evergreen-ils.org/​2.12/_developer_resources.html
13:35 Dyrcona And, the 3.0 link for that gives me a 404.
13:37 Dyrcona Jaswinder: This is a good introduction to the main technologies used in Evergreen: http://evergreen-ils.org/~denials/workshop.html
13:38 Dyrcona Some of the commands in the latter have been superseded.
13:38 Dyrcona The ones for starting and stopping services in particular.
13:38 * Dyrcona has to get bower installed, apparently.
13:39 remingtron Dyrcona: I'll email Robert about the docs errors
13:39 Dyrcona 2.12 still uses bower, right?
13:39 Jaswinder Thanks Dyrcona
13:40 Dyrcona remingtron++
13:40 Dyrcona Jaswinder: You're welcome.
13:41 kmlussier Dyrcona: Yes, 2.12 continues to use bower.
13:41 Dyrcona Hrm. My alleged 2.12.8 branch is hosed and looks like master.
13:42 Dyrcona Think I'll roll the VM back and start over.
13:42 Dyrcona I bet I messed up the checkout and the remote branch is ok.
13:46 Dyrcona Oh. "Clean" is cleaner than I thought....
13:52 Dyrcona Think I'll make another snapshot after cloning git.
14:00 collum_ joined #evergreen
14:25 pinesol_green [evergreen|Remington Steed] Docs: Fix '&' DocBook conversion error - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=df8ada1>
14:27 khuckins_ joined #evergreen
14:29 khuckins__ joined #evergreen
14:31 remingtron_ joined #evergreen
14:39 collum joined #evergreen
14:46 kmlussier Bmagic: I might be misunderstanding your question on the list, but does 0cd235b do what you need?
14:47 kmlussier hmm,...did I do that wrong?
14:48 berick commit 0cd235b
14:48 berick huh
14:48 berick 0cd235b8e8cfef608d2b27560a746591dd978706
14:49 berick @blame pinesol_green
14:49 pinesol_green berick: itself forgot to give the gerbils their chocolate-frosted sugar bombs
14:49 kmlussier Oh, I know what I did wrong.
14:50 kmlussier fd7f9041
14:50 pinesol_green kmlussier: [evergreen|Dan Pearl] LP#1308090 Relator fields and facets need normalization. - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=fd7f904>
14:50 kmlussier I was using the commit from the working repo.
14:51 berick ah
14:51 khuckins_ joined #evergreen
15:07 kmlussier dbwells++
15:08 gmcharlt dbwells++
15:11 pinesol_green [evergreen|Remington Steed] Docs: Use raw ampersand '&' to fix conversion error - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=631d2c7>
15:20 mmorgan1 joined #evergreen
15:25 Bmagic kmlussier: huh?
15:26 kmlussier Bmagic: Oh, sorry. That wasn't you.
15:26 kmlussier My brain doesn't seem to be working properly today.
15:26 Jaswinder joined #evergreen
15:28 * kmlussier redirects the question to Stompro about whether commit fd7f9041 will address his facet / browse question.
15:28 pinesol_green kmlussier: [evergreen|Dan Pearl] LP#1308090 Relator fields and facets need normalization. - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=fd7f904>
15:29 remingtron Dyrcona: FYI, the docs should be fixed now.
15:30 Dyrcona Thanks! I wanted point Jaswinder at the latest docs.
15:30 remingtron @blame remingtron and the fancy ampersands
15:30 pinesol_green remingtron: remingtron and the fancy ampersands is really just another name for autogen
15:30 Stompro kmlussier, thanks, I'll take a look at that.
15:33 Stompro kmlussier++ that seems to address the problem we are seeing.  Thanks.
15:48 miker kmlussier: highlighting branch also helps, incidentally, fwiw
15:48 miker new mods-level normalizations
15:48 kmlussier miker: Good to know!
16:18 Bmagic jeff++
16:31 Dyrcona I've got 12,926 parts that are not deleted and do not have an entry in asset.copy_part_map. Is it safe to just delete them?
16:34 Dyrcona Well, it's a test database for a reason.
16:52 Dyrcona Hmm... It's too late in the day to mess with th code. Guess I'll leave it to tomorrow.
18:06 eby does anyone have a sec to run an explain analyse for a hold query. i'm trying to figure out if we have a postgres config issue
19:05 dbwells_ joined #evergreen
20:24 jeff eby: i'm willing to try, but not sure if version mismatch will torpedo the usefulness.
20:25 eby jeff: thanks. i'll pm
21:14 jvwoolf joined #evergreen
22:12 jvwoolf left #evergreen

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