Evergreen ILS Website

IRC log for #evergreen, 2015-10-21

| 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
06:01 jeff_ joined #evergreen
07:34 mt-dev joined #evergreen
07:35 mt-dev Looking to help my school library deploy Evergreen. Are there any drawbacks or things that might prevent them from switching? They are currently using Destiny.
07:48 rjackson_isl joined #evergreen
07:57 jboyer-isl joined #evergreen
08:24 ericar joined #evergreen
08:30 ericar_ joined #evergreen
08:46 kmlussier Good morning #evergreen
08:46 kmlussier @coffee [someone]
08:46 * pinesol_green brews and pours a cup of Nectar, and sends it sliding down the bar to pinesol_green
08:46 kmlussier @tea [someone]
08:46 * pinesol_green brews and pours a pot of Huang Shan Mao Feng Green Tea, and sends it sliding down the bar to egbuilder (http://ratetea.com/tea/teavivre/hu​ang-shan-mao-feng-green-tea/6041/)
08:47 kmlussier I guess the bots need a little caffeine too.
08:55 mt-dev joined #evergreen
08:57 jeff_ morning, kmlussier
09:16 Dyrcona joined #evergreen
09:27 mt-dev Looking to help my school library deploy Evergreen. Are there any drawbacks or things that might prevent them from switching? They are currently using Destiny.
09:28 Dyrcona mt-dev: How big of a school library are you talking about?
09:30 mt-dev 600 students, not sure how many books: https://dccs.follettdestiny.com/catalog​ing/servlet/presentadvancedsearchredire​ctorform.do?l2m=Library%20Search&tm​=TopLevelCatalog&l2m=Library+Search
09:33 maryj joined #evergreen
09:35 csharp mt-dev: do you have much in the way of Linux expertise on the school IT staff?
09:35 csharp (particularly Linux server expertise/experience)
09:37 yboston joined #evergreen
09:38 Dyrcona csharp: Excellent question... I'll remember to ask that first. :)
09:42 mmorgan joined #evergreen
09:47 mrpeters joined #evergreen
09:48 csharp Dyrcona: :-)
10:01 yboston jboyer-isl: good morning. I wanted to let you now I live somewhat close to Boston's Logan airport. You can stay at my place on Friday the 6th
10:04 jboyer-isl yboston: I appreciate the offer! I’ve already got the room through the 7th, but I’ll definitely consider it since I don’t necessarily want to make kmlussier go all that way at WHAT o’clock.
10:05 kmlussier jboyer-isl / yboston: I should know by the end of the week how things are looking for Saturday transportation.
10:05 kmlussier yboston++
10:05 yboston jboyer-isl: you should be able to change the hotel reservations without penalty. BTW, I live a subway stop away from where we had the Boston/Cambridge conference
10:06 yboston kmlussier: are you coordinating the rides from Logan?
10:06 kmlussier Yes
10:07 yboston kmlussier: I was hoping to hitch a ride from Logan on Tuesday, but if it creates more work I can go with the ride you arranged for me on Wednesday morning
10:07 yboston kmlussier: what makes your life easier?
10:08 yboston kmlussier: I woudl prefer to be there Tuesday night/afternoon, but not if it makes mroe work for others
10:08 kmlussier yboston: DPearl is the one making the Tuesday afternoon airport runs. He has a minivan with a lot of space. I would recommend joining the late afternoon shuttle run. The early afternoon one has a lot of people.
10:09 kmlussier yboston: They all make my life easy. It turns out I'm transporting nobody so far. :)
10:09 yboston kmlussier: yes, I was thinking that the late afternoon ride would be less packed
10:10 kmlussier yboston: Do you have DPearl's contact information?
10:10 yboston jboyer-isl: let me know if you want me to email you my address, as well as taxi or subway transportiation options and costs
10:10 yboston kmlussier: I should eb able to track down hi email
10:10 yboston *his email
10:13 mt-dev joined #evergreen
10:16 jboyer-isl yboston: If I’m the only one leaving that early I think I will take you up on that so long as it’s not an imposition. If it turns out that one or two more people have to be there that early I may leave things as they are. So I know, is it possible to make it to BOS through the subway or will you need to drive me there/part way?
10:17 jboyer-isl (I don’t want to put out anyone, but if I can make it there via subway and you don’t have to be driving around at 5-6 AM that’s even better)
10:17 kmlussier jboyer-isl: Unless there are people attending who haven't identified themselves on the wiki yet, I think everyone who has to fly in has already put their travel plans in the spreadsheet.
10:18 jboyer-isl I was just about to ask about that.
10:18 kmlussier jboyer-isl: You can make it to the airport on the subway. You have to make one stop on the way to change trains.
10:18 yboston jboyer-isl: I live very close to the hotel where we had the cofnerence, just one more stop. So the same transportation instrucitons apply to get to Logan from my apt
10:19 yboston jboyer-isl: the subway (the "T") starts running at 5:30 AM, and the stop is a 10 minute walk from my apt. A cab ride woudl be around $35
10:19 kmlussier #announcement: If anyone is planning to attend the hack-a-way and hasn't put their name here - http://wiki.evergreen-ils.org/do​ku.php?id=hack-a-way-2015-agenda - please add your name, even if your plans are still up in the air.
10:19 jboyer-isl That sounds great. I think I’ll take you up on that yboston, thanks a lot. I have to talk to the hotel tomorrow anyway to get the travel card stuff squared away anyway.
10:20 jboyer-isl a 10 minute walk is a good excuse to eat junk food when I get to the airport. ;)
10:20 jwoodard joined #evergreen
10:20 yboston jboyer-isl: BTW, I don't have a car but we will have options
10:20 kmlussier heh
10:20 kmlussier Anything is a good excuse to eat junk food IMO
10:20 mt-dev Dyrcona, csharp: I'm in the 9th grade at the school :D The librarian doesn't like what we have now and was going to see if I can set it up myself on a VPS. And if I could figure it out for myself, I feel like I could probably set up a server for the school. The IT staff we have aren't the greatest. I've been using Linux for 3 years now. Am running
10:20 mt-dev two servers now. (one Homeserver running LAMP stack, Plex, Transmission, one VPS on Digital Ocean https://flexstart.me) I can set up web servers, do cron jobs, bash scripting, etc. I've set up many applications before that require databases, creating users, and modifying permissions (Wordpress, ownCloud, TinyTinyRSS, Transmission, Piwik, Piwigo, et
10:20 mt-dev c). Do you think I could install and maintain Evergreen by following the installation tutorial and documentation on the Evergreen site?
10:20 jboyer-isl I’m going to have to put some reminders in my phone though so I don’t forget and have my bags ejected from the room, heh.
10:22 kmlussier yboston / jboyer-isl: We will also make sure you have a way to get back to Boston on Friday. I could probably drive you on my way home.
10:22 yboston kmlussier: thanks
10:23 jboyer-isl kmlussier: that sounds much better than picking me up at the hotel at ~5 the next day. ;)
10:23 jboyer-isl kmlussier++
10:23 jboyer-isl yboston++
10:23 kmlussier yboston++
10:23 yboston kmlussier++
10:24 Dyrcona mt-dev: It really helps to have installed other software from source code on Linux beforehand, i.e. using configure, make, etc.
10:24 Dyrcona Just getting it installed is step 1.
10:27 mt-dev joined #evergreen
10:27 kmlussier mt-dev: Given your experience, you probably could install and maintain Evergreen if you follow the README, but one thing I would be concerned about is who will maintain it in a few years if the IT staff isn't strong.
10:27 yboston mt-dev: I agree with Dyrcona, after installation you need to spend time configuring EG.
10:28 yboston mt-dev: to give you a rough and slightly outdated idea of what configuration work that has to be done after isntalation, the EG community created a now somewhat outdated online manual of some of the configuration steps that have to be done after instalation http://flossmanuals.net/evergreen-in-action/
10:30 kmlussier Sorry, I meant to say you might be able to install and maintain, not probably. Basically,  EG requires a lot of ongoing care and maintenance.
10:33 Bmagic mt-dev++
10:34 Bmagic mt-dev: There are definatly concerns in the long run. I'm excited of the prospect you taking on such a big project though! Have you thought about getting the MARC out of the old system?
10:41 mt-dev joined #evergreen
10:43 Dyrcona Whee! Fun with parameters!
10:43 Bmagic parameters++
10:44 Bmagic spanish for "for meters"
10:44 Dyrcona Bmagic++
10:44 kmlussier heh
10:44 Bmagic :)
10:45 kmlussier I think that would be parametros :)
10:45 Dyrcona Sic vis itinere para meters.
10:49 Dyrcona Hmm. Should probably Sic vis iter... but anyway.
10:51 mt-dev joined #evergreen
11:03 csharp @eightball is "El Niño" Spanish for "The Niño"?
11:03 pinesol_green csharp: The outlook is hazy, please ask again later.
11:06 Dyrcona Hah.. Sounds about right.... :)
11:08 mt-dev Dyrcona: I have installed some software from source before
11:09 mt-dev What kind of configuration/maintenance does it require?
11:10 mt-dev Bmagic:   haven't got the database from the old system yet.
11:11 mt-dev Are there companies that set up/maintain Evergreen for schools?
11:11 mt-dev They're paying Destiny now, so
11:12 miker mt-dev: there are several (disclaimer: I with one of them, Equinox)
11:13 kmlussier mt-dev: There is a list available here http://wiki.evergreen-ils.org/dok​u.php?id=faqs:evergreen_companies
11:13 mt-dev Thanks kmlussier
11:15 mt-dev They wouldn't have to do anything on-premise, would they (Does it matter if they aren't anywhere near the school in Indiana?)
11:16 csharp mt-dev: some of the vendors on that list do off-site hosting, yes
11:19 mt-dev joined #evergreen
11:20 Christineb joined #evergreen
11:24 krvmga joined #evergreen
11:26 krvmga i find it can be frustrating trying to figure out why something is considered more relevant in an opac search
11:26 krvmga for instance, i'm doing a search for Title, All Books (regular print), in the library HCC for the word cooked
11:27 krvmga this record http://bark.cwmars.org/eg/opac/record/1614306​?query=cooked;qtype=title;fi%3Asearch_format=​book;locg=217;page=5;expand=marchtml#marchtml
11:27 dbwells joined #evergreen
11:27 krvmga appears before this one http://bark.cwmars.org/eg/opac/record/3085313​?query=cooked;qtype=title;fi%3Asearch_format=​book;locg=217;page=5;expand=marchtml#marchtml
11:28 krvmga but looks to be to be less relevant for the search term in a title field
11:28 krvmga it is a puzzlement to me
11:28 miker krvmga: which location is 217
11:28 krvmga Holyoke Community College - HCC
11:28 miker nm, I'll just change the URL ;)
11:28 miker ah, thanks
11:30 mt-dev joined #evergreen
11:45 miker krvmga: in a naive test that does not even come close to approximating all that evergreen does to generate a rank based on your config, they end up with the same relevance score.  I wonder if you need to adjust the class weights. (unfortunately, I'm about to turn into a pumpkin, though ... POOF)
11:46 krvmga miker: thanks for looking at this. is there anything you can share about the test you did? (however naive)
11:49 miker select ts_rank_cd(index_vector,to_tsquery('cooked')),* from test_table; ... test_ contains the id and output of to_tsvector() for the 245 from each record in a index_vector column. they both have 0.2
11:49 miker so, other factors (up to and including the numerical order of the ids) are deciding the final sort order. but, there are knobs ... :)
11:50 * miker runs away
11:50 kmlussier I'm confused. I'm looking at the xml output for those records, and, in both cases, the only time cooked (or cooking) appears in the title index is in the title proper index.
11:51 kmlussier I don't see how adjusting class weights could help with that problem.
11:51 * kmlussier doesn't expect to get an answer now since miker has turned into a pumpkin. :)
11:59 Bmagic We recently ran out of workers on our utility server. I've asked myself: can I setup another utility server and have all the same things running (hold targeter, clark, action trigger) simultaneously? How do you handle it?
12:00 jboyer-isl Bmagic: That will make you very sad.
12:00 Bmagic ok, that's what I thought
12:00 Bmagic hehe
12:01 jboyer-isl Unless you split them up so that machine 1 does all of the holds, and the other one does all of the A/T work, etc.
12:01 Bmagic so, I am considering putting some of the bigger worker consumers on another machine
12:01 jboyer-isl You can also increase the worker limit, unless you’re running out of RAM/SPU/ETC.
12:01 jboyer-isl SPU? CPU.
12:01 Bmagic What is up with the fine generator rolling back constantly!
12:02 jboyer-isl I think it tests and then discards some things. It’s been an age since I’ve looked at it.
12:02 Bmagic we have it set to  <max_children>45  <min_spare_children>1 <max_spare_children>5  on 4GB memory and 12 cpu
12:03 Bmagic At some point, I remember someone saying that once you get to a certain point, you just need to spin up another app server
12:06 jboyer-isl We have a 4 CPU, 16GB VM with most services set to a max of 75 children and it never causes us any trouble.
12:07 Bmagic interesting
12:07 Bmagic maybe I should increase the Memory and decrease the brick count
12:07 jboyer-isl (I know I could add more than 75 to some and take some away from others to make this better, but it has yet to be urgent.)
12:08 Dyrcona Bmagic: You set up another server to run drones, but only run the cron jobs on one of them.
12:08 Dyrcona Althought jboyer-isl basically said that in different words.
12:08 jboyer-isl Bmagic: yeah, you probably don’t have as many concurrent processes as you do mem-heavy processes (which then get backed up until they’re bumping into each other, etc.)
12:08 Bmagic There isn't a hypothetical limit to the number of children?
12:09 jboyer-isl INT_MAX.
12:09 jboyer-isl Theory, practice, etc.
12:09 Bmagic so, we run 8 app servers at 4GB each, with 45 children. Perhaps we could go back to 4 app servers, with 8GB and 90 children?
12:12 sandbergja joined #evergreen
12:17 jboyer-isl That may work. How many CPUs in that situaiton?
12:18 Bmagic 12 CPUs each which are hardly used
12:19 mt-dev joined #evergreen
12:19 bmills joined #evergreen
12:26 jboyer-isl Oh yeah, we don’t have more than 4 on anything here, but 12GB on every app server.
12:27 Dyrcona jboyer-isl | Bmagic: The O/S has a limit on max children and max processes.
12:27 Bmagic Dyrcona: where is that limit? Or how do I find out?
12:28 Dyrcona Off the top of my head, I don't remember because I use several variations of *NIX and its different on some of them.
12:29 Dyrcona I always have to look it up, so I recommend Google.
12:29 Dyrcona They're both likely to be in the tens of thousands by default on GNU/Linux.
12:29 Bmagic right now ps aux|grep OpenSRF|wc -l yeilds numbers like 120
12:30 Bmagic and that is with 45 children
12:30 Dyrcona I know OpenBSD sets a rather low number (2048) as a default for max processes.
12:34 Bmagic cat /proc/sys/kernel/pid_max  ?
12:34 jboyer-isl Bmagic: that 120 includes the listener for each service (not doing anything but listening for requests) and drones for every service you have turned on, whether they’re used or not. (depending on min_spare_children and some other things)
12:34 jboyer-isl Bmagic:  ulimit -a
12:34 mt-dev joined #evergreen
12:35 jboyer-isl max user processes is 128221 on one of my vms. (I only touch it on the DB servers)
12:35 Bmagic 7719 on a dev machine. What an odd number
12:36 mt-dev left #evergreen
12:36 Bmagic production machine reports 31412, going up to 90 children - I dont think im going to hit 31k processes. Considering that the total right now is less than 400 for ALL ps aux
12:46 Dyrcona GNU/Linux may try to calculate something reasonable given the available hardware.
12:47 Dyrcona On most *NIXen it is hard coded and can by modified with sysctl or a similar mechanism.
12:47 Dyrcona s/by/be/
12:47 * Dyrcona grumbles about cold fingers.
12:49 Dyrcona you also likely want something other than ulimit, as that applies to your user's totals.
12:52 csharp Bmagic: I would definitely up the RAM to at least 8 if you can
12:52 Dyrcona Oh don't mind that previous statement, that is the same value returned by sysconf CHILD_MAX or _POSIX_CHILD_MAX.
12:52 * Dyrcona agrees with csharp.
12:53 * Dyrcona listens to the sounds of demolition happening behind him.
12:55 Dyrcona Told you, I have to look this stuff up every time.
12:57 Dyrcona hmmm...
12:59 Dyrcona So, I am getting unmet dependencies from running OpenSRF/src/extras/Makefile.install ubuntu-trusty just now.
12:59 Dyrcona This is OpenSRF master, btw.
13:00 Dyrcona AFAIK, nothing changed on the OpenSRF side since this succeeded for me the last time, so I blame Ubuntu.
13:00 csharp what's it saying?
13:00 csharp @blame Ubuntu
13:00 pinesol_green csharp: Ubuntu 's bugfix broke csharp's feature!
13:03 pastebot "Dyrcona" at 64.57.241.14 pasted "unmet dependencies" (36 lines) at http://paste.evergreen-ils.org/18
13:04 Dyrcona Looks like apache2 and friends are broken, or my sources.lst is.
13:05 csharp Dyrcona: I'm cloning a stock trusty VM - I'll test momentarily
13:05 Dyrcona My sources.list looks good, no local mirrors.
13:06 Dyrcona If I try and install the missing packages by hand, I get a similar message about uuid-dev.
13:07 Dyrcona Maybe apt-get follow options (or whatever) changed recently?
13:08 * csharp waits for dist-upgrade to complete
13:09 Dyrcona Could be my VM is hosed, but I just made it.
13:10 Dyrcona Something is not right here.
13:11 csharp so far so good on my fresh trusty vm
13:11 Dyrcona libaprutil1-dev (Apache2) says it depends on libpq-dev (PostgreSQL), and that can't be right.
13:12 Dyrcona heh. I feel dumb.
13:12 Dyrcona apt-get update seems to have resolved it for me.
13:12 Dyrcona Don't usually have to do that.
13:12 csharp I wondered about that
13:13 csharp out-of-date package cache
13:13 Dyrcona Well, the strange dependency was the clue.
13:13 Dyrcona Probably corrupted.
13:13 csharp yep - I've seen that
13:13 jihpringle joined #evergreen
13:13 Dyrcona Makes me wonder... I should probably build a new VM.
13:14 Dyrcona Don't think I can trust this one.
13:14 jeff_ for trusty, libaprutil1-dev does depend on libpq-dev: http://packages.ubuntu.com/trusty/libaprutil1-dev
13:15 csharp we just build stock images of different RAM/storage configurations and have a bash wrapper around virt-clone to clone from there
13:16 Dyrcona jeff_: Ok, but that seems off.
13:16 Dyrcona I use vmbuilder to build fresh vms from "scratch."
13:16 csharp (assuming KVM/qemu as opposed to other virtualization possiblities)
13:16 csharp ah
13:16 Dyrcona yeah, it's kvm/qemu, also.
13:17 jboyer-isl Dyrcona: I’ve looked at vmbuilder recently, but it looks like it’s been set aside for years. Is everything ok with it and Trusty?
13:17 jboyer-isl (haven’t had time to test it yet, sadly)
13:17 csharp @eightball is it time for the Evergreen project to finally give in to the public pressure to support Linux From Scratch?
13:17 pinesol_green csharp: No.
13:17 Dyrcona jboyer-isl: I got it working after applying a patch. I believe it may also be used with OpenStack.
13:19 jboyer-isl Dyrcona: good to know. I’ll try to make the time to do some real research soon.
13:19 Dyrcona The patch may no longer be needed.
13:20 Dyrcona Ok. Now I smell the dust from the demolition in the office.
13:20 jboyer-isl I’ll start with the vmbuilder package in trusty and see if/where it falls over. If that doesn’t work, it’s going to get some serious side-eye from me.
13:22 Dyrcona jboyer-isl: Here's the patch that I mentioned: https://launchpadlibrarian.net/120169451/usr_share​_pyshared_VMBuilder_plugins_ubuntu_dapper.py.diff
13:23 csharp wow - Dapper
13:24 jboyer-isl Most of the vmbuilder conversations I’ve seen revolve around issues with Karmic, that’s why I was suspect of it’s continued viability. I am reminded that I have a trusty vm on this mac, I could figure this out in a few short minutes.
13:26 Dyrcona It is attached to bug 1037607
13:26 pinesol_green Launchpad bug 1037607 in vm-builder (Ubuntu) "vmbuilder completely fails on Quantal due to kernel pae detection failure" [High,Triaged] https://launchpad.net/bugs/1037607
13:26 jeff I use https://github.com/andsens/bootstrap-vz for building images. Probably doesn't help Ubuntu users.
13:28 mtate joined #evergreen
13:31 Dyrcona jboyer-isl: There's a work around to avoid the patch if you add linux-image-generic explicitly to the list of packages to install.
13:31 Dyrcona I tried that, too, and it works.
13:32 jboyer-isl Dyrcona: thanks, I’ll make some notes.
13:32 Dyrcona On an unrelated note, it will be fun when I upgrade my OpenBSD gateway to 5.8 and sudo becomes doas.
13:34 jboyer-isl Ugh, is that an OpenBSD specific utility, or is something changing further upstream?
13:35 Dyrcona It's OpenBSD specific, they've moved sudo from the base to ports.
13:35 * csharp just googled 'sudo vs. doas' and found http://www.tedunangst.com/flak/post/doas
13:36 plux_ joined #evergreen
13:36 jboyer-isl Heh, I do tend to lazysearch IRC quite a lot.
13:37 jboyer-isl But the natural language recognition is even better than Google
13:37 jboyer-isl 's!
13:37 mmorgan joined #evergreen
13:37 csharp @lazysearch doas
13:37 pinesol_green csharp: <shaggy>We're, like, doomed.</shaggy>
13:38 plux_ starting to put together a mac staff client for 2.9 …..just wondering if I should stay with xulrunner 14.0.1 or if I can pull down a newer version of xulrunner…any thoughts?
13:39 jboyer-isl plux_: I’d stick specfically with the version mentioned, some xulrunner features we absolutely require were removed not long after that.
13:39 Dyrcona plux_: You can try, but it won't work.
13:40 plux_ thanks….sticking with 14
13:41 plux_ the last few I did had issues with the infamous spine label printers but that’s not probably associated
13:42 Dyrcona I've always had some funkiness with XulRunner on Mac OS X, but I don't use Mac OS X any more.
13:43 Dyrcona csharp: BTW, a fresh VM is working fine. I occasionally get a corrupted VM, typically they refuse to start.
13:47 mllewellyn joined #evergreen
13:56 csharp Dyrcona: awesome
14:03 Dyrcona I guess we're not doing monthly maintenance releases today....
14:04 Dyrcona AFAIK only 1 bug was fixed in 2.9, and it is not a biggy.
14:06 * mmorgan is looking at credit card processing and was sure there was a variable in one of the tt2 files where a minimum amount could be set below which the pay option wouldn't appear.
14:06 mmorgan Did I dream this?
14:07 kmlussier mmorgan: We did see something that looked like it would do so. Give me a few minutes and I'll see if I can find it.
14:08 kmlussier miker: I don't know if you've changed back into a coach again, but I just realized what you meant when you suggested adjusting the class weights. A reingest isn't required after you adjust those weights, is it?
14:08 mmorgan Ok, thanks. I've been scouring the tt2 files and can't put my finger on it.
14:09 vlewis joined #evergreen
14:10 mmorgan Maybe the bit is in main_payment_form.tt2?  [% IF ctx.fines.balance_owed <= 0 %]
14:10 mmorgan I thought I remembered something different, though.
14:11 vlewis_ joined #evergreen
14:12 * mmorgan had notes on a flash drive that, apparently, is toast :-(
14:15 kmlussier mmorgan: Yes, that's what I remember. Up that number from 0 to 5 and see what happens.
14:17 maryj_ joined #evergreen
14:17 mmorgan kmlussier: Ok, will give that a try. Thanks!
14:20 csharp mmorgan: the apparent lack of a minimum payment is one of our concerns too - I'll be interested in what you discover
14:24 maryj joined #evergreen
14:24 mmorgan kmlussier: Ok, changing 0 to something higher, does make the payment not possible, but the message "You either have no fines to pay or you have selected fines whose total is non-positive.  We cannot process non-positive amounts" needs a little tweaking (for our case, anyway).
14:26 kmlussier Yeah, that should be tweakable
14:26 kmlussier Ideally, it would be a setting. But it sounds like this will work in the meantime.
14:27 mmorgan csharp: I set [% IF ctx.fines.balance_owed <= 1 %], which is the limit we're planning to use. On the Account Summary, I deselected everything but $.42. When I clicked Pay Selected Fines, I got that message instead of the form to fill in the cc info.
14:27 mmorgan Yes, I think changing it will be easier than coming up with the exact words to put there ;-)
14:28 * kmlussier checks something
14:29 mmorgan Interesting that the message says "Either you have no fines to pay..." since I don't think a patron can even get to that page if nothing is owed.
14:30 kmlussier Well, sometimes people arrive at strange places by means we never anticipated they would use.
14:31 kmlussier mmorgan: With that in mind, you might want to keep the code there as is, and then add a new one to account for fines from .01 to .99 with different language saying they haven't reached the fine minimum.
14:32 kmlussier Actually, scratch that idea. That wasn't one of my good ones.
14:32 dMiller_ joined #evergreen
14:37 mmorgan We'll probably opt for a message that says something like "The minimum amount for credit card payments is $1.00" but I agree that an option would be best.
14:40 jeff "Thank you for your successful authorization of $0.15. Since it would cost the library more than $0.15 to accept your payment, we are forgiving the $0.15 and will not capture the transaction for settlement. Have a nice day!"
14:41 vlewis joined #evergreen
14:41 miker kmlussier: a class A exact match (no stemming) on cooked (he says 3 hours later) (also, still a pumpkin...)
14:41 mmorgan "Congratulations! You have been randomly selected to have your fine of less than $1.00 forgiven!"
14:42 kmlussier miker: Yes, so it took me some time, but I got there. krvmga increased the class A weight to 5, but didn't see in a change in the order of results.
14:43 kmlussier I wasn't sure if something else needed to be done to make the change go into effect. He did your little track of adding a -ajlfjlsa type word so that it wouldn't rely on cached results.
14:44 kmlussier Sigh...search. Why do you have to be so difficult?
14:45 vlewis_ joined #evergreen
14:46 * miker catches up on scrollback ... kmlussier: no reingest required
14:47 kmlussier miker: You're not doing a very good job of staying in pumpkin mode.
14:47 * miker thinks "round!"
14:52 kmlussier krvmga: We talked about this a bit in our meeting earlier this week, but the more I think about it, the more I think a separate title and keyword index for just the 245a would be a good thing for relevance.
14:53 kmlussier My one concern is that you already have a lot of indexes. I don't know how much it will affect performance if you add even lots more rows to your metabib tables.
14:53 krvmga we have got SO many indexes right now
14:54 kmlussier Yes. As does NOBLE.
14:54 krvmga just for community interest, anyone ever want to see someone with lots of indexes, i'll be happy to share.
14:55 krvmga yes, but we have two more letters in our acronym than they do. so there.
14:55 krvmga :)
14:55 krvmga i was always bad at math.
14:56 kmlussier krvmga: Ha ha. You could say you have two more characters.
14:56 Dyrcona krvmga: Oddly enough, so are computers. :)
14:56 krvmga lol
14:57 * mmorgan scrambles for a few more letters to add to acronym. ;-)
14:57 mmorgan SNOWBLE would have been quite appropriate last year.
15:01 kmlussier krvmga: OK, so let's look at it this way, It's possible that a 245a index for the keyword/title class may do even more for relevance than some of the other indexes you've created to help with relevance. Maybe we can make some tradeoffs and get rid of some of the others.
15:01 kmlussier Of course, the ones that are less important for relevance are probably the same ones that don't generate a ton of metabib entries.
15:02 ericar_ joined #evergreen
15:07 krvmga mmorgan: rofl
15:09 csharp @marc 590
15:09 pinesol_green csharp: unknown tag 590
15:13 mmorgan joined #evergreen
15:13 jboyer-isl csharp: local notes! we just blew them all away here.
15:14 jihpringle joined #evergreen
15:14 jboyer-isl (not that “local notes!” helps, I think there’s a lot of those.)
15:15 Dyrcona We use that to differentiate "on order" or "brief" records from "regular" MARC records.
15:15 mmorgan joined #evergreen
15:16 Dyrcona My local version of the Backstage software does not export records with a date or an o/o, or similar note in any 590 subfield.
15:16 Dyrcona So, it has uses....
15:17 jboyer-isl Dyrcona: sure. We mostly had a ton of migration 590s. (and forgotten on order notes, I suspect.)
15:20 csharp jboyer-isl: yeah - they're actually desired in my current project (Georgia Governor's Mansion)
15:24 Dyrcona csharp: You can pretty much put what you want into it.
15:24 Dyrcona You might have to do some work to make it display in the catalog if desired.
15:32 mmorgan joined #evergreen
15:43 csharp Dyrcona: yeah - I just had to add it to /openils/var/templates/opac​/parts/record/contents.tt2 and it showed up fine ;-)
15:43 csharp (for the logs ^^)
15:44 Dyrcona csharp: Cool.
15:57 jlitrell joined #evergreen
16:00 mmorgan joined #evergreen
16:06 dMiller_ joined #evergreen
16:16 bmills joined #evergreen
16:48 mmorgan joined #evergreen
16:53 kmlussier I have a question about the estimated search results count on the search results page. My recollection is that the result counts start to get inaccurate once you hit the 1,000 records threshold for records retrieved.
16:53 kmlussier Because 1,000 that's the default superpage size in Evergreen. And that count includes both opac-visible and non-visible records that are retrieved in the search.
16:54 kmlussier However, in this search, we are retrieving fewer than 100 records, and getting inaccurate result counts: http://bark.cwmars.org/eg/opac/results?bool=and​&amp;qtype=keyword&amp;contains=contains&amp;qu​ery=ireland&amp;bool=and&amp;qtype=title&amp;co​ntains=contains&amp;query=&amp;bool=and&amp;qty​pe=author&amp;contains=contains&amp;query=&amp;​_adv=1&amp;fi%3Alocations=1602&amp;fi%3Alocatio​ns=2640&amp;fi%3Alocations=2639&amp;locg=247&am​p;pubdate=is&amp;date1=&amp;date2=&amp;sort=
16:54 kmlussier I'm wondering if it's related to the copy location limiters. Without those limiters, I would get more than 1000 results when searching that particular OU.
16:55 kmlussier Is that why I don't get an accurate records even though this particular search retrieves just a handful of records? Because the non-copy-location-limited search hits that 1000-record threshold?
16:59 mmorgan joined #evergreen
17:02 * kmlussier notes that the estimate seems to be accurate when limiting by format instead of copy location.
17:02 dMiller_ joined #evergreen
17:06 mtj_ joined #evergreen
17:22 mmorgan joined #evergreen
17:22 mmorgan left #evergreen
17:33 mtj_ joined #evergreen
17:41 dbwells joined #evergreen
17:49 jihpringle joined #evergreen
18:57 Stompro joined #evergreen
19:12 miker kmlussier: assuming stock settings, the first superpage had exactly 10 visible records after trimming for that filter. that filter happens after the core search select, like site and available filters do
19:13 miker so the second page of results triggered (at least) one more superpage search
19:44 dMiller_ joined #evergreen
19:59 miker I'll just leave this right here: http://blog.pgaddict.com/posts/performan​ce-since-postgresql-7-4-to-9-4-fulltext
20:00 miker tl;dr: pg 9.4 = faster search, esp on worst case searches ("the", "history", your state name, etc)

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