| Time |
Nick |
Message |
| 01:55 |
|
beanjammin joined #evergreen |
| 03:46 |
|
beanjammin joined #evergreen |
| 04:32 |
|
dbwells__ joined #evergreen |
| 06:38 |
|
Dyrcona joined #evergreen |
| 07:07 |
|
rjackson_isl joined #evergreen |
| 07:32 |
|
bdljohn joined #evergreen |
| 07:52 |
|
bos20k joined #evergreen |
| 08:22 |
* Dyrcona |
hates it when behavior changes inexplicably and when two different systems with the same software and configuration behave differently. Computing sucks. |
| 08:23 |
Dyrcona |
Well, I have to take the same configuration thing back. I thought they were the same but they're not..... |
| 08:24 |
Dyrcona |
Here's what's bugging me. I changed our open-ils.trigger configuration to use 3 collectors and 3 reactors during our upgrade. |
| 08:25 |
|
collum joined #evergreen |
| 08:25 |
Dyrcona |
All of the action triggers seems to be working fine, except for our trigger to print lost notices. It is not grouping notices by usr anymore, instead we're getting 1 notice for each user for each item they have lost/long overdue. |
| 08:25 |
Dyrcona |
The grouped action triggers are working as expected. |
| 08:28 |
Dyrcona |
So, something is wrong in the collection step? I don't have any idea at this point. |
| 08:37 |
|
mmorgan joined #evergreen |
| 09:02 |
Dyrcona |
Was the 3.2.0 release announced? |
| 09:03 |
|
stephengwills joined #evergreen |
| 09:03 |
Dyrcona |
Indeed, it was. |
| 09:05 |
* Dyrcona |
waits on lost a/t to finish on test server... |
| 09:05 |
|
remingtron joined #evergreen |
| 09:09 |
|
sallyf joined #evergreen |
| 09:11 |
Dyrcona |
Yeahp. It's definitely the parallel settings for open-ils.trigger causing the lost print notices not group properly. The question is now why does this one a/t have the problem and our others don't? |
| 09:36 |
|
kmlussier joined #evergreen |
| 09:37 |
|
yboston joined #evergreen |
| 09:38 |
Dyrcona |
@monologue |
| 09:38 |
pinesol |
Dyrcona: Your current monologue is at least 11 lines long. |
| 09:38 |
gmcharlt |
good morning, Dyrcona! |
| 09:38 |
Dyrcona |
Good morning, gmcharlt! |
| 09:39 |
kmlussier |
jeff: It seems odd to me that the keyword blob would be a display field. I may be missing something, but I'm thinking that may have been unintentional. |
| 09:52 |
|
mmorgan joined #evergreen |
| 09:57 |
Dyrcona |
So, I have been informed that our lost print notice was further customized by someone who no longer works for EOLI (IIRC). But, I don't know how this was done. Must be in the database somewhere because we would have lost any Perl changes when I did our first upgrade. |
| 09:58 |
Dyrcona |
It wasn't a custom filter, because this is the on a/t that we're not using a custom filter on. |
| 10:04 |
|
nfBurton joined #evergreen |
| 10:04 |
Bmagic |
Any theories on whether or not an active authority load will conflict with an Evergreen database upgrade script(s). I've got eg_staged_bib_overlay overlay_auths_stage1 that I would rather not have to restart |
| 10:07 |
Dyrcona |
Bmagic: The answer is it depends on what the upgrade script is doing. |
| 10:07 |
bshum |
Bmagic: Well that depends on ... |
| 10:07 |
bshum |
Yeah, what Dyrcona said :) |
| 10:07 |
Bmagic |
yeah, I've been studying the scripts. There are references to the authority schema sometimes |
| 10:07 |
Bmagic |
2.12.6-3.0.0 for example |
| 10:07 |
Dyrcona |
Then, it will probably interfere or bock until the authorities are done. |
| 10:08 |
Dyrcona |
s/bock/block/ |
| 10:08 |
Dyrcona |
Doesn't that also require some ingest after the upgrade script? |
| 10:09 |
Bmagic |
The nature of the eg_staged_bib_overlay script seems to be working in groups, in theory once it's performs it's commit and before it can start another transaction, the upgrade script would cause it to wait or visa versa? |
| 10:10 |
Dyrcona |
In theory, but again that depends, and I wouldn't count on anything. Chances are one or the other will get kicked out of the db. |
| 10:11 |
Bmagic |
I think Im just going to bite it and restart it after the upgrade :) thanks for helping me decide! |
| 10:11 |
Dyrcona |
Bmagic: You do shut Evergreen services down when doing upgrades don't you? |
| 10:11 |
Bmagic |
yep |
| 10:12 |
Dyrcona |
OK, so the same principle applies, you really don't want anything else going on when running an upgrade script. |
| 10:12 |
Bmagic |
I was just wishing |
| 10:25 |
|
beanjammin joined #evergreen |
| 10:26 |
* Dyrcona |
was mistaken about the person not working for EOLI any more. |
| 10:26 |
* Dyrcona |
should just shut up for a bit. |
| 10:32 |
csharp |
Dyrcona: we haven't seen problems with grouping notices with parallel collect/react |
| 10:33 |
csharp |
there was a bug sometime in the last 1.5 years where the group field could be null but the perl assumed it wasn't and would choke, but that's not what your issue sounds like |
| 10:34 |
csharp |
and that should be fixed upstream (not just in PINES) |
| 10:45 |
Dyrcona |
Right, and that's not my problem. There's something about our lost-print trigger that doesn't like being run parallel. |
| 10:46 |
Dyrcona |
I'm waiting on another test run to finish... |
| 11:05 |
|
jvwoolf joined #evergreen |
| 11:09 |
|
Christineb joined #evergreen |
| 11:20 |
|
khuckins joined #evergreen |
| 11:21 |
jeff |
kmlussier: the blob is a display field in the sense that it has display_field = true, but isn't mapped in config.display_field_map. Of course, it isn't the only row in config.metabib_field with that characteristic. Didn't finish digging last night. :-) |
| 11:22 |
jeff |
there are 25 cmf rows with display_field = true that don't have a mapping in cdfm, and that's probably expected. |
| 11:33 |
|
mmorgan joined #evergreen |
| 11:34 |
mmorgan |
windows-- |
| 11:35 |
Bmagic |
microsoft-- |
| 11:35 |
Bmagic |
windows_file_explorer++ |
| 11:35 |
Bmagic |
windows_task_manager++ |
| 11:36 |
* mmorgan |
wonders what's so great about file explorer and task manager |
| 11:36 |
aabbee |
"windows 10 october 2018 update no longer deletes your data" https://arstechnica.com/gadgets/2018/10/microsoft-fixes-october-update-file-deleting-bug-resumes-insider-testing/, so they got that going for them, which is nice. |
| 11:39 |
kmlussier |
@karma microsoft |
| 11:39 |
pinesol |
kmlussier: Karma for "microsoft" has been increased 0 times and decreased 1 time for a total karma of -1. |
| 11:39 |
kmlussier |
microsoft-- |
| 11:40 |
mmorgan |
microsoft-- |
| 11:40 |
mmorgan |
@karma windows |
| 11:40 |
pinesol |
mmorgan: Karma for "windows" has been increased 0 times and decreased 5 times for a total karma of -5. |
| 11:43 |
Bmagic |
mmorgan: Windows File Explorer beats every single UI file explorer that I have ever used. Konquerer, Finder, Dolphin, etc pale in comparison. Same with Task manager with advanced enabled. |
| 11:44 |
bshum |
I do enjoy that the October update brought "dark mode" file explorer. I lucked out and did not have my data destroyed. As far as I know.... |
| 11:44 |
Dyrcona |
Konqueror was originally a web browser, just sayin'. |
| 11:45 |
* bshum |
wants more "dark mode" everything |
| 11:45 |
csharp |
new Evergreen DARK MODE |
| 11:45 |
Dyrcona |
I use Nautilus, 'cause I like being able to hit Ctl-l, and then type in something like sftp://user host/path and have it just mount. |
| 11:46 |
bshum |
csharp: YES! I want to go to there |
| 11:46 |
* mmorgan |
got the October update last night, first sad face blue screen this morning :-( |
| 11:46 |
bshum |
Hackaway project idea |
| 11:49 |
Bmagic |
I've given Linux another look recently for my workstation. I would make the switch if I could get a file explorer that was worth a darn. Moving/copying files across network shares, (un)mounting usb/nfs/smb, file/folder recursive permissions. No lag, accurate copy progress with bandwidth feedback... |
| 11:50 |
berick |
csharp: bshum: https://bootswatch.com/darkly/ -- dark bootstrap theme |
| 11:51 |
Dyrcona |
Bmagic: Suit yourself. I'm not out to make converts. |
| 11:52 |
Bmagic |
:) - I really really want to use Linux for desktop UI. I'm sure it'll get there. I give it a real shot every couple of years to see what's new |
| 11:52 |
Dyrcona |
Since I work with Linux servers all the time having Linux on the desktop makes things easier for me. Not to mention that I have more accumulated work hours on Unix-like operating systems than on Windows. |
| 11:52 |
Bmagic |
Right now, I am favoring fedora with KDE (if someone made me switch today) |
| 11:53 |
bshum |
berick: Oooo pretty :D |
| 11:53 |
* Dyrcona |
used to work on KDE and wouldn't recommend it. They broke everything in KDE 4 and decided to release it half-finished. |
| 11:53 |
Bmagic |
I've heard that too but in testing, everything seems to work for me |
| 11:53 |
Dyrcona |
Well, they've had time to finish more stuff since then. :) |
| 11:54 |
Bmagic |
Starcraft 2 doesn't work so well in wine on my GPU on Gnome or KDE... so there's that |
| 11:54 |
Dyrcona |
I left about the time 2.0 was done because my daughter was born and no more time. |
| 11:54 |
Dyrcona |
The original Starcraft is the reason that WINE exists, and even if that's not true, it's the myth that I cling to. :) |
| 11:55 |
Bmagic |
wine is amazing and it blows my mind that we can execute windows binaries over the Linux kernel |
| 11:56 |
Bmagic |
but it's really hard to beat running it on Windows :) |
| 11:57 |
* Dyrcona |
doesn't play games, so doesn't need Windows. |
| 12:01 |
Bmagic |
Dyrcona: It's not too late to (re)start :) |
| 12:02 |
|
jihpringle joined #evergreen |
| 12:02 |
aabbee |
gaming on linux has gotten better. engines like unity and source are cross-platform. team fortress 2 and kerbal space program come to mind as fun examples. and, of course, tuxracer. |
| 12:03 |
Bmagic |
Oh man, tuxracer! I played that on my slackware box in 1998 |
| 12:03 |
aabbee |
meanwhile, in windowsland, why does farmville2 keep coming back?! |
| 12:04 |
aabbee |
windows-- |
| 12:04 |
Bmagic |
haha |
| 12:05 |
csharp |
extremetuxracer FTW |
| 12:06 |
csharp |
berick: nice ;-) |
| 12:12 |
Dyrcona |
So, I believe the changes/customization that I alluded to earlier are in the template code for our lost print action trigger, and after looking at it, I have a couple of questions. |
| 12:13 |
|
beanjammin joined #evergreen |
| 12:13 |
Dyrcona |
If the action trigger event definition is grouped by usr, is there any reason to sort on usr in the template itself? |
| 12:15 |
|
jvwoolf1 joined #evergreen |
| 12:32 |
mmorgan |
Dyrcona: I checked our long overdue print notice trigger and it does include a sort on usr. Not sure if it's necessary, bit I confirmed that the print notices do include multiple items on a single notice for a given usr. |
| 12:32 |
mmorgan |
s/bit/but |
| 12:36 |
Dyrcona |
mmorgan: Thanks. All of our print notices are sort on usr. |
| 12:37 |
Dyrcona |
Our lost print notice has an extra loop to sort by target_copy.call_number.owning_lib, and before I started running these in parallel it worked. |
| 12:38 |
Dyrcona |
The other od print notices work, so I assume the failure is related to this second sort, but I'm still waiting on test results. |
| 12:56 |
* Dyrcona |
thinks we're going back to the old settings... |
| 13:22 |
|
jvwoolf joined #evergreen |
| 13:58 |
jeff |
Just a few short days after the LoC MARC site outage, and I'm getting bibs from OCLC with no 245. |
| 14:09 |
|
yboston joined #evergreen |
| 14:33 |
|
nfburton joined #evergreen |
| 14:36 |
nfburton |
Does anyone have a SQL script they run to find BIBS that don't have any copies? |
| 14:36 |
nfburton |
Also, is the only link between bib and copy the call number? |
| 14:38 |
rhamby |
nfburton: yeah, the link is through the call number/volume, are you looking for someting with no current copies or no copies ever? |
| 14:39 |
nfburton |
Currently. Ever would be good to have too |
| 14:40 |
nfburton |
We are in the process of scrubbing our catalogue and noticed a bunch of empty records :( |
| 14:41 |
Dyrcona |
nfburton: You know you can't actually delete the bibs, right? They only get flagged deleted. |
| 14:41 |
rhamby |
there are several ways to do it, I'd caution you to exclude recent records that catalogers might still be working on |
| 14:43 |
rhamby |
when doing something like that I'll usually do left joins to subqueries on the asset.call_number and asset.copy looking only for undeletd copies and then join on null, it's a lot faster than a NOT IN query for any decent size data set |
| 14:45 |
nfburton |
Oh I know about the flagging. That's all I've been doing is flagging them. They're still kicking around in there. I've done a full E-Item cleanup to prepare for API usage. |
| 14:46 |
csharp |
nfburton: this is what we run on the master db in a nightly cron: https://pastebin.com/9hViAc2N |
| 14:46 |
csharp |
not sure where we got that (JBoyer? jeffdavis, maybe?) |
| 14:46 |
nfburton |
I have done a query LEFT joining copies and calls. Just didn't know if there was a better way. |
| 14:46 |
csharp |
it replaced one that I had hand-coded |
| 14:48 |
|
kmlussier joined #evergreen |
| 14:48 |
nfburton |
OOOOH I like that. Although I have to be careful because the electronic resources we have in the catalogue also have no copies/call numbers |
| 14:49 |
jeff |
nfburton: if you don't need any data from the LEFT JOINed fields, an EXISTS or NOT EXISTS clause can be speedy. |
| 14:49 |
Dyrcona |
CW MARS runs this one weekly: https://pastebin.com/isCN8c4Q |
| 14:50 |
nfburton |
Thanks! |
| 14:50 |
nfburton |
Dyrcona++ |
| 14:50 |
nfburton |
csharp++ |
| 14:51 |
nfburton |
Those can be helpful in my plight to clean this mess up lol |
| 14:51 |
jeff |
oh fun. that paste site's ads are starting to expand to obscure the "raw" link. |
| 14:52 |
Dyrcona |
jeff: Ad blockers or go pro. :) |
| 14:52 |
Dyrcona |
or both. :) |
| 14:52 |
nfburton |
I was going to say...."What ads?" :P |
| 14:53 |
csharp |
my ad blockers prevent my seeing that |
| 15:09 |
|
mmorgan1 joined #evergreen |
| 16:17 |
|
yboston joined #evergreen |
| 16:20 |
kmlussier |
@dessert |
| 16:20 |
* pinesol |
grabs some candy corn for kmlussier |
| 16:20 |
kmlussier |
Timely! |
| 16:36 |
csharp |
@dessert |
| 16:36 |
* pinesol |
grabs some coffee frappes for csharp |
| 16:42 |
bshum |
@dessert |
| 16:42 |
* pinesol |
grabs some Boston Cream doughnuts for bshum |
| 16:47 |
|
mmorgan joined #evergreen |
| 17:08 |
|
mmorgan left #evergreen |
| 17:20 |
jeffdavis |
jeff: a new library here wants to do some very basic stuff with NCIP v1 (Lookup User and Lookup Item). I'm leaning towards iNCIPit as a lightweight option, are there any big gotchas I should look out for? |
| 17:37 |
|
jvwoolf left #evergreen |
| 18:47 |
|
sandbergja joined #evergreen |
| 23:22 |
|
sandbergja joined #evergreen |
