Time |
Nick |
Message |
02:40 |
|
eglogbot joined #evergreen |
02:40 |
|
Topic for #evergreen is now Welcome to #evergreen (https://evergreen-ils.org). This channel is publicly logged. Logs for today: http://irc.evergreen-ils.org/evergreen/today |
06:25 |
JBoyer |
jeffdavis, sorry! I missed that when I posted my gist. You'll probably want to change it to ss:mc-ctx so it uses the StorageService that has buildMap=1 (I think that has additional "stuff" for Sessions) Though I should probably re-read those several pages of the wiki to see if it even matters to have one with the map and one without, especially if they both have the same prefix. |
07:17 |
|
rfrasur joined #evergreen |
07:33 |
|
BDorsey joined #evergreen |
08:21 |
|
tlittle joined #evergreen |
08:22 |
|
kworstell-isl joined #evergreen |
08:41 |
|
mmorgan joined #evergreen |
08:46 |
|
mantis1 joined #evergreen |
09:09 |
jeffdavis |
Ah ok, thanks, I couldn't figure out when to use mc vs mc-ctx |
09:12 |
|
smayo joined #evergreen |
09:44 |
|
terranm joined #evergreen |
09:45 |
|
Dyrcona joined #evergreen |
10:11 |
|
BDorsey joined #evergreen |
10:25 |
|
kworstell_isl joined #evergreen |
10:27 |
Dyrcona |
"...that'll be gone soon enough." Made me chuckle. |
10:28 |
berick |
heh |
10:28 |
berick |
progress is being made |
10:31 |
Dyrcona |
Yeah, albeit slowly. |
10:41 |
Dyrcona |
ubuntu-- |
10:42 |
Dyrcona |
So, I've switched to Arc on my main personal laptop, and Ubuntu feels like a prison in more ways than one. |
10:42 |
Dyrcona |
s/Arc/Arch/ |
10:42 |
Dyrcona |
Also, busted packages lately on Ubuntu 20.04, but this is off topic. |
10:43 |
berick |
Arch is fun. but also, xubuntu++ |
10:44 |
Dyrcona |
Arch will test/improve your actual Linux chops. I've come to appreciate systemd. I still don't like, but I don't dislike it as much as I used to. :) |
10:58 |
Dyrcona |
I'm trying to install vanilla Gnome on Ubuntu 20.04, but package dependencies are out of sync. |
11:02 |
Dyrcona |
I like how apt says "you have held broken packages." But I have no held packages. Neither of the ways to check (aptmark, dpkg) say I have held packages. |
11:03 |
berick |
Dyrcona: maybe start w/ a server install so you don't have all that existing Gnome stuff? |
11:04 |
Dyrcona |
Well, I'm trying to update my work laptop to regular gnome. I don't want to reinstall the O/S. |
11:05 |
Dyrcona |
One issue is adwaita-icon-theme-full wants adwaita-icon-theme 3.36.0 but I've got adwaita-icon-theme 3.36.1. The vanilla Gnome packages are obviously ignored by the maintainers. |
11:06 |
Dyrcona |
The other package issues are similar. Some package wants an older version than what I've got installed. I don't feel like fixing all of that right now. |
11:09 |
Dyrcona |
Back to a more Evergreen-related topic, and related to my first comment this morning: I just want to clarify something about websockets and OpenSRF/Evergreen. If I were to write some client code to use remotely, I could do everything with websockets, correct? I wouldn't need to implement client code for the HTTP translator for anything, would I? |
11:13 |
|
mdriscoll joined #evergreen |
11:16 |
berick |
Dyrcona: you can do everything w/ websockets, but the data it passes around are low-level opensrf messages, similar to the translator |
11:16 |
berick |
iow, not like the json gateway, which just gives you an array of responses |
11:18 |
Dyrcona |
berick: Thanks! I might as well implement both if the message parsing is going to be the same. (I'm not saying that I'll do it soon.) |
11:18 |
scottangel_ |
I've got a new dev question. Where would I find information about retrieving data from the database from within javascript? This is what I have but I'm getting null. $scope.hide_strict_barcode_checkbox = setting['ui.circ.hide_patron_strict_barcode']; Not really sure where to look on the wiki. I tried doing some searches but the results were irrelevant. |
11:19 |
scottangel_ |
Sorry, that's all within the egCore.org.settings('ui.circ.hide_patron_strict_barcode').then(function(setting) { // code here } |
11:23 |
Dyrcona |
Well there's a typo on this line: https://git.evergreen-ils.org/?p=working/Evergreen.git;a=blob;f=Open-ILS/src/eg2/src/app/staff/circ/checkin/checkin.component.ts;h=f95481f00fbba91313dca69286cebf24f4312b84#l134 |
11:23 |
berick |
scottangel_: egCore.org.settings(...) is pulling data from the database. beware org setting values are cached, though |
11:24 |
berick |
so if the value is changed elsewhere, it may require a login/out to clear the cache |
11:24 |
Dyrcona |
Oh, never mind me. I missed the line where "sets" is actually set. |
11:25 |
scottangel_ |
crap. that's right. I forgot about the login/logout. lol I bet that works now. |
11:27 |
|
Christineb joined #evergreen |
11:37 |
StomproJ |
bmagic, did the setup for bug #1862834, the CNAMES for the hostname with eg or staff prefix get into any of the Mobius servers? |
11:37 |
pinesol |
Launchpad bug 1862834 in Evergreen 3.11 "regex based url building that can match hostnames" [Medium,Confirmed] https://launchpad.net/bugs/1862834 |
11:38 |
Dyrcona |
StomproJ: If not, you should be able to fake it with host file entries, even on Windows. |
11:38 |
Dyrcona |
Or, are you asking about them being added to the Apache/nginx configuration? |
11:40 |
Dyrcona |
I should stop jumping to conclusions. |
11:41 |
StomproJ |
I'm not sure what the plan was... I was just going to check if bmagic had setup something before I work on it myself. |
11:42 |
|
BrianK joined #evergreen |
12:02 |
|
jihpringle joined #evergreen |
12:05 |
|
rfrasur joined #evergreen |
12:09 |
Bmagic_ |
StomproJ: I forgot about that! I'll get on it |
12:10 |
|
Bmagic joined #evergreen |
12:10 |
StomproJ |
Bmagic, it is ok, I have a test system setup now to test it. |
12:10 |
Bmagic |
don't need me to do anything? |
12:10 |
StomproJ |
Nope, but thank you. |
12:11 |
Bmagic |
sorry! I know I said I would do that |
12:12 |
berick |
fwiw, Dyrcona's suggestion of just adding a hosts entry should work too |
12:14 |
Bmagic |
berick: but the certificates are a problem? |
12:14 |
Dyrcona |
Not if you use a wildcard certificate. |
12:14 |
Bmagic |
these are letsencrypt FQDN certs |
12:14 |
Dyrcona |
or you can click through the warnings. |
12:15 |
Dyrcona |
Well, clicking through the warnings it is. :) |
12:15 |
Dyrcona |
cnames won't help in that case. |
12:15 |
Bmagic |
I've noticed that I can no longer do that with Firefox/Chrome |
12:15 |
berick |
Bmagic: type "thisisunsafe" on the cert error page |
12:16 |
berick |
in Chrome |
12:16 |
Bmagic |
if the URL that you are on is a name (not an IP) - the browser won't let you click through the ssl warning. Unless you go into the deep settings about::config |
12:16 |
Bmagic |
berick: there's a magic word you can just "type" whilst on the error page? |
12:16 |
berick |
yup |
12:16 |
Bmagic |
lol |
12:17 |
Bmagic |
berick++ |
12:19 |
Dyrcona |
Y'know its the combination of snaps and apparmor that made me flee Ubuntu on my personal laptop. I tried to open some files in Firefox, and it said, sorry I can't do that. I did some digging and found out it was mostly apparmor, but disabling apparmor didn't fix it because snaps are special. Changing the apparmor config only lasts until the next snap refresh. So I kicked off the training wheels and switch to a distro with |
12:19 |
Dyrcona |
wer. ;) |
12:19 |
Dyrcona |
It's related because "developers know better than you" what's safe and what's not safe. :) |
12:20 |
Bmagic |
what distro did you move to? |
12:20 |
Dyrcona |
Arch |
12:20 |
Bmagic |
A popular choice |
12:20 |
Bmagic |
I think you're in good hands over there |
12:20 |
Dyrcona |
It's almost a DIY, but not so much as LFS. |
12:20 |
StomproJ |
Is there a config switch to go back to using AngularJS interfaces? |
12:20 |
Dyrcona |
Yeah, last I checked it's as popular as Ubuntu and Debian. |
12:20 |
Bmagic |
I played with this flavor for a little while https://xerolinux.xyz/ |
12:20 |
berick |
Dyrcona: not jumping straight to *BSD? |
12:21 |
Dyrcona |
berick: I used FreeBSD on the desktop for over a decade. I find its more painful on a laptop. |
12:21 |
berick |
yeah... |
12:22 |
Bmagic |
I noticed all of the laptops that were running Linux, couldn't use the projector at the conference. |
12:22 |
Dyrcona |
I still run FreeBSD on my peronsal mail/web/file server and OpenBSD on my home router. |
12:22 |
Dyrcona |
Mine had almost zero trouble. |
12:22 |
berick |
another one i've been meaning to try https://voidlinux.org/ |
12:23 |
berick |
yeah i don't have issues w/ projectors generally |
12:23 |
Dyrcona |
StomproJ: I don't think there is such a switch in the config. |
12:23 |
StomproJ |
Thanks, I wasn't sure if that was a thing or not. |
12:24 |
Dyrcona |
StomproJ: You can mangle the templates, and there's a way to switch out the Angular staff catalog, IIRC. |
12:24 |
Bmagic |
voidlinux! Awesome. reading their splash page, I'd imagine I would be in for a gentoo expierence |
12:25 |
Dyrcona |
Bmagic: I also have an ARM laptop with Manjaro KDE preinstalled. Manjaro is an Arch derivative. I'm considering migrating it to Arm for ARM, but that's a bigger project because I have to make an image and flash the machine. |
12:26 |
Bmagic |
https://upload.wikimedia.org/wikipedia/commons/a/ad/2023_Linux_Distributions_Timeline.svg |
12:26 |
Dyrcona |
Heh, "Arm for ARM" should be Arch for ARM. |
12:26 |
Bmagic |
voidlinux is on there! At the bottom with all the other little lines of origins |
12:27 |
* Dyrcona |
has considered trying to install Evergreen on an Arch virtual machine, but rolling release distros may not be suitable for production use. |
12:29 |
Dyrcona |
My first distro was Slackware, but it has fallen too far behind for my taste these days. |
12:44 |
* Dyrcona |
tries to get 3 auto-renewals that got error state after successfully renewing to generate the notification events. This is gonna require some Perl. |
12:46 |
Dyrcona |
I don't think I need to auth to do it, either. |
13:03 |
sharpsie |
I started with Debian in 2007 and moved to Ubuntu and landed on Fedora since 2014 or so |
13:04 |
sharpsie |
one of these days I will get back to my RHEL/Rocky branches :-/ |
13:09 |
Dyrcona |
Actually, I misspoke. My first distro was MkLinux on a PPC 601 PowerMac back in '96 or so. I didn't get to Slackware until '99. (I tend to discount MkLinux, 'cause it was Linux on a Mach microkernel and developed mostly by Apple and OSF at first, then the community took over.) |
13:10 |
Dyrcona |
I still have the old machine in the attic. |
13:11 |
Dyrcona |
Man, 60MHz and 64MB of RAM seemed like a lot back then, even if it took 3 days to compile all of KDE. :) |
13:17 |
Dyrcona |
Weird. My work laptop just hung, and I thought I'd fixed that with a kernel switch. |
13:20 |
jeffdavis |
heh, 27 years of Linux on the desktop and it still hangs on us sometimes |
13:20 |
Bmagic |
Dyrcona: it detected those keystrokes which, when composed together, spelled sentences that spoke ill of it. So it kernel paniced |
13:20 |
Dyrcona |
Heh. |
13:21 |
Dyrcona |
Well, this is different. Looks like there was a timeout talking to the NVMe drive or the bus, I'm not sure which. The kernel parameter fixes a known issue with Intel graphics on some Dell models where the screen freezes briefly every now and then. |
13:22 |
Dyrcona |
Log says it reset the nvme bus, so that was likely why it froze. |
13:23 |
Dyrcona |
Don't think there's a kernel parameter for that one. :) |
13:23 |
jeffdavis |
I stick to stock Ubuntu on Thinkpads and it usually goes pretty smoothly, but I ran into some trouble with suspend last time I tried a fresh 22.04 install which was aggravating (surely this shouldn't be a problem anymore?) |
13:23 |
jeffdavis |
15 years on Ubuntu, I'm getting old |
13:23 |
Bmagic |
I think the babylonians had it right: use clay tablets |
13:24 |
jeffdavis |
definitely have days where I think libraries should go back to card catalogues |
13:24 |
Dyrcona |
Heh. I was ranting last night about how computing seems to be going backwards when it comes to literacy, compared to the history of human writing. |
13:25 |
berick |
jeffdavis: same. https://askubuntu.com/questions/1347924/20-04-freezes-after-waking-up-from-suspend |
13:25 |
berick |
easy fix, fortunately |
13:26 |
jeffdavis |
sweet, thanks! bookmarked for my next attempt. |
13:26 |
Dyrcona |
I got suspend to disk working on the Arch laptop. Even my "normal" suspend does hibernate thanks to a little systemd configuration. Works great. Suspend is basically turning it off. Also, kernel 6.4 rocks. The laptop seems faster than when it was on 5.15. |
13:29 |
Dyrcona |
Also, nothing wrong with card catalogs and paper records. I'm not always convinced that "progress" is progress. |
13:30 |
Dyrcona |
All right, what was I doing? Oh, yeah, devising a cstore search to find error state event for event_def 124. |
13:34 |
|
rfrasur joined #evergreen |
13:36 |
berick |
all this Bug Squash Pie has me feeling saucy. https://bugs.launchpad.net/opensrf/+bug/2032835 |
13:36 |
pinesol |
Launchpad bug 2032835 in OpenSRF "Discussion: Merge OpenSRF Into Evergreen?" [Wishlist,New] |
13:36 |
Dyrcona |
If I want to cast a timestamp to a date, I suppose I have to use the cast function? |
13:37 |
Dyrcona |
I imagine '::date' will confuse things, but I'll give a try. Worst thing is I'll get an error. |
13:39 |
|
rfrasur joined #evergreen |
13:39 |
Dyrcona |
Yeah, 'add_time::date' => 'somevalue' doesn't work. |
13:42 |
Dyrcona |
ooh. That's gonna be tricky. I want to cast the field value as well as the comparison value, and the latter is gonna be the function now(). |
13:44 |
Dyrcona |
berick: Do you know if any of these work with cstore: https://www.postgresql.org/docs/10/sql-expressions.html#SQL-SYNTAX-TYPE-CASTS ? I've never tried before. |
13:45 |
berick |
Dyrcona: for date specifically, i know you can do stuff like this w/ json_query... |
13:45 |
berick |
{"select":{"bre":[{"column":"create_date","transform":"date"}]} |
13:46 |
berick |
IIRC you can give the column an alias then use the alias the in the "where" part... mabye, I'd have to verify that |
13:47 |
Dyrcona |
Ok. Thanks. I'm about to try that in a where {"add_time":{"=":{"transform":"date"},"value":{"now":{"transform":"date}}}} |
13:48 |
Dyrcona |
Except in Perl syntax. |
13:48 |
|
terranm joined #evergreen |
13:51 |
Dyrcona |
That's not quite right. |
13:51 |
berick |
Dyrcona: another approach... "where":{"create_date":{"between":["2022-12-01", "2022-12-02"]}} |
13:51 |
Dyrcona |
Yeah, but I don't want to hardcode the dates or have to calculate them in the Perl. |
13:52 |
Dyrcona |
I might do the latter if I can't get this to work. |
13:54 |
Dyrcona |
An array works: {"add_time":{"=":{"transform":"date"},"value": ["dae", "now"]}} |
13:54 |
berick |
nice |
13:54 |
Dyrcona |
Well dae should be date. I didn't copy and paste, 'cause of Perl syntax. |
13:58 |
Dyrcona |
Now that I have my events, I just need to figure out what user data I need for the notify event, and what state the autorenew event needs to be in, etc. :) |
14:00 |
Dyrcona |
Hmm.. Just to make sure that I'm getting ones that have renewed, maybe I should join with action circulation.... |
14:09 |
jeffdavis |
I went ahead and added "Merge OpenSRF into EG?" to the agenda for the next dev meeting, seems worth some discussion. (I think it's a good idea.) |
14:12 |
Dyrcona |
This is a bad sign: "nvme nvme0: controller is down; will reset" my laptop did the freeze again. I suspect a hardware issue is developing. |
14:12 |
berick |
jeffdavis++ |
14:12 |
Dyrcona |
jeffdavis++ I think so, too. It has been suggested before. |
14:13 |
* Dyrcona |
runs a quick backup. |
14:21 |
Dyrcona |
Looking through journalctl, there are a lot more nvme errors than just for the two times the laptop "froze." |
14:22 |
Dyrcona |
I'm still not sure if it's the SSD or the controller, but the last error implies the controller has an issue. |
14:26 |
* Dyrcona |
is all over the place today. |
14:28 |
Dyrcona |
Back to my cstore/Perl utility. I started this to just fix autorenewals that had renewed, but got stuck in error state, and didn't create the notify event. I then modified the main query slightly so I can fix all of our stuck aurorenewal events including those that did not renew. |
14:29 |
Dyrcona |
Now, I'm considering changing it more to fix all of our stuck events. I've got a crude process for that already, but this one could do a more thorough job. |
14:30 |
Dyrcona |
Assuming of course that the laptop holds together. |
14:32 |
Dyrcona |
Eh, no. If I do the latter, I'd want to only look at events from the day before or older because we might run at a bad time and pick up events that haven't yet processed because of delay or that are in the middle of reacting or something. |
14:34 |
Dyrcona |
I just figured out how to do "add_time::date = now()::date" and I don't want to figure out "add_time::date = now()::date - '1 day'::interval" |
14:34 |
Dyrcona |
Y'know just for grins.... |
14:39 |
Dyrcona |
Yeah, that latter one is not so obvious. |
14:48 |
Dyrcona |
It seems kind of crazy that the start time and update time on this one event are an hour and 25 minutes apart, but I guess the start_time is set when the event is picked up, the system runs through a bunch of events along the way. |
15:06 |
|
mantis1 left #evergreen |
15:41 |
|
jihpringle joined #evergreen |
15:57 |
* Dyrcona |
is forgetting basic stuff, like do I need a transaction for $editor->update_schema_table($obj)? I don't think so. I'll try it I can always fix i in the database if I got it wrong. |
15:58 |
Dyrcona |
Never mind. That was dumb. I do need a transaction. |
16:05 |
Dyrcona |
I think it worked! |
16:05 |
Dyrcona |
At least on the three that renewed. |
16:08 |
Dyrcona |
The AutorenewNotify events were created and the error events were set to complete. |
16:56 |
|
bgillap_ joined #evergreen |
17:03 |
|
bgillap__ joined #evergreen |
17:06 |
|
bgillap joined #evergreen |
17:07 |
|
bgillap_ joined #evergreen |
17:14 |
|
mmorgan left #evergreen |
18:35 |
jeffdavis |
here's a new one: EG services suddenly started failing with "Transport error in recv()" errors in the EG logs. Our ejabberd logs show errors like this: |
18:35 |
jeffdavis |
[error] <0.17325.30> CRASH REPORT Process <0.17325.30> with 0 neighbours exited with reason: {process_limit,{max_queue,10003}} in p1_server:terminate/7 line 864 |
18:35 |
jeffdavis |
[error] <0.579.0> Supervisor ejabberd_c2s_sup had child undefined started with {ejabberd_c2s,start_link,undefined} at <0.17325.30> exit with reason {process_limit,{max_queue,10003}} in context child_terminated |
18:37 |
jeffdavis |
It happened on multiple bricks pretty much simultaneously, even though they don't share services. Things worked fine again after restarting EG services. |
18:49 |
jonadab |
Hmm. What would cause it to reach a process limit? Leftover processes not exiting for some reason maybe? |
18:51 |
jonadab |
Roughly how long had EG been up and running when the problem occurred? |
18:53 |
jeffdavis |
1 week at most since the last EG service restart |
19:17 |
jonadab |
Hmm. That hardly seems like it would have accumulated more of a long-term process leak than ever before on your setup. |
19:20 |
jeffdavis |
looking further, the cause appears to be tens of thousands of open-ils.auth.session.delete calls for a single authtoken over a 4-minute period :( |
19:30 |
jeffdavis |
we've seen this once before, where the client (I presume) floods us with a high volume of repeat open-ils.auth.session.delete calls until open-ils.auth falls over |
19:30 |
jeffdavis |
not sure how to reproduce |
19:34 |
jeffdavis |
worth noting that the open-ils.auth service didn't die, it just stopped working once ejabberd freaked out - some sort of graceful recovery in that case might be good |
22:19 |
|
bgillap__ joined #evergreen |