| Time |
Nick |
Message |
| 07:18 |
|
collum joined #evergreen |
| 08:08 |
csharp_ |
@decide localhost or localhosot |
| 08:08 |
pinesol |
csharp_: That's a tough one... |
| 08:20 |
csharp_ |
berick: for when you're around, I got a fascinating Redis error: https://pastebin.com/7Dz1Smhf |
| 08:41 |
|
mmorgan joined #evergreen |
| 08:56 |
|
Dyrcona joined #evergreen |
| 09:00 |
JBoyer |
Reminds me of this gem: |
| 09:00 |
JBoyer |
root dev:/tmp/systemd-private-b89d9ba0a9714077b0a309f33029b8cd-apache2.service-KXwEDg/tmp/eg_template_cache# ls |
| 09:00 |
JBoyer |
'#' Comment legacy line openils out the this to TPAC use |
| 09:01 |
JBoyer |
(harmless, but apache holds no truck with comments that come after config lines) |
| 09:06 |
|
dguarrac joined #evergreen |
| 09:15 |
|
sandbergja joined #evergreen |
| 09:21 |
csharp_ |
offending line: $logger->activity("No skip check: barcode=$barcode seen_barcode=".$seen_barcode{$1}." status_changed_time=".$c->status_changed_time." ou_setting=".ol_get_org_setting('circ.offline.skip_checkout_if_newer_status_changed_time')) |
| 09:22 |
csharp_ |
it's specifically dying on $seen_barcode{$1} because $1 is uninitialized - I'm not deep enough into perl hashes to know, but I can't find anything about what $1 would do |
| 09:22 |
csharp_ |
also, about to just comment out all logger lines, but I actually want to see logs for offline stuff |
| 09:23 |
csharp_ |
I also learned today about "//", which I have inserted where I can |
| 09:23 |
csharp_ |
I guess I could just initialize $1 somewhere too, but that seems very bad |
| 09:25 |
csharp_ |
is $1 some kind of idiomatic syntax? |
| 09:25 |
csharp_ |
or an error that didn't matter pre-Redis? |
| 09:26 |
csharp_ |
all this so we can see "seen_barcode=" in the log... :-( |
| 09:29 |
csharp_ |
regex backreference? |
| 09:30 |
csharp_ |
no regex in that subroutine I can find |
| 09:38 |
csharp_ |
actually trying to trace commit e3c90505d2a to find out how %seen_barcode is used and I'm more confused than before |
| 09:38 |
JBoyer |
In perl $1 should only be a reference to a group match (unless I'm mistaken and perl has more ugly warts than I thought) so I would wonder if it's supposed to be $seen_barcode{$barcode} or similar, having not looked at that file yet |
| 09:39 |
csharp_ |
I actually don't see where it would ever have a value at that point in the sub |
| 09:42 |
csharp_ |
I s/\$1/$barcode/ -ed the file, so at least it might have some validity |
| 09:42 |
Dyrcona |
csharp_: Well there you go, you found your bug. |
| 09:42 |
JBoyer |
Yeah, looking at it I would assume it's intended to be 1 or undef, so it should probably be $seen_barcode{$barcode} |
| 09:43 |
csharp_ |
also adding ($var // '') to anything we ever concatenate might be a good practice going forward |
| 09:43 |
Bmagic |
csharp_: that's a known issue with offline and redis.. just a second and I'll link the patch |
| 09:43 |
csharp_ |
Bmagic++ |
| 09:44 |
csharp_ |
looks like it would be an issue with any perl that emits any warning and redis |
| 09:44 |
Dyrcona |
If it's the patch I'm thinking of, it does no harm if you're still using Ejabberd. |
| 09:44 |
csharp_ |
good |
| 09:44 |
berick |
yeah there's an offfline issue |
| 09:45 |
Dyrcona |
It might slow down offline processing, but not so anyone would notice, I think. |
| 09:45 |
* csharp_ |
renames the file offfline.pl |
| 09:45 |
csharp_ |
if we can go from "no processing" to "slow processing" I'll call it a win |
| 09:46 |
mmorgan |
Bmagic: Is it bug 2091644 |
| 09:47 |
pinesol |
Launchpad bug 2091644 in Evergreen "Offline Perl CGI Interacts Poorly with Redis" [Undecided,Confirmed] https://launchpad.net/bugs/2091644 |
| 09:47 |
Dyrcona |
If you're processing thousands of transactions in a single batch, you might notice, but for "average" workloads it should be fine. |
| 09:47 |
Bmagic |
mmorgan++ |
| 09:48 |
berick |
difference is, the back end processes at the same rate, but the connection to the apache process stays open and the progress dialog in the staff client persists throughout, instead of immediately closing |
| 09:49 |
Bmagic |
right, letting go of the process allows the brick to continue without hanging the interface |
| 09:50 |
csharp_ |
mmorgan++ |
| 09:50 |
Bmagic |
csharp_: we probably should have told you about that bug before your upgrade |
| 09:51 |
berick |
yeah, didn't think about that |
| 09:51 |
csharp_ |
thanks - this is very helpful |
| 09:51 |
csharp_ |
hey, and I already have KCLS evergreen-pub as a remote! |
| 09:52 |
Bmagic |
another little tid bit csharp_: if you're dealing with it right now, the processes that are still running offline, need to be killed, then offline.pl patched, then update the offline.session table to reset the progress of any offline sessions in flight, and re-started |
| 09:52 |
berick |
Bmagic++ |
| 09:59 |
csharp_ |
Bmagic: thanks - will do |
| 09:59 |
csharp_ |
Bmagic++ berick++ Dyrcona++ JBoyer++ mmorgan++ |
| 10:01 |
|
mmorgan1 joined #evergreen |
| 10:04 |
|
sandbergja joined #evergreen |
| 10:39 |
|
Christineb joined #evergreen |
| 10:55 |
|
redavis joined #evergreen |
| 11:07 |
* Dyrcona |
considers dusting off Lp 1933984 |
| 11:07 |
pinesol |
Launchpad bug 1933984 in OpenSRF "Reporter Logging Improvements" [Undecided,Confirmed] https://launchpad.net/bugs/1933984 |
| 11:07 |
Dyrcona |
This one is bugging us again. |
| 11:10 |
|
mantis joined #evergreen |
| 11:16 |
Dyrcona |
Hmm. That OpenSRF branch has stuff not in rel_3_2..... Musta been based on something later. |
| 11:40 |
Dyrcona |
Hm.. Looks like we already have the Evergreen patch applied in production. |
| 12:06 |
Dyrcona |
Bmagic: Do you already have a way to apply OpenSRF patches baked into your build system for vms/docker? |
| 12:07 |
Bmagic |
ansible vars.yml git_cherry_picks: [] |
| 12:07 |
Dyrcona |
For OpenSRF? |
| 12:07 |
Bmagic |
oh, OpenSRF - opensrf_git_cherry_picks |
| 12:08 |
Bmagic |
copy and pasted from the config for accuracy: OpenSRF_cherry_picks |
| 12:08 |
Dyrcona |
Cool! I might just add one for the utility VM. |
| 12:08 |
Bmagic |
you da man |
| 12:08 |
Bmagic |
Dyrcona++ |
| 12:08 |
csharp_ |
@band add OpenSRF Cherrypix |
| 12:08 |
pinesol |
csharp_: Band 'OpenSRF Cherrypix' added to list |
| 12:09 |
csharp_ |
offline.pl-- |
| 12:11 |
csharp_ |
at least it's working now |
| 12:16 |
Dyrcona |
csharp_++ |
| 12:16 |
Dyrcona |
I might as well patch the bricks, too. Not sure they use the daemonize function, but just in case. |
| 12:22 |
Bmagic |
Dyrcona: sure |
| 13:59 |
|
Dyrcona joined #evergreen |
| 14:35 |
csharp_ |
@decide me or your own eyes |
| 14:35 |
pinesol |
csharp_: go with your own eyes |
| 14:57 |
csharp_ |
berick: is KCLS using this for offline? https://github.com/kcls/Offline-CKO---Windows-App |
| 15:01 |
berick |
csharp_: yes. |
| 15:01 |
csharp_ |
very interesting |
| 15:01 |
berick |
IIR, it just does checkout |
| 15:02 |
csharp_ |
berick: awesome |
| 15:02 |
csharp_ |
pretty sure we don't do offline checkin at most libraries |
| 15:02 |
csharp_ |
we warn them against it |
| 15:09 |
jonadab |
Every time we've done offline circulation, in any ILS we've had, we've always regretted it. There's never enough feedback about what succeeded and what failed. |
| 15:10 |
mantis |
jonadab++ |
| 15:10 |
mantis |
complete agree |
| 15:11 |
redavis |
I dunno. In very narrowly defined circumstance, EG has worked beautifully offline. But with no checkin. Only checkout. And a fair amount of handholding. |
| 15:29 |
|
mantis left #evergreen |
| 15:46 |
csharp_ |
jonadab: yes |
| 15:49 |
csharp_ |
AH01215: unexpected end of string while parsing JSON string, at character offset 462 (before "(end of string)") at /usr/local/share/perl/5.34.0/OpenSRF/Utils/JSON.pm line 109, <F> line 43.: /openils/var/cgi-bin/offline/offline.pl, referer: https://gapines.org/eg/staff/offline-interface |
| 15:50 |
csharp_ |
that <F> looks an awful lot like the filehandle within offline.pl :-/ |
| 15:50 |
csharp_ |
oh - it is, in fact |
| 15:51 |
csharp_ |
rubberducks++ |
| 16:04 |
Christineb |
has anyone run into issues with ISBN's starting with 979 . We get our cover images from Content Cafe and we've just noticed that the cover image does not display for bib records with 020 $a 979xxxxxxxxxxx |
| 16:06 |
|
Dyrcona joined #evergreen |
| 16:07 |
csharp_ |
Christineb: we're using Syndetics and I've not heard of trouble |
| 16:08 |
mmorgan |
Christineb: I seem to remember we had that issue and used the guidance in bug 2061568 as a workaround |
| 16:08 |
pinesol |
Launchpad bug 2061568 in Evergreen "Added Content: B&T Fake ISBNs no longer working" [Undecided,New] https://launchpad.net/bugs/2061568 |
| 16:09 |
Christineb |
thank you both!! |
| 16:09 |
Christineb |
2061568 sounds like that might work for us |
| 16:09 |
berick |
Christineb: had a similar problem not long ago and updating Business::ISBN perl mod fixed it |
| 16:10 |
berick |
as in: sudo cpan Business::ISBN |
| 16:15 |
Christineb |
thank you berick!! |
| 16:32 |
briank |
berick++ updating that perl module seemed to have solved our issue, thank you! |
| 16:34 |
Christineb |
berick++ love an easy solution :D |
| 16:36 |
berick |
yay! |
| 16:41 |
|
redavis joined #evergreen |
| 17:01 |
|
mmorgan left #evergreen |
