| Time |
Nick |
Message |
| 00:06 |
|
abneiman joined #evergreen |
| 00:06 |
|
drigney joined #evergreen |
| 00:06 |
|
sallyf joined #evergreen |
| 00:06 |
|
lbarry joined #evergreen |
| 02:33 |
|
Jillianne joined #evergreen |
| 06:00 |
pinesol_green |
News from qatests: Test Success <http://testing.evergreen-ils.org/~live> |
| 07:20 |
|
rjackson_isl joined #evergreen |
| 07:35 |
|
JBoyer joined #evergreen |
| 07:51 |
|
agoben joined #evergreen |
| 08:14 |
|
stephengwills joined #evergreen |
| 08:30 |
|
kmlussier joined #evergreen |
| 08:34 |
|
Dyrcona joined #evergreen |
| 08:40 |
|
mmorgan joined #evergreen |
| 09:06 |
|
bos20k joined #evergreen |
| 09:10 |
|
collum joined #evergreen |
| 09:16 |
|
yboston joined #evergreen |
| 09:18 |
|
krvmga joined #evergreen |
| 09:33 |
Dyrcona |
krvmga Don't bother me, I'm working on it. :) |
| 09:38 |
|
rlefaive_ joined #evergreen |
| 09:49 |
|
collum joined #evergreen |
| 09:50 |
Dyrcona |
JBoyer: Are you running NCIPServer on Ubuntu 16.04? |
| 10:02 |
Dyrcona |
So, it looks like Dancer is trying to run methods on a XML blob.... |
| 10:08 |
|
rlefaive joined #evergreen |
| 10:43 |
stephengwills |
is there a way to get the control_number_maint to set 001/003 to branch for something other that serials? right now all our bibs get the consortial name? |
| 10:50 |
stephengwills |
I have the cat.marc_control_number_identifier set properly for my branch and the global option for MARC21 spec is turned on but all bibs imported still see consortium name. |
| 10:57 |
|
Christineb joined #evergreen |
| 10:58 |
|
jvwoolf joined #evergreen |
| 11:04 |
|
jvwoolf1 joined #evergreen |
| 11:17 |
|
annagoben joined #evergreen |
| 11:45 |
stephengwills |
sorry, I see the update 0329, nm till I do more reading. |
| 11:51 |
|
rlefaive joined #evergreen |
| 12:01 |
|
khuckins joined #evergreen |
| 12:19 |
|
jihpringle joined #evergreen |
| 12:33 |
|
rlefaive_ joined #evergreen |
| 13:27 |
Dyrcona |
JBoyer: Have you tried NCIPServer on Ubuntu 16.04? I'm getting bizarre errors in production and so far have no solution. |
| 13:39 |
berick |
in case it hasn't been mentioned, here's another gitlab-ish thing: https://www.phacility.com/ -- I've never used it. |
| 13:39 |
berick |
oh hah, that's from Dyrcona |
| 13:40 |
berick |
thought I was looking at someone else g+ post |
| 13:40 |
Dyrcona |
Ha! |
| 13:40 |
berick |
Dyrcona++ |
| 13:40 |
Dyrcona |
I noticed it when looking at the latest version of the Enlightenment window manager. They apparently use it. |
| 13:41 |
Dyrcona |
Right now, I'm trying to figure out why Plack::Handler::Apache2 and Dancer::Request are not playing nice together. |
| 13:55 |
Dyrcona |
So, I'm trying Plack from CPAN. |
| 14:02 |
Dyrcona |
No one can answer me about running NCIPServer on Ubuntu 16.04? It is broken for me, but I swear I tested is once before and it worked. |
| 14:21 |
Dyrcona |
Oh, lovely. |
| 14:21 |
Dyrcona |
It's broken in production, but it just works on a test vm. |
| 14:41 |
Dyrcona |
And, it just works for me in production when I run my test script using LWP against it. |
| 14:41 |
jeff |
problem with the data production is receiving? |
| 14:42 |
|
stephengwills joined #evergreen |
| 14:51 |
Dyrcona |
Nope. Not the data. |
| 14:51 |
Dyrcona |
The log messages make it look like Plack::Handler::Apache2 is doing something wrong. |
| 14:52 |
Dyrcona |
I use LWP to post a LookupUser for myself, and it works. |
| 14:52 |
Dyrcona |
Autographics does a LookupUser, it bombs out trying to use the XML data as a filehandle object. |
| 14:53 |
Dyrcona |
So, I should probably move on to checking the production database and fixing replication. |
| 14:56 |
Dyrcona |
Autographics must have a broken POST implementation, IDK. |
| 15:11 |
JBoyer |
Dyrcona, sorry, I forgot to mark myself away and spent the day in meetings. Everything here is still 14.04. |
| 15:15 |
Dyrcona |
Looking through the apache logs, every request I make with libwww-perl gets a 200 response. |
| 15:15 |
Dyrcona |
Every request autographics makes gets a 500, without (AFAICTL) a corresponding line in the error.log. |
| 15:16 |
Dyrcona |
2017-10-10 14:39:32 bh3 root: 192.168.100.101 - - [10/Oct/2017:14:39:31 -0400] "POST /NCIP/ HTTP/1.1" 200 5839 "-" "libwww-perl/6.15" |
| 15:16 |
Dyrcona |
2017-10-10 14:39:52 bh2 root: 69.43.75.87 - - [10/Oct/2017:14:39:51 -0400] "POST /NCIP/ HTTP/1.1" 500 738 "-" "-" |
| 15:17 |
Dyrcona |
And, it's not working on bricks and not others, 'cause I've got enough log messages from them all to know. |
| 15:21 |
Dyrcona |
There's a missing word in there, but I think you can figure it out. |
| 15:21 |
JBoyer |
I don't see anything too unusual about their requests (As a debugging aid I dump all incoming and outgoing messages to the logs) |
| 15:21 |
JBoyer |
cat osrfe |
| 15:21 |
JBoyer |
Oops |
| 15:25 |
pastebot |
"JBoyer" at 64.57.241.14 pasted "General LookupUser message from AG today" (22 lines) at http://paste.evergreen-ils.org/857 |
| 15:25 |
Dyrcona |
I'm getting their requests in the logs today, because they're failing and getting dumped. |
| 15:27 |
Dyrcona |
Along with: [31842] error @0.864041> [hit #1]request to POST / crashed: Can't locate object method "seek" via package "<?xml version="1.0" encoding="utf-8"?> |
| 15:27 |
Dyrcona |
<message body> |
| 15:27 |
Dyrcona |
(perhaps you forgot to load "<?xml version="1.0" encoding="utf-8 |
| 15:27 |
Dyrcona |
"?> ... <message body> |
| 15:28 |
Dyrcona |
?) at /usr/share/perl5/Dancer/Request.pm line 110. in /usr/share/perl5/Dancer/Handler.pm l. 98 |
| 15:28 |
Dyrcona |
I look at the referenced code, and it is reading the POST from the handler. |
| 15:29 |
Dyrcona |
So, it has not even gotten to processing the actual message, yet. |
| 15:29 |
Dyrcona |
When I POST with libwww-perl, it just works. |
| 15:29 |
jeff |
where's current NCIPServer code, again? |
| 15:29 |
jeff |
on git.evergreen-ils.org or elsewhere? |
| 15:30 |
Dyrcona |
git.evergreen-ils.org |
| 15:30 |
Dyrcona |
I've been chatting with some folks in #dancer on irc.perl.org. |
| 15:30 |
Dyrcona |
The consensus there was some weird interaction of Plack::Handler::Apache2 and Dancer, but that was before I had done my own tests. |
| 15:31 |
Dyrcona |
I even removed the packaged plack and installed it from CPAN. |
| 15:32 |
jeff |
did it break suddenly, or is this after a known change was made? |
| 15:32 |
Dyrcona |
Now, it think it's something "weird" in their POST requests that Apache 2.4 or the Plack handler doesn't like. |
| 15:32 |
Dyrcona |
Yeah, we upgraded to Evergreen 2.12. and Ubuntu 16.04. |
| 15:36 |
Dyrcona |
Nice. These have picked up since the upgrade: WARN: 10 NOT CONNECTEDs returned this hour: (Top server this hour: 3 bd2-bh5) |
| 15:39 |
mmorgan |
Dyrcona: We have noticed the increase in NOT CONNECTEDs, also. |
| 15:40 |
Dyrcona |
Someone should Lp that. |
| 16:09 |
jeff |
JBoyer: do you happen to have headers for that NCIP message in the paste above? |
| 16:24 |
JBoyer |
jeff, unfortunately because of the hackish way I'm able to log them at all there aren't any headers. I did wonder if a busted MIME type might have something to do with it, but that's a WAG. |
| 16:27 |
Dyrcona |
It might be this: https://rt.cpan.org/Public/Bug/Display.html?id=111876 |
| 16:59 |
|
Jillianne joined #evergreen |
| 17:01 |
stephengwills |
Do you have a recommended sql debugger for insight into the postgresql triggers at runtime? |
| 17:02 |
stephengwills |
do we like pgAdmin for this? |
| 17:03 |
|
mmorgan left #evergreen |
| 17:09 |
Bmagic |
I just got a request for a feature that would attempt to prevent the patron from selecting the wrong format when placing holds. Something like "preferred format" on the patron account......??/ |
| 17:10 |
Bmagic |
interesting thought, anyone else had something similar? |
| 17:10 |
Bmagic |
stephengwills: Explain analyze is what I use |
| 17:11 |
Bmagic |
stephengwills: of course, you have to have a query. Having the postgres logs, then running an example query with analyze turned on might help? |
| 17:12 |
jeff |
stephengwills: are you trying to step through a trigger function, or do something else? |
| 17:12 |
stephengwills |
stepping through trigger function |
| 17:13 |
stephengwills |
still working on that 001/003 thing from this morning. everytime I try to get out…it pulls me back in! |
| 17:13 |
Bmagic |
stephengwills: I have been known to replace a function with one that puts more in the output using things like NOTICE along the way. On a test machine of course. |
| 17:14 |
stephengwills |
I was thinking of doing that. just lots of watchdog lines |
| 17:14 |
Bmagic |
in some cases, adding columns to tables to write the debug output during execution. Maybe jeff has something better |
| 17:17 |
|
Jillianne2 joined #evergreen |
| 17:20 |
jeff |
there is a debugger, but i'm not well-versed enough to recommend or guide you on its use. there's a server-side extension maintained by the EnterpriseDB folk, and the only client I'm aware of is pgAdmin. |
| 17:21 |
stephengwills |
@jeff I am trying to leverge the maintain_control_numbers trigger to fix the 003 to indicate at the BR that made them. |
| 17:21 |
pinesol_green |
stephengwills: Yeah, well, you know, that's just like uh, your opinion, man. |
| 17:21 |
jeff |
so, starting with some print-style debugging might be best. |
| 17:21 |
stephengwills |
ok |
| 17:21 |
jeff |
going back to your problem, though... what are you trying to figure out? |
| 17:21 |
jeff |
or what isn't working? |
| 17:21 |
stephengwills |
all 003s seem to be assigned the top level ou name |
| 17:21 |
stephengwills |
even when I have local settings setup |
| 17:22 |
stephengwills |
so I jut wanted to follow the $_TD and see what, in it, is causeing this route through the function |
| 17:23 |
jeff |
do these records have a value in biblio.record_entry.owner? |
| 17:23 |
stephengwills |
not sure. |
| 17:24 |
stephengwills |
... |
| 17:24 |
jeff |
if they don't, then they will fall back to org unit id 1, and use either that org unit's setting for cat.marc_control_number_identifier or they will use the shortname if org unit 1 does not have a value set for that setting. |
| 17:25 |
stephengwills |
so what i really want to do is run the bre and update the owner? on save the 003 will get fixed? |
| 17:25 |
jeff |
where "they will fall back to org unit id 1" is better written as "then the control number identifier handling for those records will fall back to org unit id 1" |
| 17:25 |
stephengwills |
that’s easy enough to test |
| 17:26 |
jeff |
the global flag cat.maintain_control_numbers also needs to be enabled, but i suspect you have that set already. |
| 17:27 |
stephengwills |
yup that’s good |
| 17:27 |
stephengwills |
almost all of our records have a null owner |
| 17:27 |
stephengwills |
:/ |
| 17:28 |
stephengwills |
you understand I only come here to embarress myself in from of my better, right? |
| 17:29 |
stephengwills |
thanks |
| 18:01 |
pinesol_green |
News from qatests: Test Success <http://testing.evergreen-ils.org/~live> |
| 18:11 |
|
stephengwills left #evergreen |
| 21:18 |
pinesol_green |
[evergreen|Jane Sandberg] Docs: updating copy buckets docs - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=ec21eb8> |
| 23:14 |
pinesol_green |
[evergreen|Jane Sandberg] Docs: documentation for Record Buckets - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=55edc1f> |
