| Time |
Nick |
Message |
| 05:00 |
pinesol_green |
News from qatests: Test Success <http://testing.evergreen-ils.org/~live> |
| 07:06 |
|
rjackson_isl joined #evergreen |
| 07:20 |
|
JBoyer joined #evergreen |
| 07:39 |
|
agoben joined #evergreen |
| 08:08 |
|
collum joined #evergreen |
| 08:27 |
|
kmlussier joined #evergreen |
| 08:44 |
|
mmorgan joined #evergreen |
| 08:52 |
|
bos20k joined #evergreen |
| 08:57 |
|
jvwoolf joined #evergreen |
| 09:05 |
|
Dyrcona joined #evergreen |
| 09:29 |
|
yboston joined #evergreen |
| 09:30 |
|
Stompro joined #evergreen |
| 09:30 |
|
Stompro joined #evergreen |
| 09:35 |
|
Stompro joined #evergreen |
| 09:46 |
|
ppound joined #evergreen |
| 10:03 |
|
mmorgan1 joined #evergreen |
| 10:54 |
|
mmorgan joined #evergreen |
| 10:55 |
|
bos20k joined #evergreen |
| 11:32 |
|
Christineb joined #evergreen |
| 11:36 |
|
jvwoolf joined #evergreen |
| 11:38 |
|
jvwoolf1 joined #evergreen |
| 11:55 |
|
khuckins joined #evergreen |
| 11:56 |
|
mmorgan1 joined #evergreen |
| 12:09 |
|
jihpringle joined #evergreen |
| 12:26 |
|
mmorgan joined #evergreen |
| 12:49 |
|
plux joined #evergreen |
| 12:51 |
ppound |
Is anyone using Relais for ILL configured to use Evergreen patron lookups with NCIP? I can't seem to find any documentation on how to configure evergreen with NCIP. |
| 12:54 |
ppound |
that is evergreen as an NCIP server |
| 13:04 |
abneiman |
ppound: at my former library we used Relais with SIP2. I don't believe there is a fully functional Evergreen NCIP server, though work has been done off-and-on over the years. |
| 13:05 |
abneiman |
Dyrcona ^^ didn't you do some work on this? |
| 13:05 |
Dyrcona |
Yes. I was just about to say that NCIPServer exists and basically just works with AutoGraphics. |
| 13:06 |
Dyrcona |
NCIPServer needs real documentation on how to set it up. I have a PDF that I send to those who ask, but it needs to be made into a README. |
| 13:07 |
Dyrcona |
There's also something called iNCIPit, that works with III's whatever it is. |
| 13:07 |
ppound |
Dyrcona: I would be interested in seeing the pdf, could you send it to ppound upei.ca |
| 13:08 |
Dyrcona |
It's on the way. |
| 13:09 |
ppound |
thanks |
| 13:09 |
Dyrcona |
jeff would know more about iNCIPit, though it's based loosely on something that I wrote. |
| 13:10 |
ppound |
abneiman: Can SIP2 do patron lookups? |
| 13:10 |
Dyrcona |
ppound: Yes. |
| 13:11 |
Dyrcona |
:) |
| 13:11 |
abneiman |
ppound: it was able to do patron & item lookups |
| 13:11 |
Dyrcona |
It can also do circulations and pay bills. |
| 13:11 |
ppound |
ok thanks |
| 13:13 |
Dyrcona |
NCIPServer can do patron lookups, item lookups, place requests, checkin, and checkout, but it is written for how a specific vendor's product uses NCIP 2. |
| 13:14 |
Dyrcona |
That's the fun part of NCIP 1 and 2. Each vendor does their own thing. |
| 13:15 |
Dyrcona |
Sorry, I misspoke: It can accept incoming items, not lookup items. :) |
| 13:26 |
|
krvmga joined #evergreen |
| 13:44 |
|
dbwells joined #evergreen |
| 14:31 |
gmcharlt |
berick: a question upon getting an instance of Hatch working... since I imagine you've thought on the matter, do you see any possibility of embedding a JRE with the Hatch installer |
| 14:32 |
gmcharlt |
or is that too much of a Java faux pas (or incipient security issue) |
| 14:35 |
berick |
gmcharlt: that was the original plan, but I punted because of the (oracle) licensing mess and because openjdk is not trivial to install on windows. |
| 14:35 |
berick |
i'm sure we could figure something out, it was more than I had time to figure out though |
| 14:35 |
gmcharlt |
gotcha |
| 14:36 |
gmcharlt |
our kingdom for a lighter-weight language that (a) could produce native executables everywhere we need it and (b) have sufficent access to native printing APIs |
| 14:37 |
berick |
you said it, brutha |
| 15:03 |
|
mmorgan1 joined #evergreen |
| 15:10 |
* kmlussier |
wonders how many times she can use the word 'cake' in the 2016 annual report. |
| 15:10 |
gmcharlt |
kmlussier: as you times as you want, so long as the cake is never a lie |
| 15:11 |
kmlussier |
FAKE NEWS! |
| 15:17 |
Dyrcona |
So, I don't know acq work flows. |
| 15:18 |
Dyrcona |
I've been given a screen shot with FILE_UPLOAD_ERROR. |
| 15:19 |
Dyrcona |
That occurs 1 time in the code in a function call upload_files, which actually appears to run after the file is uploaded (by Vandelay?). |
| 15:21 |
kmlussier |
Yeah. I've been able to find a more useful error in the logs when I've seen that. |
| 15:22 |
Dyrcona |
Well, it is supposed to log unable to read MARC file $filename, but I can't find that anywhere. |
| 15:23 |
* Dyrcona |
thinks his logging is still broken, despite JBoyer's help. |
| 15:23 |
Dyrcona |
I was looking for other messages last week and couldn't find them, either. |
| 15:24 |
Dyrcona |
So, I'm asking for hints of what else to look for. |
| 15:26 |
Dyrcona |
I can find osrf_http_translator messages with the queue name. |
| 15:26 |
Dyrcona |
Or maybe that's the PO name. They're the same in the screen shot I was given. |
| 15:27 |
Dyrcona |
Supposedly, this has been happening a lot lately. |
| 15:31 |
Dyrcona |
Records were queue 5 minutes at the time on the PC in the screen shot. (They sent the whole screen.) |
| 15:33 |
Dyrcona |
What I don't see, yet, is the link between the po and the vandelay queue. |
| 15:35 |
Dyrcona |
So, the Vandelay stuff is stored i the cache and that's it? |
| 15:36 |
Dyrcona |
I mean the filename and that stuff that I don't see in the tables. |
| 15:43 |
Dyrcona |
So, we have 2 memcached servers configured. Could it be that they sometimes hit one that doesn't have the same data? |
| 15:43 |
Dyrcona |
I've wondered how using more than 1 memcached server works. |
| 15:44 |
Dyrcona |
I notice that they have different values for bytes and curr_items. |
| 15:44 |
JBoyer |
Poof. And now I'm here. |
| 15:45 |
JBoyer |
Dyrcona, re: multiple memcached servers, I was under the impression that the key name was hashed somehow to determine which server to store/retrieve from, but I've never looked. It has to be deterministic or a close approximation thereof or chaos. |
| 15:45 |
JBoyer |
Also, how did I help with logging, turning off limits and so on, or something else? |
| 15:46 |
Dyrcona |
Turning off limits, but I think we're still dropping messages. |
| 15:46 |
Dyrcona |
I can worry about that later. |
| 15:46 |
Dyrcona |
What I don't know is the whole key to look up in memcached. |
| 15:47 |
Dyrcona |
And the servers both have about 2GB cached with over 200,000 items each. |
| 15:47 |
JBoyer |
Hmm. Things have to be done on both the sending and receiving machines, if that helps. But yeah, if your primary issue is related to Acq or EDI I'm not much help there. |
| 15:48 |
Dyrcona |
It's acq creating a po/pikclist from a MARC file upload. I'm told that a couple of libraries report often having to do it more than once lately to get it to work. |
| 15:48 |
|
wsmoak joined #evergreen |
| 15:50 |
JBoyer |
I could see how that could potentially point to memcache. Do you have all of the same servers listed in the same order everywhere you specify memcache connections? Depending on how and when things load them and then process them, this could happen: |
| 15:51 |
Dyrcona |
They should be. I use the same config files everywhere, AFAIK. Unless someone has been messing with them behind my back or some change didn't make it everywhere in the recent server migration. |
| 15:52 |
JBoyer |
A machine with both servers listed could store a key related to this operation, but if the next step (provided this isn't a single, one-shot job) takes place on a different machine, or in a service on the same machine with it's own memcache settings and only 1 is specified, you've got about a 50-50 chance of that working. |
| 15:52 |
Dyrcona |
(That was not fun with people asking for changes in the middle of moving 20 servers and replacing them with VMS and hardware issues with the NFS server to boot.) |
| 15:52 |
Dyrcona |
Both servers are listed in each config file, or they were when I started here. |
| 15:53 |
Dyrcona |
I'll download them all and check. |
| 15:54 |
JBoyer |
They're also specified in the apache configs, which would probably cause additional problems if all of them didn't match, but it's one more thing to consider. |
| 15:59 |
jihpringle |
Drycona: for your acq issue, there are two errors that occur on the Load MARC Order Records interface |
| 16:00 |
jihpringle |
1. Acq Import Error generally indicates that there is an issue in the file - if holdings info is being loaded there can be a typo in the shelving location, circ modifier etc. |
| 16:00 |
jihpringle |
a single typo will kill the file |
| 16:01 |
jihpringle |
2. Database Error - usually indicates that something is missing on the upload interface (Provider, Context Org Unit, etc.) |
| 16:03 |
jihpringle |
the other issue we've run into is https://bugs.launchpad.net/evergreen/+bug/1675196 |
| 16:03 |
pinesol_green |
Launchpad bug 1675196 in Evergreen "Acq: Load MARC Order Records Interface Sometimes Times Out While Loading Records and Creating a PO" [Undecided,New] |
| 16:03 |
Dyrcona |
jihpringle: The screen shot says "FILE_UPLOAD_ERROR ->" |
| 16:06 |
Dyrcona |
So, the file from brick head 1 looks just like my canonical version from git, except for path substitutions. |
| 16:06 |
Dyrcona |
It has a single block in the globabl config where the two cache servers are set. |
| 16:07 |
Dyrcona |
I'll compare the others to the file from brick head 1. |
| 16:13 |
Dyrcona |
Does this count as an "anonymous" cache request: my $data = $cache->get_cache("vandelay_import_spool_$key"); |
| 16:13 |
Dyrcona |
I notice we have a different time out for anon cache entries. |
| 16:17 |
Dyrcona |
But that's not it, 'cause it is longer than 5 minutes. |
| 16:18 |
|
Jillianne joined #evergreen |
| 16:25 |
Dyrcona |
Well, it's not NFS, either. They all have the tmp directory mounted the same. |
| 16:26 |
Dyrcona |
And I checked that it is mounted on all of the servers. |
| 16:27 |
Dyrcona |
Even if I could figure out the full key, it would be too late not if it's an anon key. |
| 16:27 |
Dyrcona |
s/not/now/ |
| 16:36 |
Dyrcona |
Ah ha! |
| 16:36 |
Dyrcona |
JBoyer++ |
| 16:36 |
Dyrcona |
The Apache config only lists 1 memcached server. |
| 16:36 |
Dyrcona |
Can you list more than 1? |
| 16:38 |
Dyrcona |
I don't think it ever had more than 1, though. |
| 16:38 |
gmcharlt |
Dyrcona: yeah, you can |
| 16:41 |
Dyrcona |
Just two entries or is it a list? |
| 16:51 |
Dyrcona |
Even if you set more than 1, it looks like only 1 of them will be used... |
| 17:00 |
pinesol_green |
News from qatests: Test Success <http://testing.evergreen-ils.org/~live> |
| 17:10 |
|
mmorgan joined #evergreen |
| 17:18 |
|
mmorgan left #evergreen |
| 20:24 |
|
kmlussier joined #evergreen |
| 21:27 |
|
wsmoak joined #evergreen |
| 21:29 |
|
genpaku joined #evergreen |
| 23:21 |
pinesol_green |
[evergreen|Jane Sandberg] Docs: adding to the list of config.tt2 variables - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=e7a5763> |
