| Time |
Nick |
Message |
| 05:48 |
|
rjackson_isl_hom joined #evergreen |
| 06:00 |
pinesol |
News from qatests: Failed Running perl live tests <http://testing.evergreen-ils.org/~live//archive/2022-04/2022-04-15_04:00:02/test.49.html> |
| 08:31 |
|
rfrasur joined #evergreen |
| 08:40 |
|
mmorgan joined #evergreen |
| 09:16 |
|
Stompro joined #evergreen |
| 09:18 |
Stompro |
Bmagic, didn't you present a few years back on AWS hosting? I'm trying to track down slides or notes on that presentation to share with an MN Library that is interested in the topic. |
| 09:22 |
Bmagic |
Stompro: yes, it was Google though. We are hosting Evergreen servers on AWS as well but I didn't present on it. Just a minute, I think the link to the slides is currently broke |
| 09:24 |
Bmagic |
Strompro: does this link work for you? http://slides.mobiusconsortium.org/blake/evergreengoogledocker/#/ |
| 09:24 |
Stompro |
Yes it does, thank you. |
| 09:25 |
|
Dyrcona joined #evergreen |
| 12:11 |
miker |
@later tell jamin re action.all_circulation, that's a view that combines "live" and "anonymized" circs into a single list. so it's true that the circs that have been anonymized and written to action.aged_circulation have all their values locked in place, but not the "live" circs in action.circulation. circulation aging is a local process that has to be set up. |
| 12:11 |
pinesol |
miker: The operation succeeded. |
| 12:16 |
berick |
Dyrcona: FYI just found a fix for the npm / rxjs 100% cpu problem. LP en route |
| 12:28 |
miker |
Dyrcona: I won't be able to finish it for you before 3.9.0, i don't think, but so you know that I know (that you know that...) the "deadlock fix" does introduce some serialization of ingest (WITH DELETE ... RETURNING * INSERT ... ON CONFLICT guarantees deadlock-free ACID semantics, but by forcing some locking), but(!) it also introduces a place for us to hook into that we can use avoid that serialization in the main DYM part of the ingest code. if we |
| 12:28 |
miker |
invent a way of skipping the call to search.symspell_dictionary_reify() in the trigger (global flag, a new "Skip this part" bool for the function, etc) then we can just run a reification function (slightly different than the current reify function, because it should not care about which transaction id generated the change(s)) after all are done, and push DYM changes into the dictionary in one go at the end. |
| 12:28 |
miker |
Dyrcona's getting all kinds of good(?) news today ;) |
| 12:40 |
Dyrcona |
berick++ miker++ |
| 13:52 |
|
Guest4933 joined #evergreen |
| 14:41 |
jeff |
last month I asked about workflows for physical library card distribution at schools: http://irc.evergreen-ils.org/evergreen/2022-03-22#i_503879 |
| 14:41 |
jeff |
at the time, I said something like: |
| 14:41 |
jeff |
> I'm leaning toward (roughly) "print two labels, one to go on the envelope, one to go on the library card, scan the card and the label to establish a link between "this envelope received this card", batch create cards based on that data, spot check, distribute. |
| 14:41 |
jeff |
(and that we might skip the label on the card) |
| 14:43 |
jeff |
we ended up doing something very similar, but we actually skipped the sticker on the envelope: we put the sticker (with temporary linking barcode) on the card, put a letter (with no variable per-patron data on it) in the envelope, put the card in the envelope so that it was visible through the large window on the front, and we did the "scan barcodes to link the cards" step after everything was all sealed |
| 14:43 |
jeff |
in. |
| 14:45 |
jeff |
paying a copy shop to fold the letters was something we decided to do very early on in the process, and we made good use of the envelope sealing features of our postage meter ("seal only", no postage requred because these are being distributed through the schools) |
| 14:55 |
|
stephengwills joined #evergreen |
| 15:35 |
berick |
any way to see conf. proposals we've already submitted? or, i guess, all of them? |
| 15:56 |
rhamby_ |
berick: https://docs.google.com/spreadsheets/d/1mh6ipfBJNp59uI0DMqiLF1Or0MU-3QeWjXLnbWSRqwk/edit?usp=sharing |
| 16:03 |
JBoyer |
I wish I were as excited about anything I've ever done as Google is about those silly PeopleChips. Every time I open a Sheets doc it's /still/ all "HAVE YOU HEARD ABOUT THESE AWESOME PEOPLECHIPS?" |
| 16:30 |
Bmagic |
when hosting NCIP behind nginx. Are there any considerations for nginx configuration? Or just standard proxy pass rules like the rest of Evergreen traffic? |
| 16:34 |
Dyrcona |
Bmagic: Nothing special needs to be done. It just works. |
| 16:35 |
Bmagic |
Awesome. Maybe the "Require" IP lines are a problem when all the traffic "looks" like it's from localhost (to Apache) ? |
| 16:36 |
jeff |
also, which NCIP code are you working with? the answers may vary. :-) |
| 16:39 |
Dyrcona |
Bmagic: There's some config for getting the remote IP. It may have made it to Launchpad. |
| 16:39 |
Bmagic |
working with our repo: https://git.evergreen-ils.org/?p=NCIPServer.git;a=summary |
| 16:42 |
Bmagic |
I think I may have got it. Just putting these lines in the apache ncip config seems to have taken care of it (both for 7080 and 7443) RemoteIPInternalProxy 127.0.0.1/24 RemoteIPInternalProxy ::1 RemoteIPHeader X-Forwarded-For |
| 16:43 |
Bmagic |
I'm testing with a web browser, and correct me if I am wrong, but I know that I've gotten through to the NCIP server when hitting the URL: /NCIP and the browser gives me "XML Parsing Error: not well-formed" error? |
| 16:43 |
Dyrcona |
Bmagic: Yeah, that should fix it. I was just typing to say that I haven't found a bug for it on Lp. You also need the mod_remoteip module enabled. |
| 16:43 |
Bmagic |
Which is better than what I was getting: Forbidden |
| 16:43 |
Stompro |
What version of NCIP does that repo support? |
| 16:43 |
Dyrcona |
You have to send it a XML NCIP2 message. |
| 16:44 |
jeff |
with the disclaimer that we don't use that code, I do most of my NCIP testing using curl to throw XML files at the NCIP endpoint. |
| 16:44 |
Dyrcona |
Stompro: NCIP 2.02, but like all NCIP responders it only responds to a limited subset of NCIP. |
| 16:44 |
Bmagic |
thanks yall, it was the RemoteIPHeader stuff that I needed |
| 16:45 |
Stompro |
Dyrcona, hmmm, I was told we had to pay OCLC a couple thousand for a v1 to v2 translator service... |
| 16:46 |
Dyrcona |
Bmagic: You can try this, edited for your environment: https://gist.github.com/Dyrcona/34f36ba7aacd7b94e31f35c8a54f7f7e |
| 16:46 |
Stompro |
Oh, nevermind, it is the other way around, they only support V1, so it is a V2 to V1 translator service. |
| 16:47 |
Bmagic |
Dyrcona++ thanks for the perl NCIP test script |
| 16:47 |
jeff |
depending on the subset and dialect of NCIP that you need to talk, paying OCLC may or may not make sense for you. It's unclear if that figure you just quotes is one-time or recurring. :-) |
| 16:47 |
jeff |
s/quotes/quoted/ |
| 16:48 |
Dyrcona |
Stompro: Evergreen's NCIPServer really only supports 1 remote vendor: Autographics ILL Product (I forget the product's name). |
| 16:48 |
jeff |
Auto-Graphics SHAREit |
| 16:48 |
Dyrcona |
Share-It... That's what it's callde |
| 16:48 |
Dyrcona |
:) |
| 16:48 |
Dyrcona |
jeff++ |
| 16:50 |
Stompro |
It is a very reasonable fee for setup and a small ongoing fee. (I was just mis-remembering the why behind the gateway.) |
| 16:50 |
Dyrcona |
Bmagic: You may also need the patch from bug 1732485. (It's not a patch for Evergreen, but for a Perl module used by NCIPServer.) |
| 16:50 |
pinesol |
Launchpad bug 1732485 in NCIPServer "Crash with mime type application/xml on recent distros" [High,Confirmed] https://launchpad.net/bugs/1732485 |
| 16:51 |
Bmagic |
yep, got that! The README was helpful |
| 16:52 |
Bmagic |
afk |
| 16:52 |
Dyrcona |
I want to remove the dependencies on Plack and Dancer at some point. That was how rangi started the Koha implementation, so I stuck with it thinking we'd end up being compatible. |
| 16:56 |
Dyrcona |
Going back to the RemoteIP config, I've added those at the Apache server level configuration so that they work for all parts of Evergreen and remote IPs get logged. It doesn't work for Websockets, however. We've discusses that a few times and people should be able to find it in the IRC logs if they want more information. |
| 16:58 |
Dyrcona |
Maybe I'll get some time after we move to Mobius hosting to work on some of these cleanup bugs. |
| 17:04 |
|
Stompro joined #evergreen |
| 17:14 |
|
mmorgan left #evergreen |
| 18:00 |
pinesol |
News from qatests: Failed Installing Angular web client <http://testing.evergreen-ils.org/~live//archive/2022-04/2022-04-15_16:00:02/test.29.html> |
| 18:23 |
|
rjackson_isl_hom joined #evergreen |
| 19:13 |
|
rjackson_isl_hom joined #evergreen |
