Evergreen ILS Website

IRC log for #evergreen, 2021-01-28

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat

All times shown according to the server's local time.

Time Nick Message
01:49 troy__ joined #evergreen
01:49 troy__ joined #evergreen
06:00 pinesol News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>
07:15 rjackson_isl_hom joined #evergreen
07:48 csharp Bmagic: re EBSCO - assuming it's EDS, we don't allow them real-time updates, but rather do a manual export that we upload to them monthly (but that may not be feasible for your situation)
07:51 collum joined #evergreen
07:58 JBoyer I don't think it's as simple as that, berick was saying he's seen a ton of them and they're not even using EBSCO for anything. (at least that's what I took away from the last time we talked about it)
07:59 mantis1 joined #evergreen
08:12 rfrasur joined #evergreen
08:22 alynn26 joined #evergreen
08:32 mmorgan joined #evergreen
08:44 collum_ joined #evergreen
08:46 troy___ joined #evergreen
08:46 troy___ joined #evergreen
08:50 Dyrcona joined #evergreen
09:04 Bmagic csharp: we do the same
09:06 jvwoolf joined #evergreen
09:06 jvwoolf jeff++
09:06 jvwoolf JBoyer++
09:07 jvwoolf For the barcode printing information yesterday. Had to step away, but I appreciate it!
09:09 berick csharp: JBoyer: we also do monthly batch exports to ebsco, which I hadn't put 2-2 together at the time, but in the end we also don't need/want the real time data.
09:11 JBoyer Ah, ok. That is a lot closer to ok than "we're being scraped by a vendor we don't even use." Still not great but the eyebrow has been lowered.
09:11 berick exactly
09:17 Dyrcona We do a weekly ISBN dump for Novelist and a monthly MARC dump for EDS and send them to EBSCO.
09:20 Dyrcona I suspect that the EBSCO "attacks" are them looking up data for the on-the-shelf feature of Novelist.
09:33 Bmagic I can say with 95% certainty that unapi was the catalyst for the DOS yesterday. That's 6 bricks at 8 CPU's each, 48 CPU's maxed out. Today, we have a limit on that URL at nginx
09:35 berick Bmagic: mind sharing your nginx rule?  just curious
09:36 Dyrcona We blocked the EBSCO IPs at the firewall because we don't use on the shelf.
09:39 Bmagic berick: yep, I was thinking of committing it to a working branch just for sharing, and maybe inclusion for options for devops
09:39 Bmagic I'll get something up today
09:39 berick Bmagic++ wasn't trying to make you do a bunch of work, but thanks
09:40 Bmagic You think it's worth an associated bug?
09:40 berick Bmagic: i think that could be useful, yes, if nothing else for documentation purposes
09:42 Dyrcona We should adjust the sample configuration for nginx. It should have TLSv1.3 added.
10:26 Dyrcona Backporting sucks.
10:27 Dyrcona I'm told that curbside isn't working on our training server with 3.5.2, but it works fine with 3.2.10 on production. I need to see if we're missing commits or I did a bad conflict resolution. It may be easier just to build a new branch.
10:27 Dyrcona Not to mention the other stuff backported to training.....
10:28 Dyrcona So much for working on Perl test today.....
10:28 Dyrcona s/test/tests/
10:31 Dyrcona Doesn't help that I have a list of 18 commits for the curbside feature. I'll have to check that list is accurate. I suppose I could try the 3.5 backport branch, but I seem to recall not using it because it was missing something.
10:33 Dyrcona Oh, that's right. There isn't a 3.5 curbside branch.
10:38 Dyrcona Oorah! git log working/collab/gmcharlt/lp​1879983_curbside_pickup_pr ^35fca97a4ec4f5009735544a26160060eaff7610 | grep -c ^commit
10:38 pinesol Dyrcona: [evergreen|Bill Erickson] 3.5 Post-Beta Release Notes Additions - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=35fca97>
10:49 Bmagic berick: bug 1913610
10:49 pinesol Launchpad bug 1913610 in Evergreen "Evergreen needs a way to mitigate DOS" [Undecided,New] https://launchpad.net/bugs/1913610
10:59 Dyrcona I wonder if bisect will help? I haven't found the missing code, yet.
11:02 Dyrcona Bmagic: Your bug title sounds dire.
11:02 Bmagic :)
11:02 alynn26_away joined #evergreen
11:19 Dyrcona OK. Where does curbsideDisabled() come from?
11:20 Dyrcona grep -R curbsideDisabled Open-ILS/ turns up 1 use and 0 definitions for it.
11:22 AFloyd__ joined #evergreen
11:34 Dyrcona Well, after all that, it looks like a bad conflict resolution with the access key fix.
11:35 csharp @who is under attack from EBSCO?
11:35 pinesol yeats is under attack from EBSCO.
11:36 berick heh
11:49 Dyrcona BTW, https://git.evergreen-ils.org/?p=working/E​vergreen.git;a=shortlog;h=refs/heads/colla​b/dyrcona/rel_3_5%2Bcurbside%2Bstripe_v3
11:49 Dyrcona That's curbside and Stripe V3 backported to rel_3_5 for those following along at home.
11:55 Dyrcona Guess it's close enough to lunchtime.
12:07 Cocopuff2018 joined #evergreen
12:28 rfrasur joined #evergreen
12:30 Bmagic csharp berick Dyrcona: with the new limit imposed, I'm trapping a lot of limits on URI's like this one: "GET /opac/extras/unapi?id=tag:acq.open-ils.org:bib​lio-record_entry/877750/-&format=marcxml-uris
12:30 Bmagic The origin is from the bricks themselves...
12:31 sandbergja_ joined #evergreen
12:36 jeff (obligatory "the call is coming from inside the house!")
12:37 berick Bmagic: that does look like the ones we were getting as well
12:37 berick they were def. coming from ebsco, though
12:38 Bmagic it seems like the Evergreen code.. perl? or somewhere is perhaps relaying an external request back as an internal request?
12:38 Dyrcona I think the Evergreen bug should be marked a duplicate of the OpenSRF.
12:39 Dyrcona Bmagic: What?
12:39 Bmagic the Evergreen Z39.50 service comes to mind
12:39 berick Bmagic: what makes you think it's coming from the brick?
12:39 Bmagic Dyrcona: just spit balling, but since the requests are from itself
12:39 Bmagic berick: the NGINX IP address recorded in the limiting error log
12:40 berick Bmagic: hm, i wonder what the access log shows.  ours was showing external IP
12:40 Dyrcona Bmagic: Do you have the proper set up for apache to log the external IPs? That could actually use better documentation. I've had to figure it out twice.
12:40 berick also what Dyrcona said
12:40 Dyrcona Imagonna bug that.
12:41 Bmagic these are nginx logs
12:42 Dyrcona Z39.50 does talk to the local apache/nginx. We run ours on its own server without nginx.
12:43 Bmagic I'm starting to think these are the aws dynamic load balancer internal IP's.... googling
12:43 Dyrcona simple2zoom does that: converts Z39.50 to SRU and back again.
12:50 Dyrcona launchpad-- bugs--
12:52 Bmagic yep, those requests are likely external, but my logs show the internal IP space because the load balancer isn't configured to reveal the source IP in the headers... Working on a solution. But yeah, these are external requests like you thought
12:55 Bmagic the external IP needs to traverse the load balancer, then nginx, then apache2. :) pretty sure the last two are correct. The X-Forwarded and X-Real NGINX configs are in place like the ones shown in my opensrf working commit 08a7299b762ed366496a6a4fc14
12:57 Dyrcona Oh, I already did: bug 1908540
12:57 pinesol Launchpad bug 1908540 in Evergreen "Add Apache proxy logging example" [Wishlist,New] https://launchpad.net/bugs/1908540
12:58 collum joined #evergreen
13:00 sandbergja_ joined #evergreen
13:05 Bmagic Dyrcona++
13:08 Bmagic Dyrcona: the commit message is confusing me. CustomLog "|/usr/bin/logger -p local7.info" common   -> CustomLog "|/usr/bin/logger -p local7.info" proxy     niether of the LogFormat definitions are named "proxy", is that a mistake/typo. Should it be "proxy-common" or "proxy-combined"
13:10 Bmagic asciidoc message rather
13:13 Dyrcona Bmagic: Fixed, rebased, and force-pushed.
13:13 Bmagic wowsa
13:13 Bmagic Dyrcona++
13:14 Dyrcona I see it's gonna be one of those event-driven days where I don't get to what I planned to do.
13:16 Dyrcona When I see bugs related to double clicking, I feel like saying, "Hello! Welcome to the Worldwide Web!"
13:16 Dyrcona Double click to open an application was a dumb idea, because it is pretty much single click everywhere else.
13:18 Dyrcona Now, to figure which thing to deal with next....
13:34 Bmagic format c:
13:39 Dyrcona Command 'format' not found, did you mean....
13:41 mantis1 left #evergreen
13:41 Dyrcona So, I'm leaning toward skipping this month's point releases again and going with the scheduled date in February.
13:43 mantis1 joined #evergreen
13:49 Bmagic Dyrcona: like that term: event-driven day. That has been my last 8 days. This new unapi nginx limitation has made a huge difference today already. It's stopped over 6k requests in the last 12 hours alone. So my event-driven days might be coming to an end. ::crosses fingers::
13:50 pinesol Showing latest 5 of 6 commits to Evergreen...
13:50 pinesol [evergreen|Jane Sandberg] LP 1857351: set basic admin interface field order using route - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=7f23544>
13:50 pinesol [evergreen|Jane Sandberg] LP 1857351: reorder a few other fm-editors - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=05abce1>
13:50 pinesol [evergreen|Jane Sandberg] LP 1857351: admin-page grid column order respects specified fieldOrder - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=cedd847>
13:50 pinesol [evergreen|Jane Sandberg] LP 1857351: fix field order for Recurring Fine Rules admin interface - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=85bfd65>
13:50 pinesol [evergreen|Jane Sandberg] LP 1857351: fix field order for Course Terms - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=ced3a85>
14:04 sandbergja_ joined #evergreen
14:23 berick grabbing 1243
14:24 pinesol [evergreen|Bill Erickson] LP1894131 Sticky catalog holdings org select - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=1a9f22f>
14:24 pinesol [evergreen|Bill Erickson] LP1894131 DB Stamp: Sticky Holdings Org - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=a8d4d1b>
14:24 pinesol [evergreen|Bill Erickson] LP1901760 Shared Worker minor lint repair - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=45b0e0f>
15:04 jihpringle joined #evergreen
15:30 mantis1 left #evergreen
16:04 khuckins joined #evergreen
17:11 mmorgan left #evergreen
17:33 Cocopuff2018 joined #evergreen
18:01 pinesol News from qatests: Failed Installing Angular web client <http://testing.evergreen-ils.org/~live//arch​ive/2021-01/2021-01-28_16:00:03/test.29.html>
18:31 pinesol [evergreen|Terran McCanna] LP1855761 Typo when creating new carousel mapping - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=600eedd>
18:31 pinesol [evergreen|Terran McCanna] LP1855761 Another typo fix - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=e7b5cdc>
18:31 pinesol [evergreen|Michele Morgan] LP1855761 Fix additional occurrences of 'Succeessfully' - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=2066633>
18:37 Cocopuff2018 joined #evergreen
19:21 sandbergja Weird that the test was only failing on Chrome
19:21 sandbergja But I'm glad that it failed somewhere, because it did rightly find a goofy typo
19:23 pinesol [evergreen|Jane Sandberg] LP1857351 (follow-up): Fix typo that led to failing test - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=43a802a>
19:34 sandbergja_ joined #evergreen
20:12 troy__ joined #evergreen
20:12 jonadab joined #evergreen
20:17 kip joined #evergreen
22:29 sandbergja__ joined #evergreen
22:32 berick_ joined #evergreen
22:33 yeats_ joined #evergreen
22:34 dbs_ joined #evergreen

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat