Evergreen ILS Website

IRC log for #evergreen, 2019-11-13

| 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
00:59 sandbergja joined #evergreen
01:11 sandbergja joined #evergreen
01:39 sandbergja joined #evergreen
03:30 haishu joined #evergreen
06:49 agoben joined #evergreen
06:51 agoben joined #evergreen
07:08 rjackson_isl joined #evergreen
07:46 csharp jeff: yep, it was vmotion - our DB servers are supposed to be pinned to a physical host and moved anyway
07:46 csharp jeff: postgres saved the day by STAYING ALIVE through it all :-)
07:47 csharp but I thought for a few minutes there that we were going to have to switch to the secondary master replica
07:48 csharp and the third DB (second replica) was also vmotioned and postgres died off, but resumed without trouble
07:48 csharp just remembering the bad old days of slony and being very grateful
07:51 rfrasur joined #evergreen
08:08 remingtron joined #evergreen
08:15 haishu joined #evergreen
08:16 alynn26 bmagic: that is correct.  then the line "|| COALESCE(collectionHQ.quote(genres), '');
08:16 alynn26 " is added to the output.
08:39 mmorgan joined #evergreen
08:55 sandbergja joined #evergreen
09:25 yboston joined #evergreen
09:45 remingtron joined #evergreen
10:20 jeff csharp: do you know what caused postgres to die on the third db server?
10:20 haishu joined #evergreen
10:31 Dyrcona joined #evergreen
10:32 Dyrcona systemd-- # to infinity and beyond!
10:35 rfrasur lol
10:40 jvwoolf joined #evergreen
10:46 Dyrcona systemd caused my problems with the db2 server that I mentioned earlier this week. (At least, I think I did.)
10:53 jeff what was the problem?
10:54 Dyrcona Well, pid 1 stopped responding, so sytemctl could do nothing, so after that anything that wasn't already running was basically hosed.
10:55 Dyrcona Specifically, I ran apt upgrade.
10:55 Dyrcona Upgrades installed, but some would not configure: keyboard-configure got stuck for instance.
10:56 Dyrcona Turns out it was trying to talk to plymouth. Plymouth wouldn't start because systemctl couldn't communicate with pid 1.
10:56 Dyrcona I modified it's post.install to exit 0;....Next package wouldn't configure either.
10:57 Dyrcona I removed the exit 0 from the keyboard post install script and gave up on that.
10:58 Dyrcona I tried rebooting. Got some message from systemctl about not being able to connect or communicate with something.
10:58 Dyrcona I tried stopping a service. Same thing.
10:58 Dyrcona I googled a little bit. I tried restarting systemd itself. No connection.
10:59 Dyrcona I did a force reboot: systemctl --force reboot.... NO connection.
10:59 Dyrcona systemctl --force --force reboot.... Hey! Something works......
11:00 Dyrcona VFS: Cannot open root device "md0" or unknown-block(9,0)  # Except mine ends in error -5
11:01 Dyrcona More googling: initramfs is hosed or the raid is.....
11:02 Dyrcona It's approaching bedtime on a Sunday night, the server is 50 miles away, it's not responding to Ctrl-Alt-Del over the KVM remote software. (I forgot about iDRAC.) I decided to wait until the next day.
11:02 cmalm joined #evergreen
11:02 Dyrcona Or maybe that was Monday morning....
11:02 pinesol News from qatests: Failed Installing AngularJS web client - Expected 6 errors but encountered 5. <http://testing.evergreen-ils.org/~li​ve/test.28.html#2019-11-13T11:00:50,598274535-0500 -0>
11:02 pinesol News from qatests: Failed Installing Angular web client <http://testing.evergreen-ils.org/~li​ve/test.29.html#2019-11-13T11:00:50,640152635-0500 -2>
11:03 Dyrcona Anyway, drama continues on Tuesday with the iDRAC. No remote console 'cause the license either expired or was never purchased, but the hardware looks good.
11:04 Dyrcona It's still 50 miles away, and I don't feel like driving, so I try what I can try remotely, and no luck. It's only the reports server and that's all running on the main DB server.
11:04 Dyrcona So, I get into the office today, about 4 miles from the colo facility.
11:04 Dyrcona I suspect my problem is the latest kernel's initramfs not being configured because systemd.....
11:05 csharp jeff: well, looking at the logs for that DB, it didn't actually die, but replication had stopped as far as the master server was concerned and it resumed when I restarted postgres on the replica
11:05 Dyrcona So, I boot it through the iDRAC, and watch it through the kvm remote. I tried selecting the advanced options on the grub menu, but the network delay was too long.
11:07 Dyrcona Get the access card and keys for the colo. Drive over. Got inside. Get on the kvm. Power the machine off, then back on, select an older kernel in grub. It boots. I login. Run dpkg --configure -a. That works, this time, because systemd restarted of course. Reboot. All is working on the new kernel.
11:08 jeff Dyrcona: much fun.
11:08 Dyrcona csharp: I won't speculate publicly about that, but I plan to restart replication manually. I actually told the master not to try log shipping or replication while db2 was down.
11:09 Dyrcona jeff: Yes, and Devuan is looking might tempting right about now.
11:09 Dyrcona s/might/mighty/
11:11 Dyrcona In hindsight, there were some other things that I could have tried, but didn't think of until after. Not sure that they would have worked given systemd.
11:12 Dyrcona csharp: I am not surprised that replication picked back up after restarting postgres on the second db server.
11:14 jeff csharp: interesting. thanks for sharing!
11:16 Dyrcona I think I'll check the logs to see what csharp and jeff were talking about.
11:51 yboston joined #evergreen
11:52 cmalm_ joined #evergreen
11:52 cmalm_ joined #evergreen
12:22 yboston joined #evergreen
12:53 alynn26 Quick question about EDI.  We are looking at the EDI failed messages, and the only one that can see them is the Admin.  No other Global Admin or ACQ admin can see them. I am thinking it is a permission problem.  But looking at the permissions there is nothing about EDI messages.
12:53 alynn26 Any hints.
12:54 khuckins joined #evergreen
13:24 berick alynn26: are you on the page at /eg2/en-US/staff/admin/acq/edi_message ?
13:25 berick looks like that would respect any of the ADMIN_PROVIDER MANAGE_PROVIDER VIEW_PROVIDER permissions
13:25 alynn26 Yes,
13:26 cmalm joined #evergreen
13:27 berick using the 'owner' field of the linked ->account->provider as the permission context org unit
13:28 alynn26 Checked those permissions and those are set to Consortium.  hmmm
13:28 jeff and the users in question (who can't see the messages) have a working location set?
13:29 alynn26 That is one of the first things I checked.
13:29 jeff (asking basic things, since maybe the users were newly created for this task/testing and haven't yet been used in a way where a lack of working location would have been noticed)
13:29 jeff good deal.
13:33 gerson joined #evergreen
13:46 berick alynn26: could be related to https://bugs.launchpad.net/evergreen/+bug/1847805
13:46 pinesol Launchpad bug 1847805 in Evergreen "pcrud search can fail to retrieve rows that the user has access to" [High,Confirmed]
13:47 berick one experiment would be to increase the grid row count to 100 and see if anything shows
13:47 jeff berick++ good idea. i should have thought of that.
13:48 alynn26 That might be the issue, or related.
13:49 alynn26 I did increase it to 100, but the failed messages still doesn't who. All EDI messages that are successful messages do show.
13:49 alynn26 failed messages don't show.
13:49 sandbergja joined #evergreen
13:52 berick oh...
13:53 berick alynn26: what exactly is a "failed" message in this scenario?  any chance it doesn't actually exist in the database?
13:54 alynn26 put FAILED: put to ftp.baker-taylor.com failed with error: Can't call method "put" on an undefined value at /usr/local/share/perl/5.18.2/O​penILS/Utils/RemoteAccount.pm line 584.
13:54 csharp I remember that we had a lot of trouble getting perms right with those users at the beginning
13:55 csharp and that particular error is probably a bad FTP address
13:55 alynn26 logged as admin, we delete this message, and then it processes like normal.
13:56 alynn26 we checked that first.  just to make sure.
13:59 berick huh, the success of an EDI message should be irrelevant for a simple grid render
14:02 alynn26 That is what has be really puzzled.
14:05 berick alynn26: looks like the query has no ORDER-BY unless manually applied in the grid.
14:05 berick maybe sort on one of the columns (e.g. time of error)
14:05 berick it's possible they are sorted to the end by pure happenstance
14:06 berick or Time Processed, etc.
14:07 gerson_ joined #evergreen
14:08 sandbergja joined #evergreen
14:08 gerson left #evergreen
14:10 gerson_ left #evergreen
14:16 berick alynn26: lastly, confirm the edi_message has a value set in the 'account' field.  no reason why it shouldn't, but it's null-able, and would affect the pcrud permission filtering
14:17 gerlevi joined #evergreen
14:21 alynn26 The failed messages do not have a value in the account field.
14:23 alynn26 That is why there not showing up.  I placed an account in one of the failed messages and it shows up now.
14:23 alynn26 Now, how to get the failed messages to show up automatically.
14:26 gerlevi92 joined #evergreen
14:38 jvwoolf1 joined #evergreen
14:40 gerlevi left #evergreen
14:43 gerlevi joined #evergreen
14:46 gerlevi left #evergreen
14:54 pinesol [evergreen|Jane Sandberg] LP1607922: Receiving items on page 2 of search results - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=f3ad285>
14:57 gerlevi joined #evergreen
15:04 gerlevi left #evergreen
15:05 yboston joined #evergreen
15:06 gerlevi joined #evergreen
15:15 gerlevi Hello, I would like to ask if any of you have already configured a carousel for the OPAC. If that is case, could you please let me know the process to make it visible?
15:20 csharp gerlevi: you've read section 4.6.1 of https://evergreen-ils.org/documentation/rel​ease/RELEASE_NOTES_3_4.html#_public_catalog, right?
15:21 csharp gerlevi: there's also a recent fix that needs to be applied: https://git.evergreen-ils.org/?p=wor​king/Evergreen.git;a=commit;h=e0d528​92b1aa320c7b4ff2c276b334dce59cc86c
15:21 pinesol csharp: [evergreen|Galen Charlton] LP#1851524: fix adding a bucket when saving a new carousel - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=e0d5289>
15:27 gerlevi csharp: I had no knowledge of the recent fix, i will try that. Thanks for your help
15:35 yar joined #evergreen
15:38 csharp gerlevi: without the fix, a bucket (required) wasn't being created.  Another thing we saw was that the shelving location is required but it lets you save without it and the error only appears in the logs
15:39 csharp gerlevi: https://bugs.launchpad.net/evergreen/+bug/1851879 is what we experienced trying to get them going
15:39 pinesol Launchpad bug 1851879 in Evergreen "Carousel Usability and Performance Issues" [Undecided,New]
15:50 gerlevi csharp: ok, one more question, do you know of any library that already has a carousel in its opac?
16:02 tlittle joined #evergreen
16:06 gerlevi joined #evergreen
17:02 khuckins joined #evergreen
17:05 mmorgan left #evergreen
17:09 sandbergja joined #evergreen
19:24 Christineb joined #evergreen
20:19 sandbergja joined #evergreen
21:36 sandbergja joined #evergreen
22:18 sandbergja joined #evergreen
23:02 pinesol News from qatests: Failed Installing AngularJS web client - Expected 6 errors but encountered 7. <http://testing.evergreen-ils.org/~li​ve/test.28.html#2019-11-13T23:02:29,616028449-0500 -0>
23:02 pinesol News from qatests: Failed Installing Angular web client <http://testing.evergreen-ils.org/~li​ve/test.29.html#2019-11-13T23:02:29,656218194-0500 -2>

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