Time |
Nick |
Message |
03:11 |
|
jeffdavis joined #evergreen |
05:02 |
|
book` joined #evergreen |
07:39 |
|
BDorsey joined #evergreen |
08:21 |
|
rfrasur joined #evergreen |
08:26 |
|
kworstell-isl joined #evergreen |
08:27 |
phasefx |
csharp_: looks like evergreen-ils.org and friends are down? |
08:27 |
csharp_ |
an issue at ITS is affecting incoming connections to PINES/GPLS which affects most Evergreen community assets, including evergreen-ils.org, git.evergreen-ils.org, and list.evergreen-ils.org |
08:27 |
csharp_ |
phasefx: you beat me to it! |
08:27 |
phasefx |
oof, csharp++ |
08:28 |
csharp_ |
oddly, I'm chatting from an ITS server - I've seen from ITS discussions that existing connections are apparently ok in some cases |
08:40 |
|
mmorgan joined #evergreen |
08:58 |
jeff |
7:42: everything's down! ... 7:50: all better! ... 8:21: everything's down! |
09:00 |
phasefx |
we didn't start the fire... |
09:01 |
jeff |
This incident affects: [3.3 KB of text] |
09:08 |
mmorgan |
:-( |
09:10 |
csharp_ |
I think we're back |
09:10 |
csharp_ |
ITS is warning that we might not be done |
09:16 |
|
Dyrcona joined #evergreen |
09:17 |
|
mantis1 joined #evergreen |
09:26 |
mmorgan |
csharp_++ |
09:36 |
|
dguarrac joined #evergreen |
09:54 |
Stompro |
jeffdavis, I will add the query info for the hold shelf query. It seems to be related to the left join for asset.call_number. |
10:05 |
Dyrcona |
Stompro: Is it a left join with a null clause or a negative logic like a not equals? |
10:07 |
Stompro |
There is an or condtion in there. But even when I remove that part it still takes a long time. I have to attempt to install Sage 50 for our finance director, and then I should have time to post the info. |
10:12 |
|
mantis1 joined #evergreen |
10:20 |
Stompro |
Wohoo, sage installed this time. |
10:27 |
pinesol |
News from commits: LP1955403-Bootstrap opac - Google Books preview not rendering. <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=1762a17d803a77a88fa3d2d60e80f68036ccca1c> |
10:43 |
Stompro |
Dyrcona, changing that one join to having a less open ended criteria brought the run down from 30-50 to 2 seconds. |
10:52 |
Dyrcona |
Stompro: Cool! Are you going to Lp it? |
11:05 |
|
jvwoolf joined #evergreen |
11:09 |
Stompro |
Dyrcona, yep, just trying to gather everything and put it in a gist. |
11:09 |
Stompro |
Dyrcona, and my co worker brought in a cheescake, so that is also slowing me down. |
11:13 |
Dyrcona |
cheesecake++ |
11:14 |
Dyrcona |
Stompro: This is about Bug 1971745 right? |
11:14 |
pinesol |
Launchpad bug 1971745 in Evergreen "Holds shelf list can fail to retrieve results" [Undecided,New] https://launchpad.net/bugs/1971745 |
11:21 |
Dyrcona |
Sometimes there are too many ways to do something.... |
11:21 |
mmorgan |
Dyrcona++ |
11:21 |
mmorgan |
Always... |
11:22 |
Stompro |
Dyrcona, yes. |
11:24 |
Dyrcona |
Stompro: OK. Just making sure. |
11:26 |
Stompro |
Dyrcona, I finished adding the query data now to bug 1971745. |
11:26 |
pinesol |
Launchpad bug 1971745 in Evergreen "Holds shelf list can fail to retrieve results" [Undecided,New] https://launchpad.net/bugs/1971745 |
11:33 |
Dyrcona |
Stompro++ |
11:33 |
Dyrcona |
I'll have to try that with the explains later to see if I can reproduce the same. I'd like to see what happens on more recent Pg releases as well. |
11:37 |
Stompro |
I struggle with choosing which PG version, Now 10 is out of support... but the 3.9 install docs say "Evergreen support for PostgreSQL server versions greater than 10 is still exprimental." |
11:38 |
Stompro |
So, I'm not sure if I should be using 11-15. I'll have to see what people are using for their recent 3.9 upgrades. |
11:39 |
Dyrcona |
Yeah, that hasn't been updated. Master still recommends Pg 10. I started looking at Pg 15 and Evergreen last week. |
11:41 |
|
Christineb joined #evergreen |
11:42 |
Dyrcona |
I suppose I could stop banging my head against cherry-pick conflicts and work on Pg 15 + Lp1871745, but first, lunch! |
11:42 |
Dyrcona |
Oops. Typoed the bug number. |
11:47 |
Stompro |
Dyrcona, I took a look at the work you did on moving that forward, the benchmarks of the various versions. Thanks for sharing that. |
11:52 |
Stompro |
Content Cafe seems to still be down. But they have another cover art service that still seems to be up... the one that titlesource 360 uses. https://contentcafeau.azureedge.net/Jacket.svc/UUID/ISBN/Medium/Logo |
12:03 |
|
dmoore joined #evergreen |
12:04 |
|
sleary joined #evergreen |
12:10 |
|
jihpringle joined #evergreen |
12:15 |
|
sleary26 joined #evergreen |
12:23 |
Dyrcona |
Stompro: Yeah. One of the things I have yet to do is build plprofiler for Pg 15. Guess I could try that now. |
12:29 |
Dyrcona |
Oof. My previous installation is no longer in the bash history. This is gonna be interesting. There's also an upgrade from plprofiler 4.1 to 4.2.... |
12:30 |
jeff |
Stompro: ironic, since TS360 remains down. :-) |
12:36 |
mmorgan |
Content Cafe may be back. Seeing images in our dev system and in Novelist |
12:39 |
jihpringle |
mmorgan: we're seeing them on our live system this morning |
12:50 |
jeff |
Yep, looks like contentcafe2.btol.com is responding once more. |
12:51 |
jeff |
OCS and TS360 still appear down. Not sure of the status of EDI. |
12:52 |
Dyrcona |
mmorgan++ jihpringle++ |
12:52 |
mmorgan |
jihpringle++ |
12:53 |
mmorgan |
jeff++ |
12:53 |
Dyrcona |
So, installing plprofiler for different Pg versions was relatively easy. I should document it for future me... |
12:53 |
Dyrcona |
jeff++ |
13:01 |
Stompro |
mmorgan++, jihpringle++ thanks, just tested again and seems to be back for us also. I'm going to wait a few hours before I re-enable though. |
13:01 |
mmorgan |
Stompro: Probably a good idea :) |
13:01 |
mmorgan |
Stompro++ |
13:42 |
jmurray-isl |
jeff++ Verified on my dev server. |
13:51 |
Dyrcona |
Gotta love cascading failure... |
13:58 |
|
mantis1 joined #evergreen |
14:05 |
jmurray-isl |
Only the bjackson user (or root) can write to the bjackson folder. |
14:05 |
jmurray-isl |
oops |
14:08 |
Dyrcona |
:) |
14:30 |
Dyrcona |
Oh, joy.... Looks like we'll have to add version checks in a couple of the schema SQL scripts. It looks like coalesce is not immutable in Pg 15, and Pg 15 allows NULLs to be not distinct for unique indexes. We've got two unique indexes that use coalesce and need to be converted for Pg 15. |
14:32 |
Dyrcona |
I can't find anything in the release notes about changes to coalesce, but I found "Allow unique constraints and indexes to treat NULL values as not distinct" |
14:44 |
Dyrcona |
Turns out this is our problem: Mark the interval output function as stable, not immutable, since it depends on IntervalStyle (Tom Lane). This will, for example, cause creation of indexes relying on the text output of interval values to fail. |
14:45 |
Dyrcona |
I can probably fix it without anything version-specific being needed. |
15:15 |
Dyrcona |
It's weird. Some of the emails sent to gitadmin go through my spam filter and others don't. I haven't figured out why that is. |
15:16 |
Dyrcona |
Anyway... /me checks on the progress of the db build on Pg 15. |
15:17 |
Dyrcona |
Success! |
15:17 |
Dyrcona |
It still builds on Pg 14. Guess I'll try 10 through 13 before I commit any changes. |
15:19 |
Stompro |
Dyrcona++ |
16:33 |
|
jvwoolf left #evergreen |
17:02 |
|
mmorgan left #evergreen |
17:26 |
|
jihpringle joined #evergreen |
17:58 |
|
abneiman_ joined #evergreen |