| Time |
Nick |
Message |
| 00:56 |
|
mrisher joined #evergreen |
| 03:05 |
|
sandbergja joined #evergreen |
| 04:48 |
|
alynn26 joined #evergreen |
| 05:41 |
|
khuckins joined #evergreen |
| 05:51 |
|
khuckins joined #evergreen |
| 05:53 |
|
khuckins_ joined #evergreen |
| 07:11 |
|
rjackson_isl_hom joined #evergreen |
| 07:57 |
|
rfrasur joined #evergreen |
| 08:01 |
|
Dyrcona joined #evergreen |
| 08:07 |
|
mantis1 joined #evergreen |
| 08:41 |
|
mmorgan joined #evergreen |
| 08:58 |
Dyrcona |
So, after vacuum full on training it dropped from 476GB to 383GB as reported by df -h. I think we're doing stuff in the database that autovacuum can't keep up with. |
| 08:58 |
Dyrcona |
BTW, that's only after 2 weeks of being up and very light use. |
| 08:58 |
Dyrcona |
I reloaded the db from a dump around January 20. |
| 09:05 |
JBoyer |
I think it's probably necessary to change the autovacuum_*_scale_factor values for our larger tables; the vacuum scale factor defaults to 20% and the analyze to 10%. For something like metabib.* those can be ridiculous. |
| 09:06 |
JBoyer |
But they need to be changed on a per-table basis or AV will be running when changing ~51-55 rows in other smaller tables. |
| 09:08 |
Dyrcona |
OK. I'll look into that. We have some huge tables. |
| 09:08 |
Dyrcona |
Search is working again, so it was that the db needed a vacuum. |
| 09:09 |
Dyrcona |
I do see autovacuum working on the hold_request and hold copy map tables in production a lot. They get heavy use throughout the day as people process pull lists and circulations. |
| 09:17 |
|
collum joined #evergreen |
| 09:23 |
JBoyer |
A quick explanation for the logs: The autovacuum_vacuum_threshold is the number of rows that have to be changed in a table for AV to run on it, and the autovacuum_vacuum_scale_factor is the percentage of the table size that is added to that to come up with the actual threshold. |
| 09:24 |
JBoyer |
So a table like money.billing with, lets say, 21770000 rows and default AV settings isn't vacuumed until you change 4354050 or more rows. |
| 09:24 |
JBoyer |
That's not great. |
| 09:59 |
Dyrcona |
I'm trying to figure out from the database if certain tables/views are used by reports. Has anyone got any magic sauce for that already? |
| 10:00 |
csharp |
you mean at the reporter.template level? |
| 10:02 |
Dyrcona |
Yeah. I looked at a couple and I think I can search for the field mapper class, but I'm uncertain if that is thorough enough. |
| 10:06 |
Dyrcona |
Ayup. This seems to work for one of our custom report views: select id from reporter.template where data ~ 'cash_report_mimic' |
| 10:13 |
csharp |
yeah |
| 10:36 |
|
malexander joined #evergreen |
| 10:54 |
Dyrcona |
Anyone opposed to a commit to add an Emacs modeline to the bottom of the fm_IDL.xml, similar to the vim line that is already there? |
| 10:57 |
Dyrcona |
bug 1914625 |
| 10:57 |
pinesol |
Launchpad bug 1914625 in Evergreen "Add Emacs mode to fm_IDL.xml" [Wishlist,New] https://launchpad.net/bugs/1914625 - Assigned to Jason Stephenson (jstephenson) |
| 10:58 |
Dyrcona |
No branch, yet, but will have in a few minutes. |
| 11:14 |
Dyrcona |
Actually, looking at the IDL, it seems that it has a mix of tabs and spaces, a.k.a. "smart" tabs. I wonder if that vim line is working or those of using Emacs have hosed it. :) |
| 11:19 |
berick |
yeah, and other editors |
| 11:19 |
berick |
the vim line is working, can confirm |
| 11:24 |
csharp |
drone exhaustion update - applying the patch from bug 1913811 has improved things somewhat, but still seeing high drone counts |
| 11:24 |
pinesol |
Launchpad bug 1913811 in Evergreen "Too many parallel open-ils.actor.user.itemsout.notices requests" [High,Confirmed] https://launchpad.net/bugs/1913811 |
| 11:24 |
csharp |
not having to restart every hour anymore though |
| 11:28 |
pinesol |
[evergreen|Bill Erickson] LP1913811 Items out notice summary API batching - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=cf2b64e> |
| 11:28 |
pinesol |
[evergreen|Jeff Davis] LP1913811 progress dialog on Items Out grid - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=faf48f6> |
| 11:29 |
berick |
csharp++ |
| 11:29 |
berick |
csharp: do you have bug 1913458 merged as well? |
| 11:29 |
pinesol |
Launchpad bug 1913458 in Evergreen "Record bucket 'Add To Bucket' actions should be batched or serialized." [Undecided,New] https://launchpad.net/bugs/1913458 |
| 11:34 |
csharp |
berick: as it turns out, we don't! |
| 11:34 |
csharp |
applying it now |
| 11:35 |
berick |
cool, it addresses a couple of scenarios |
| 11:48 |
Dyrcona |
berick++ csharp++ |
| 11:49 |
Dyrcona |
berick: Should I untabify the IDL as part of this branch? |
| 11:49 |
|
mrisher joined #evergreen |
| 11:50 |
Dyrcona |
It looks like just setting tab-width in emacs doesn't do the trick. I am looking for the xml offset variable that I need to set. |
| 11:50 |
berick |
Dyrcona: i have no strong feeling either way. not really sure why decided on tabs for fm_IDL.xml to begin with... |
| 11:50 |
berick |
maybe it started as tabs and modeline just helped set the width |
| 11:51 |
csharp |
"fm_IDL.xml Part II: The Whitespace" coming soon to servers near you |
| 11:53 |
Dyrcona |
Well, XML doesn't care about tabs v. spaces. This started because I was cleaning up our custom entries and noticed the indentation was off. |
| 11:55 |
Dyrcona |
On "tabs v. spaces," I have personally flip-flopped favoring one or the other every few years, though I think I'm stuck on spaces for good, now. :) |
| 11:55 |
Dyrcona |
I'll leave the bug up and figure something out. This is not urgent, obviously. |
| 11:58 |
Dyrcona |
"Tabs for indentation. Spaces for alignment." :) |
| 12:02 |
|
jihpringle joined #evergreen |
| 12:29 |
|
csharp joined #evergreen |
| 12:59 |
|
collum joined #evergreen |
| 13:23 |
|
khuckins joined #evergreen |
| 13:31 |
|
sandbergja_ joined #evergreen |
| 13:32 |
|
jihpringle joined #evergreen |
| 13:38 |
|
khuckins joined #evergreen |
| 13:40 |
|
khuckins left #evergreen |
| 13:41 |
|
malexander joined #evergreen |
| 13:44 |
csharp |
berick: with both patches in place, I've seen a vast improvement in drone health - thanks! |
| 13:46 |
berick |
csharp: nice! |
| 13:52 |
|
csharp joined #evergreen |
| 13:57 |
|
Guest50176 joined #evergreen |
| 13:59 |
|
akilsdonk joined #evergreen |
| 13:59 |
|
miker joined #evergreen |
| 13:59 |
|
lisacarlucci joined #evergreen |
| 13:59 |
|
abneiman joined #evergreen |
| 14:00 |
|
rhamby joined #evergreen |
| 14:01 |
|
lstratton joined #evergreen |
| 14:17 |
|
sandbergja_ joined #evergreen |
| 14:35 |
Dyrcona |
gmcharlt: Are you still looking at bug 1890629? |
| 14:35 |
pinesol |
Launchpad bug 1890629 in Evergreen 3.5 ""Email checkout receipts by default?" Doesn't Always Display in Patron Self-Registration" [Medium,Confirmed] https://launchpad.net/bugs/1890629 |
| 14:45 |
|
jvwoolf1 joined #evergreen |
| 14:49 |
|
malexander_ joined #evergreen |
| 15:43 |
|
malexander__ joined #evergreen |
| 15:52 |
|
mantis1 left #evergreen |
| 16:16 |
jeffdavis |
open-ils.actor #drones=202/200 101% |
| 16:16 |
jeffdavis |
^ didn't realize that was possible |
| 16:16 |
csharp |
jeffdavis: oof |
| 16:58 |
|
sandbergja_ joined #evergreen |
| 16:58 |
|
mmorgan left #evergreen |
| 17:19 |
|
jihpringle joined #evergreen |
| 17:20 |
Bmagic |
csharp: the DOS "fix" for us was the nginx limitation on the unapi URL OpenSRF Bug 1913617 |
| 17:20 |
pinesol |
Launchpad bug 1913617 in OpenSRF "NGINX could use a DOS mitigation example" [Undecided,New] https://launchpad.net/bugs/1913617 |
| 18:00 |
pinesol |
News from qatests: Testing Success <http://testing.evergreen-ils.org/~live> |
| 18:20 |
|
sandbergja_ joined #evergreen |
| 18:34 |
|
sandbergja_ joined #evergreen |
| 18:44 |
|
sandbergja_ joined #evergreen |
| 19:41 |
|
rhamby joined #evergreen |
| 19:41 |
|
akilsdonk joined #evergreen |
| 20:58 |
|
malexander joined #evergreen |
