| Time |
Nick |
Message |
| 05:01 |
pinesol_green |
News from qatests: Test Success <http://testing.evergreen-ils.org/~live> |
| 07:15 |
|
rjackson_isl joined #evergreen |
| 07:16 |
|
agoben joined #evergreen |
| 08:41 |
|
mmorgan joined #evergreen |
| 10:05 |
|
Dyrcona joined #evergreen |
| 11:14 |
|
Christineb joined #evergreen |
| 11:31 |
|
jvwoolf joined #evergreen |
| 12:06 |
|
bmills joined #evergreen |
| 12:12 |
|
mmorgan left #evergreen |
| 12:31 |
|
dbs joined #evergreen |
| 12:55 |
berick |
FYI at this cstore exhaustion vector I stumbled on today: http://git.evergreen-ils.org/?p=working/Evergreen.git;a=commitdiff;h=e8139016b921541ffd53f3e431d01e5faeef7994 |
| 12:55 |
berick |
in short, in authoritative API calls, every new_editor() call opens a new connection/transaction that persists until the call is done. |
| 12:56 |
berick |
if the api call does a lot of repetitive work (e.g. batch lookups), it can mean gobbling up lots of cstore backends |
| 12:57 |
berick |
so share editor handles whenever possible |
| 12:59 |
|
jihpringle joined #evergreen |
| 12:59 |
Dyrcona |
berick: Do you think the hold_targer could trigger this? I had a cstore exhaustion event at the beginning of the week. |
| 13:04 |
berick |
Dyrcona: not in the same way. targeter doesn't make any authoritative api calls. |
| 13:04 |
Dyrcona |
OK. |
| 13:07 |
tsbere |
I got curious, looks like generally when calls using editors are in use they pass an editor in, and I don't see any obvious "the editor is made in a loop" situations |
| 13:07 |
tsbere |
At least for everything I believe is normally used by the hold targeter |
| 13:10 |
Dyrcona |
Thanks. |
| 13:11 |
|
jvwoolf left #evergreen |
| 14:04 |
JBoyer |
Happy new year's Evergreen, see you on the other side of the sun and our upgrade. |
| 14:14 |
rjackson_isl |
Jboyer++ |
| 14:15 |
Dyrcona |
JBoyer++ |
| 17:01 |
pinesol_green |
News from qatests: Test Success <http://testing.evergreen-ils.org/~live> |
| 18:00 |
|
bmills joined #evergreen |
