| Time |
Nick |
Message |
| 09:43 |
|
Dyrcona joined #evergreen |
| 09:47 |
Dyrcona |
I am running the cron jobs manually on the test server with the low RAM errors. I think badge_score_generator.srfsh is one of the culprits for causing PostgreSQL to use a lot of buffer. |
| 10:09 |
|
mantis1 joined #evergreen |
| 10:12 |
Dyrcona |
Looks like just running most of the scheduled jobs is enough to cause Pg to use a lot of buffers/cache. |
| 10:38 |
Dyrcona |
Slashing shared_buffers didn't really help. |
| 10:39 |
Dyrcona |
Hm.. Guess I could try 32GB. |
| 10:39 |
Dyrcona |
I meant 32MB... |
| 10:42 |
|
kworstell_isl joined #evergreen |
| 14:41 |
|
phasefx joined #evergreen |
| 14:41 |
|
jweston joined #evergreen |
| 15:17 |
Dyrcona |
Interesting: Running a large query causes Pg to free up some RAM. |
| 15:18 |
Dyrcona |
Well, temporarily anyway. |
| 15:29 |
|
mantis1 left #evergreen |
| 15:38 |
|
akilsdonk joined #evergreen |
| 15:38 |
|
phasefx joined #evergreen |
| 16:03 |
Dyrcona |
Heh. Memory pressure is pretty tight. Some programs don't want to run, like osrf_control. |
| 16:05 |
Dyrcona |
Nifty! "Could not connect to Redis server at private.localhost:6379: Connection refused" Wonder if that's because of low memory or because I was trying to run it with sudo. |
| 21:44 |
|
book` joined #evergreen |
