| Time |
Nick |
Message |
| 01:29 |
|
erlehmann_ joined #minetest-dev |
| 01:37 |
|
ANAND joined #minetest-dev |
| 01:53 |
|
nephele_ joined #minetest-dev |
| 02:18 |
|
reductum joined #minetest-dev |
| 02:24 |
cheapie |
So I've been poking at the indexing on the rollback SQLite database some - almost 5 years ago, a PR of mine was merged that sped up /rollback_check *dramatically* (as in, often takes less than 1% of the time it did before then) but we've realized since then that it basically kills the performance of /rollback itself in the process. |
| 02:26 |
cheapie |
Some poking around with SQLite's "EXPLAIN QUERY PLAN" function shows that adding a second index on just the timestamp column is beneficial, and from the limited testing I have done it does seem to help. |
| 02:26 |
cheapie |
The "action" column could probably also be removed from the existing index, since that part of the index doesn't seem to actually be used (I didn't really understand indexes that well back then). |
| 02:31 |
cheapie |
I know there's likely to be some "I don't see a problem, it works fine for me" responses here - the issues only arise when the rollback DB starts to get pretty large, which doesn't always take long - on servers with heavy use of mesecons and such, it can take under a day to go over 1GB, and this will likely keep happening until #6577 is addressed. |
| 02:31 |
ShadowBot |
https://github.com/minetest/minetest/issues/6577 -- Option to prevent rollback recording excess actions |
| 05:30 |
|
calcul0n joined #minetest-dev |
| 06:30 |
|
calcul0n_ joined #minetest-dev |
| 06:40 |
|
Icedream joined #minetest-dev |
| 06:41 |
|
nerzhul joined #minetest-dev |
| 06:57 |
|
nerzhul joined #minetest-dev |
| 07:07 |
|
Kimapr joined #minetest-dev |
| 07:22 |
|
karamel joined #minetest-dev |
| 09:04 |
|
Beton joined #minetest-dev |
| 09:45 |
p_gimeno |
rubenwardy: remember there's https://gitlab.com/pgimeno/irrlicht-mirror |
| 10:30 |
|
_Zaizen_ joined #minetest-dev |
| 10:31 |
|
ShadowNinja joined #minetest-dev |
| 10:32 |
|
proller joined #minetest-dev |
| 10:42 |
|
Fixer joined #minetest-dev |
| 10:49 |
|
ShadowNinja joined #minetest-dev |
| 11:00 |
|
ShadowNinja joined #minetest-dev |
| 11:00 |
|
ShadowNinja joined #minetest-dev |
| 11:50 |
|
nerzhul joined #minetest-dev |
| 12:56 |
|
calcul0n joined #minetest-dev |
| 13:22 |
|
mizux joined #minetest-dev |
| 13:48 |
|
Taoki joined #minetest-dev |
| 13:49 |
|
turtleman joined #minetest-dev |
| 14:36 |
|
calcul0n joined #minetest-dev |
| 15:25 |
|
theenemystando[m joined #minetest-dev |
| 15:28 |
|
Darcidride joined #minetest-dev |
| 16:01 |
nerzhul |
sfan5, rubenwardy #9719 is ready for a merge, we tested it with ruben yesterday on his grafana and it's fine :) |
| 16:01 |
ShadowBot |
https://github.com/minetest/minetest/issues/9719 -- Add MetricsBackend with prometheus counter support by nerzhul |
| 16:07 |
sfan5 |
I'll look at it later |
| 18:13 |
nerzhul |
rubenwardy: i fixed your comments, it should be better, also using the metrics interface permit to drop some local duplicate values we got before :) |
| 18:16 |
|
proller joined #minetest-dev |
| 18:35 |
|
calcul0n_ joined #minetest-dev |
| 18:48 |
Krock |
will merge game#2372 in 15 minutes |
| 18:48 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/2372 -- Fix door model UV for open and close by An0n3m0us |
| 19:02 |
sfan5 |
merging game#2659 in 10m |
| 19:02 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/2659 -- Fix flammable item entities crashing by sfan5 |
| 19:02 |
Krock |
merging |
| 19:27 |
|
vesper11 joined #minetest-dev |
| 20:11 |
|
mizux joined #minetest-dev |
| 20:17 |
|
calcul0n joined #minetest-dev |
| 20:22 |
|
fluxflux joined #minetest-dev |
| 21:13 |
|
jonadab joined #minetest-dev |
| 22:23 |
|
turtleman joined #minetest-dev |
| 23:02 |
|
GreenXenith joined #minetest-dev |
| 23:04 |
|
proller joined #minetest-dev |
| 23:29 |
|
turtleman joined #minetest-dev |
