Issue Information
-
#005968
-
3 - Medium
-
Fixed
Issue Confirmations
-
Yes (12)No (0)
0
Originally posted by jTynne
Using trunk 16195 and Harmony, our map-server_sql keeps spiking randomly at 99% CPU usage even with all of my scripts disabled.
I've been monitoring my server for the past three days straight, modifying/updating all of my scripts in hopes that the issue was my fault. Even had the hosting company look into it and they came up with noticing that map-server_sql keeps spiking at 99% which effectively locks everything up for my players in-game.
This has been happening periodically all day long since I first started my server last Friday afternoon. We have ~40-60 players online average, however, the CPU spikes take place regardless of how many players are online.
If ANYONE can help, I will be forever grateful. Thank you!
Using trunk 16195 and Harmony, our map-server_sql keeps spiking randomly at 99% CPU usage even with all of my scripts disabled.
I've been monitoring my server for the past three days straight, modifying/updating all of my scripts in hopes that the issue was my fault. Even had the hosting company look into it and they came up with noticing that map-server_sql keeps spiking at 99% which effectively locks everything up for my players in-game.
This has been happening periodically all day long since I first started my server last Friday afternoon. We have ~40-60 players online average, however, the CPU spikes take place regardless of how many players are online.
If ANYONE can help, I will be forever grateful. Thank you!
Originally posted by MarkZD
Try malufett's tomb system:
http://rathena.org/b...added-features/
Try malufett's tomb system:
http://rathena.org/b...added-features/
Edited by MarkZD, 19 June 2012 - 07:31 PM.
Originally posted by benzanaz
still lag...
T-T
still lag...
T-T
Originally posted by benzanaz
@bumb
follow the update at r16318
the server is still LAG
not reslove.... need to fix again
@bumb
follow the update at r16318
the server is still LAG
not reslove.... need to fix again
Originally posted by GreenBox
That wasn't a fix, just an optimization
Sincerely I don't know what can cause this on mvp tomb unloading as it only unload the npc.
That wasn't a fix, just an optimization
Sincerely I don't know what can cause this on mvp tomb unloading as it only unload the npc.
Originally posted by benzanaz
@Greenbox
i'm still waiting you for fix this problem
now my server temp to turn off this option
@Greenbox
i'm still waiting you for fix this problem
now my server temp to turn off this option
Originally posted by benzanaz
bumb need to fix...
bumb need to fix...
Originally posted by spiritualkid
Anyone found a solution to this?
From my testings, this only occurs on UNIX (couldn't reproduce on Windows).
Also, the lag spikes were minimal when the player who previously killed the MVP was within the tomb's range (minimal as in barely noticeable), but the lag spikes are much worst (few seconds freeze) when the player isn't in the same map as the tomb.
Anyone found a solution to this?
From my testings, this only occurs on UNIX (couldn't reproduce on Windows).
Also, the lag spikes were minimal when the player who previously killed the MVP was within the tomb's range (minimal as in barely noticeable), but the lag spikes are much worst (few seconds freeze) when the player isn't in the same map as the tomb.
Originally posted by Ind
this makes no sense at all, the tombs are ordinary npcs with no special handling or anything like that at all.
I've tried to reproduce by killing several mvps that spawn tombs -- doppelganger/baphomet/moon/gtb/maya/others and watched the cpu usage closely, it never gone beyond 1%
this makes no sense at all, the tombs are ordinary npcs with no special handling or anything like that at all.
I've tried to reproduce by killing several mvps that spawn tombs -- doppelganger/baphomet/moon/gtb/maya/others and watched the cpu usage closely, it never gone beyond 1%
Originally posted by jTynne
Perhaps in conjunction with something else besides "just" killing them. Unfortunately, this is still an issue for us.
Perhaps in conjunction with something else besides "just" killing them. Unfortunately, this is still an issue for us.
Originally posted by benzanaz
not test/ but i'm thinking about x64 compile (linux)
not test/ but i'm thinking about x64 compile (linux)
Originally posted by Raizen
in CentOS 6.2 and 5.8 x64 i cannot reproduce the spike.
in CentOS 6.2 and 5.8 x64 i cannot reproduce the spike.
Originally posted by Lighta
We had that spike on debian64 bit too, but it was before optimisation I'll try again, but it occur more on tomb deletion than spawning new one.
We had that spike on debian64 bit too, but it was before optimisation I'll try again, but it occur more on tomb deletion than spawning new one.
Originally posted by Ind
I identified the problem and believe I fixed it, check [rev=16460]
Before my fix: cpu spike at 36% when unloading tomb
after my fix: cpu spike at 0.6% when unloading tomb.
I identified the problem and believe I fixed it, check [rev=16460]
Before my fix: cpu spike at 36% when unloading tomb
after my fix: cpu spike at 0.6% when unloading tomb.
