| *SOLVED* Unrealistic privvmpages usage [message #11226] |
Thu, 15 March 2007 16:34  |
mephisto
Messages: 34
Registered: February 2006
|
Member |
|
|
Hi,
I'm running F-Secure Policy Manager 6.0 inside a CentOS VE and have also tried this with FSPM 7.0 inside a Debian 3.1 VE.
In both cases the F-Secure Policy Manager Server seems to consume extreme amounts of memory (as seen in user_beancounters), while this is not reflected on the hardware node (running free -m). FSPM seems to be based on Apache, so this might be something that can be seen with Apache only, but my Apache installations don't show this problem. Here are the details:
FSPMS not running:
Inside the VE:
Version: 2.5
uid resource held maxheld barrier limit failcnt
104: kmemsize 772084 2515183 33556806 36875611 0
lockedpages 0 0 4490 4490 0
privvmpages 3522 152370 279000 300000 0
shmpages 0 299 23293 23293 0
dummy 0 0 0 0 0
numproc 11 75 4000 4000 0
physpages 1031 3557 0 2147483647 0
vmguarpages 0 0 125952 2147483647 0
oomguarpages 1031 3557 125952 2147483647 0
numtcpsock 1 10 4000 4000 0
numflock 1 19 1000 1100 0
numpty 1 1 400 400 0
numsiginfo 0 5 1024 1024 0
tcpsndbuf 8944 121832 14272375 30656375 0
tcprcvbuf 16384 180600 14272375 30656375 0
othersockbuf 11180 19560 7136187 23520187 0
dgramrcvbuf 0 1212 7136187 7136187 0
numothersock 7 10 4000 4000 0
dcachesize 0 0 20078353 20680704 0
numfile 220 1349 35904 35904 0
dummy 0 0 0 0 0
dummy 0 0 0 0 0
dummy 0 0 0 0 0
numiptent 4 4 200 200 0
On the hardware node:
total used free shared buffers cached
Mem: 1516 1452 63 0 55 746
-/+ buffers/cache: 649 866
Swap: 972 5 966
FSPMS running:
Inside the VE:
Version: 2.5
uid resource held maxheld barrier limit failcnt
104: kmemsize 2004664 2515183 33556806 36875611 0
lockedpages 0 0 4490 4490 0
privvmpages 148100 152370 279000 300000 0
shmpages 299 299 23293 23293 0
dummy 0 0 0 0 0
numproc 69 75 4000 4000 0
physpages 2864 3557 0 2147483647 0
vmguarpages 0 0 125952 2147483647 0
oomguarpages 2864 3557 125952 2147483647 0
numtcpsock 7 10 4000 4000 0
numflock 17 19 1000 1100 0
numpty 1 1 400 400 0
numsiginfo 0 5 1024 1024 0
tcpsndbuf 62608 121832 14272375 30656375 0
tcprcvbuf 114688 180600 14272375 30656375 0
othersockbuf 6708 19560 7136187 23520187 0
dgramrcvbuf 0 1212 7136187 7136187 0
numothersock 5 10 4000 4000 0
dcachesize 0 0 20078353 20680704 0
numfile 1247 1349 35904 35904 0
dummy 0 0 0 0 0
dummy 0 0 0 0 0
dummy 0 0 0 0 0
numiptent 4 4 200 200 0
On the hardware node:
total used free shared buffers cached
Mem: 1516 1461 54 0 56 747
-/+ buffers/cache: 658 858
Swap: 972 5 966
How can this be explained? How could it be resolved?
Best Regards,
Frederik
[Updated on: Tue, 20 March 2007 06:56] by Moderator Report message to a moderator |
|
|
|
|
|
|
|
OpenVZ Forum
Members
Search
Help
Register
Login
Home
