[EMAIL PROTECTED] wrote: > i! > > Is it possible that VSZ becomes smaller/lower than RSS in 'vserver-stat' > after starting some applications (JBoss and Tomcat with Active-BPEL > engine, see #1)?
In order for a page to be in RAM, it would have to mapped, I'm quite sure... > ------------------------------------------------------------------------ > ~ # vserver-stat > CTX PROC VSZ RSS userTIME sysTIME UPTIME NAME > 0 58 296.4M 22.3M 0m02s42 0m04s55 6m33s88 root server > 1151 5 1.3G 283.4M 0m38s45 0m00s82 4m41s19 lvg-1-151 > 1152 16 393.6M 793.8M 1m14s76 0m01s93 4m36s71 lvg-1-152 <-- #2 > 1153 7 1.3G 195.7M 0m19s50 0m00s62 4m31s70 lvg-1-153 <-- #1 > 1154 3 42.2M 2.7M 0m00s00 0m00s00 4m21s36 lvg-1-154 You're using a vserver-stat which just sums up the values in /proc. This can break in any number of ways, like e.g. overflowing. You could try using a 2.2.0-rc7+ kernel with util-vserver-0.30.213-pre1+ to get vserver-stat to look at the values in /proc/virtual/<xid>/limit, which should be correct (although I suppose they can still overflow). -- Daniel Hokka Zakrisson _______________________________________________ Vserver mailing list Vserver@list.linux-vserver.org http://list.linux-vserver.org/mailman/listinfo/vserver
