From: Herbert Poetzl (herbert_at_13thfloor.at)
Date: Sun 30 May 2004 - 15:36:09 BST
On Sun, May 30, 2004 at 11:32:12AM +0200, Dimitri Roschkowski wrote:
> Herbert Poetzl wrote:
> >yep, for example using the 'nproc' flag and
> >setting the # of processes to 100, should prevent
> >a load higher than 100 ...
> ok, thx
> >well, except from logging the process count
> >in a well defined interval (for example with
> >vserver-stat) into a local or better remote
> >log file could give you sone hints ...
> Well, this are the first lines, vserver-stat give:
> CTX PROC VSZ RSS userTIME sysTIME UPTIME NAME DESCRIPTION
> 0 38 138MB 4kB m05.00 m24.03 55m27.95 root server
> 49152 31 386MB 9kB m27.59 m06.32 50m10.92 v01
> 49153 34 535MB 15kB m31.60 m07.03 50m07.79 v02
> 49154 64 3GB 181kB m44.48 m08.99 50m00.72 v03
> 49155 32 479MB 12kB m26.76 m06.10 49m54.56 v04
v03 for example is already running 64 processes,
which seems a little high for me ...
> What does VSZ and RSS mean? The --help doesn't realy help me
VSZ is the total sum of the virtual memory
(address space) of each process
RSS is the total sum of the resident set size
(in RAM pages) of each process
it's basically the same as in ps auxww
only summed up for all processes belonging
to a context ...
Vserver mailing list