Out of logical swap space / UMEM?
From: Erik Baigar (erik_at_baigar.de)
Date: 11/04/05
- Previous message: Peter Ilieve: "Re: Restore from tape"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Fri, 04 Nov 2005 16:14:40 +0100
Hi together,
on my O2 (R12k/400, 640M, 6.5.22m) I updated my nessus security
tool to the latest version (2.2.5). In running this I encounter the
following message in the SYSLOG:
Nov 4 15:22:47 1A:unsu unix: ALERT: nessusd [4762] -
out of logical swap space during fork/sproc - see swap(1M)
The nessus consists of a daemon, nessusd which
forks subprocesses for each host it has to scan.
Investigation shows, that the message disappears and
everything works fine if I tell nessus not scan more
hosts simultaeneously. With the old version of nessus
(2.0.5), which worked similar, everything was fine even
in scanning 10 hosts simultaneously.
According to "man swap" I checked the swap space via
"swap -l" during execution and for safety I added another
400M of swapspace. Nothing helped - the message persists.
There are even 300M of real memory free when the message
apperars and 0% of swap is used...
OK, I could restrict myself to single host scanning but
I am curious what the reason might be! The Adminstration
handbook from SGI mentions a variable UMEM, but without
further explanation. Maybe there is a kernel-variable
preventing fork from allocating more memory for a
process and it's childs beyond a certain amount? Where
can this be changed?
Any hints are welcome,
Erik.
- Previous message: Peter Ilieve: "Re: Restore from tape"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
|
