RE: [SSI-users] Ptrace_unlink: vpop_reclaim failed

[Brown, L. <Larry.Brown@NielsenMedia.com>]

Well I spoke too soon.  It only appeared to solve the problem.  It
persists.  We got the same message again as the server locked up.  The
messages are the same even though we are not running strace the message
generated is from kernel: ptrace_unlink:  Any ideas?  It has happened
while running gdb and as tested this morning it happened without gdb
running.  It is happening in the normal bash shell, not bash-ll.

>Thank you for the response Laura.  After some investigation
> we found that by turning off some of the printf statements
> on the application we were running it no longer exhibited=20
>the problem.  I'm not sure why standard out messages would=20
>cause a problem, but we are pressed for time in the develop
>ment of the app and can not investigate further right now.=20
> By turning the output to stdout off everything is working=20
>smoothly again.=20


Larry Brown
Network Engineer
Nielsen Media Research

-----Original Message-----
From: Laura Ramirez [mailto:lau...@hp...]=20
Sent: Monday, July 18, 2005 5:32 PM
To: Brown, Larry
Cc: ssi...@li...
Subject: Re: [SSI-users] Ptrace_unlink: vpop_reclaim failed


Hi Larry,

What process were you stracing?

ptrace_unlink() is called to stop tracing a process that was being
straced.

Unfortunately, the error message doesnt give too much detail as to which
process it was or what the errno was.
=09
Maybe a ps output might be helpful to see what processes are running.

laura

Brown, Larry wrote:
> I am running openssi-fc2-1.2.2.i686 on Fedora Core 2.  I have 6 nodes=20
> total with one master and the rest booting PXE off the master.  We are

> having a problem where we get thousands of syslog statements to the=20
> effect...
> =20
> ptrace_unlink: vpop_reclaim failed
> ptrfailed
> ptrace_unlink:failed
> =20
> in a cyclical fashion.  Has anyone seen this before?  Any clues on=20
> where to look for the culprit?
> =20
> Larry Brown
> Network Engineer
> Nielsen Media Research
> =20
>=20