On Fri, 2005-05-27 at 14:49 -0400, Jason Frisvold wrote: > On 5/27/05, Samuel Benzaquen <[EMAIL PROTECTED]> wrote: > > If the process is trying some I/O to some not-available, hard-mounted NFS > > filesystem, then the process will not die with 'kill -9'. > > > > Just a thought. > > I use NFS to mount the user mail directories, but clam should never > touch that. The mail comes in, simscan places it into > /var/qmail/simscan and then clamdscan is run on that to detect virii. > Once completed, control passes back to simscan... If it should lock up again and you would happen to running Solaris, I would do a pstack <pid> > <some-file> to see what all threads within the process were at or maybe a few other p-cli like pfile. If you're on Linux, which you properly are, I wouldn't be sure if it got similar tools, but maybe others on this list known how to dump the internal of a running process?
Otherwise try Google, some Linux lists or dig into the sysfs or procfs. -- Steffen Winther Soerensen <[EMAIL PROTECTED]> private luser _______________________________________________ http://lurker.clamav.net/list/clamav-users.html
