On Thu, Jun 05, 2008 at 03:56:59PM +0100, Stephen Gran wrote:
> Hi there,
>
> Could you please retest with the packages from volatile and let me know
> if this is still an issue?
seems gone:
$ clamdscan --version
ClamAV 0.93/7369/Thu Jun 5 15:19:11 2008
$ find /usr/ -type f|wc -l
91978
$ time find /usr/ -type f -exec ash -c '[ -r "{}" ] && {
clamdscan --no-summary - < "{}" || echo ErrC=$? ;
}' \;
real 40m15.694s
user 2m18.320s
sys 2m27.890s
got 5 errors writing to the socket, but this is a busy P4 2.4G / SCSI
mailserver; the test kept 4..6 clamd threads busy.
I think the 'resource unavailable' rate is resonable in such stress test,
and looks consistent with the very rare -2 ret.code I see, perhaps 1/1000
msgs which is likely due to clamdscan being called while housekeeping is
going on (error times are always around scheduled restart etc., on rc=-2
a rescan is tried again in a while and usually succeeds, else scanning
falls back to clamscan which succeeds - so far).
If that's still due to an heisenbug, I think I can live with it for now ;)
Pls note this is a backport on Woody.
thanks
--
paolo
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
