> Now we changed from Gentoo based systems (which did not use
> sa > 3.02) to Debian based systems (with 3.03 initially), still using
> the same version/config of exim/exiscan. When used in combination with
> Spamassassin 3.03, we got the said memory problems. Since we downgraded
> to 3.02 yesterda
> Are you limiting the size of msgs that exim is sending to spamd to scan?
>
> For folks using Exim, please see Justin's msg to the users list the
> other day:
> http://mail-archives.apache.org/mod_mbox/spamassassin-users/200505.mbox/[EMAIL
> PROTECTED]
>
> You really need to be limiting the ms
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
David W Smith wrote:
>> --- Spamd is called via an '/etc/init.d' startup script. The
>> server has 2Gb RAM. I'm not sure if I was seeing a memory leak or
>> a highwater mark. I just recall watching the 'free' column of the
>> Swap output of 'watch -n
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Thomas Jacob wrote:
> Exact same problem here.
>
> We upgraded our spam scanners to Debian/Sarge during the last days,
> with a manual out-of-the-box installation of spamassassin 3.03, but
> since around after 10:00 CEST (08:00 GMT) today (2005-06-01)
On Wed, 1 Jun 2005, Michael Parker wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
David W Smith wrote:
So I reverted to SA 3.0.2 and the memory leak ceased rapidly plus
the load on the server normalized.
The SA 3.0.3 install had both the URIDNSBL and the SPF module
enabled (the enablin
> I was wondering if anyone had noticed a possible memory leak with SA
3.0.3
> (under Linux/Debian)?
> I upgraded SA from 3.0.2 to 3.0.3 on a Debian server about 7 days ago
and today
> twice the load sky-rocketed followed by the server seizing up. In the
logs were
> lots of 'kernel: __alloc_pag
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
David W Smith wrote:
> So I reverted to SA 3.0.2 and the memory leak ceased rapidly plus
> the load on the server normalized.
>
> The SA 3.0.3 install had both the URIDNSBL and the SPF module
> enabled (the enabling of the latter was new (on my server
Hi,
I was wondering if anyone had noticed a possible memory leak with SA 3.0.3
(under Linux/Debian)?
I upgraded SA from 3.0.2 to 3.0.3 on a Debian server about 7 days ago and today
twice the load sky-rocketed followed by the server seizing up. In the logs were
lots of 'kernel: __alloc_pages:
