*********** REPLY SEPARATOR ***********
On 4/13/2005 at 5:12 PM [EMAIL PROTECTED] wrote: >Carl Thompson wrote: >> I believe the problem is a hard code issue. The setting --server >> isn't telling clamav-milter to listen on a specified IP for >> connections its telling clamav-milter to connect to clamd on the >> specified IP thus making the INSTALL documentation in error (unless >> i'm seriously confused) >> >> Sendmail must connect to clamav-milter not clamd directly but >> clamav-milter will not listen on a TCP port it will only connect to >> clamd on a TCP port. >> >> Carl > >OH... > >You're correct, clamav-milter won't listen on a TCP port, only on a local >socket. This is a limitation of clamav-milter. > >You could probably write a simple daemon to listen on a TCP port and then >relay all commands it hears to the socket, and all responses back to the >TCP port. > >But why? Why not have sendmail and clamav-milter on one machine, and tell >clamav-milter to communicate with clamd on the other machine? > If clamav-milter is nothing but a portal so to speak and doesn't cause much load thats fine but I thought clamav-milter now has clamd inherant in it (no longer needing to communicate with clamd). This would increase the load of clamav-milter because it is now performing the task of both clamd and clamav-milter. Basically I have a server thats loading down more than I would like and another server that is barely used. The other server doesn't have the redundancy and capacity of the primary server but if I could offload some of the resource usage to the secondary machine by having it perform all the scans then that would help my plight a little bit while I work on hardware upgrades. Also why does the INSTALL documentation imply you can connect to a remote clamav-milter if clamav-milter itself doesn't support a TCP socket? Possibly something that was simply overlooked when clamd tasks were integrated into clamav-milter in .82 As a coder myself I know how easy it is to overlook something like that when doing a merger of processes of sorts. Carl _______________________________________________ http://lurker.clamav.net/list/clamav-users.html
