Can't get clamd working???

[Rick Cooper]

> -----Original Message-----
> From: mailscanner-bounces at lists.mailscanner.info 
> [mailto:mailscanner-bounces at lists.mailscanner.info] On Behalf 
> Of Julian Field
> Sent: Wednesday, June 06, 2007 7:13 AM
> To: MailScanner discussion
> Subject: Re: Can't get clamd working???
> 
> Attached is a new SweepViruses.pm (gzip-ed) which will make
>     MailScanner --lint
> detect the clamd scanner.
> 
> I am not at all sure we should be checking for the 
> /var/lock/subsys/clamd file at all. This is a very "Linuxy" 
> feature, I'm 
> inclined to take it out or default it to be blank (has the 
> same effect).
> 
> What do you think?
> 

It the larger scheme of things it's doubtful that checking for the
existance, even on a system that is using the lock file, will save much. The
thought was to check for the lock file and short circuit the sockets part if
the lock file was missing and supported. I have the ping timeout at 90
seconds so the most it would save would be 90 seconds per message (of course
that's a lot in a moderate batch) if clamd was shutdown properly and on
purpose. If it died or was killed (-9) the lock file would likely be there
anyway. My feeling is default it to blank and the systems that already use
it can use it.

Another thing from yesterday is the threading before 0.90. I assume you want
to mention the minimum release to use threading in the config file, do you
want to add a version check, warning, and switch to CONTSCAN if someone
shoots themselves in the foot? Pretty easy, switch the PING section to
SESSION (so only one connection) and after a PONG send VERSION, which
returns the current clamd version, database version and last update. If
ScanType is MULTISCAN and version is < 0.90 warn the operator and change the
ScanType to CONTSCAN. Would add probably 25 30 lines of code.

Rick


--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.