There is definitely a significant mem leak when the databases are reloaded. It might be specific to hdb and ndb files (which are delimited text).
With the sanesecurity and msrbl databases, in addition to the two which freshclam updates, I get about 140M VM usage at startup on an amd64 box. That balloons each reload until it uses all swap and crashes. It does not seem to ballon as much when using clamd via amavis as it does when using clamav-milter. The milter may also leak while scanning. It is hard to tell. I do not have a test box to use for a valgrind run, so I cannot give any better details. I currently run lenny and amd64; the leak was also visible on x86. -JimC -- James Cloos <[EMAIL PROTECTED]> OpenPGP: 1024D/ED7DAEA6 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]