Fritz Borgstedt wrote:
> ASSP will not call clamd if it is not working, but..
> when ASSP is calling clamd and clamd does not give back control, ASSP
> will hang. There is no timeout mechanism to get control back. clamd
> must be supervised by other means - watchdog etc. every 5 minutes or
Ah, good to know. This explains a lot! Would it be possible/viable to
have an internal timer? Perhaps if clamd doesn't return control in a
certain amount of time (5 or 10 minutes?), ASSP will disable clamd
support [, send a warning to the postmaster] and continue?
Watch dog scripts are a feasible work-around, but it seems like a
programmatic control issue exists with ASSP. In the meantime, 1.3.5
seems to run fine with clamd disabled.
On a side note: This seems to be the issue I was experiencing in our
previously off-line communication. I wasn't aware that clamd
*required* a watchdog script like this. And it wasn't until I allowed
ASSP to stay "hung" for an extended period of time that I noticed that
the clamd.exe process in memory grew to (3) times its normal size in memory!
But, also to note, this clamd issue has only appeared in my environment
since ~18.104.22.168. I have no way of knowing if its a coincidence or not -
but I do no believe I had a clamd issue before. So, I'm going to still
see if I can figure out what may have changed - but in the meantime a
watchdog script will get he job done.