Dave Warren wrote:
In message <[EMAIL PROTECTED]> Per Jessen <[EMAIL PROTECTED]> wrote:
[ ... ]
There is no reason that monitors can't monitor other monitors too, in the software world.
Sure. But there's a setup cost in terms of admin time, and at least some resources being spent regardless of the approach. Because people value different things, they may well make different choices in approach.
I've used freshclam daemonized on dozens of machines without any significant problems, with one exception-- once where it was started when the machine was without external network connectivity (no default route set), in which case it sat there without ever attempting to talk with the outside world again, as far as I could tell.
On the other hand, I find it reasonable that someone might want to put a periodic task such as running freshclam under the aegis of cron or the /etc/periodic mechanism, in which case checking whether the new versions of the databases are sane before installing them seems entirely useful.
I don't think that freshclam needs another watcher to keep it running as a daemon; it's reliable enough that the most normal cause of disruption is human intervention to install the latest version. But if you want to invoke it via init or launchd or DJB's daemontools, by all means. :-)
However, if you have a distributed process-aware monitor like BigBrother or nagios or whatever handy, sure, it's reasonable to add the ClamAV daemons to the list of processes which are expected to be running on hosts.
In the hardware world, an unnoticed overheat will result in the equipment going down, which would trigger whatever monitors that box to report failures.
Equipment makers are getting better about having equipment throttle back to a low-power but still operational mode in the face of a thermal sensor warning; the better ones (ie, later model P3s & the Core/2/Duos) will keep running even if the CPU heatsink is removed or the associated fan completely fails.
Is it self-healing? No -- But not everything can heal itself. Whether the outage is noticed by the users of the equipment is another matter entirely and will depend on your redundancy.
And your monitoring/human-notification system. -- -Chuck _______________________________________________ Help us build a comprehensive ClamAV guide: visit http://wiki.clamav.net http://lurker.clamav.net/list/clamav-users.html
