We're currently using 2.1.3 release for production work. Periodically
(every few days) at apparently random times, the fbserver process
becomes unresponsive and will sometimes crash. The guardian service
then fails to restart the service. There is rarely any specific error
mentioned (although we did see one regarding unable to allocate slot
that appeared to relate to a bug that was fixed with the advent of 2.1.2).

The main database we are using is around 40Gb. It is being served from
a win2k3 server using a professional DAS array. It is accessed by
upwards of 50 people at once.

Without a way to reproduce this reliably or with specific error messages
it is very hard to find a way to begin to assess what is causing this.
At the moment the only option I can see if to upgrade to the snapshots
and see if the bleeding edge includes fixes for the as yet undetermined
issue.

Can anyone advance where we should begin looking?

Karl Geppert