On Thu, 21 Jun 2012, Matthijs Mekking wrote:
I have now implemented I think a reasonable strategy in trunk, r6450.
If this happens, the signer will fail the sign task, back off the task
and reload the engine. This will stop workers and drudgers, reopen the
connection to the HSM and start workers and drudgers again. The next
time, the backed off task runs (or another task that is scheduled
earlier), the drudgers try to create a new context and do the sign jobs.
Let me know if you are satisfied with this approach.
If it avoids the crash I'm happy, though it still does not resolve our
issue of the large amounts of RAM needed, but since that was happening
for us on a test deployment without optout (and with less RAM) we won't
be hitting this issue again, though other people might.
Paul
_______________________________________________
Opendnssec-user mailing list
[email protected]
https://lists.opendnssec.org/mailman/listinfo/opendnssec-user
