Hi all,

At what stage does the writes to the recovery log database become a
bottleneck? Currently we are bench marking our platform and hitting
about 200-300 writes per second. What we have noticed is after
finishing the frontend bench marking (using jmeter) against our
application, their appears to be a backlog of writes continuing to
happen to the recovery log for quite a period after the bench marking
is finished. In this period their appears to only be 1 connection
Established to the local recovery log database. 

We are currently using a local instance of postgresql for the recovery
log, the documentation pointis at HSqldb, is this the preferred database
for the recovery log?

Regards


Stuart
_______________________________________________
Sequoia mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/sequoia

Reply via email to