On 1/10/2014 12:22 PM, J Franz wrote:
From a clp written years ago, our nightly save (while active) is ending tcp, hostservers, many subsystems,
then starting it all back up-
apparently in past had trouble getting a clean save, with locks
the save active did not get past.
We have no interactive users at that time, but may
have web clients jobs (web server is not on i, these are all
sql requests into QUSRWRK.
It has been suggested we stop ending everything,
and just end qusrwrk and qinter, then restart them,
and start save active.
I'm struggling with the same issue of web users on during backups.
Don't I need to end QUSRWRK, start SAVACT(), wait for the checkpoint
message and only then restart QUSRWRK?