It died with OOM, and has been restarted.
This does *not* appear to be a recurrence of the problem where postgres
invaded the master node.

-Andrew

Reply via email to