Database Software

We have a 2 node RAC cluster running on VMware, each node having 2 VCPUs , 8 gig RAM and 2 gig swap space. Earlier the memory was 4 gig and we had seen scenariois where the node was getting evicted due to insufficient resources . So we bumped up the memory to 8 gig. But even now we do see alerts in the log files about insifficient swap space.

PRVF-7573 : Sufficient swap size is not available on node "mwlsvdwcrac002" [Required = 7.7959GB (8174592.0KB) ; Found = 1.9375GB (2031608.0KB)]

Are these just warning or should we be actually setting up the swap space equivalent to the available memory.

The sys admins have different views on the allocated swap. The Oracle documentation says ->

The minimum required RAM is 1 GB, and the minimum required swap space is 1 GB. Oracle recommends that you set swap space to twice the amount of RAM for systems with 2 GB of RAM or less. For systems with 2 GB to 8 GB RAM, use swap space equal to RAM. For systems with over 8 GB RAM, use .75 times the size of RAM.

2> What is the correlation between no of VCPUs and RMAN parallelism. We have 2 VCPUs defined and I am running a RMAN backup with parallelism 4 . Obviously , I am seeing load on the VM and the Cluster health monitor alerts on high CPU usage.

So should I get the no os allocated CPUs increased or reduce the degree of parallelism.

Amith