For few Solaris versions, a high CPU usage is reported impacting production database clusters when Infoscale VxVM is configured.
There are instances of free memory depletion and high paging (vmstat) to disk possibly caused due to vxconfigd
1) Locate the pid for vxconfigd
# ps -ef |grep vxconfigd
2) Issue pmap on the pid for vxconfigd and verify if the the value for "Total" changes
# pmap -x |egrep "Kbytes|total"
# pmap -x 53 |egrep "Kbytes|total"
Address Kbytes RSS Anon Locked Mode Mapped File
total Kb 23112 22424 13864 -
#
3) Repeat this command over the course of 5 or 10 minutes
If the Kbytes total steadily increases, this is an indication that vxconfigd is not cleaning up used memory; this is memory leak
To determine which vxconfigd component is responsible, please refer How to troubleshoot a suspected memory leak in the vxconfigd process on Solaris
On Solaris servers, when using Infoscale Volume Manager (VxVM), high memory usage is reported for vxconfigd