Low polling_interval ( Multipath ) cause high CPU activity / system hang with high number of LUNs while performing Storage failover test (Doc ID 1684696.1)

Last updated on AUGUST 04, 2018

Applies to:

Symptoms

Customer was performing failover test for his 400 LUns ( 1600 paths in total ), two paths were bring down while two still remains active ( 4 paths in total )

While test was performed and paths were restored after failover test Server became unresponsive and large CPU activity was observed, after 30 minutes Server hang and to restore functionality reboot was required.

Information from top command while failover was performed:

Changes

 No changes were made under OS

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.