MiniCluster: SD Command Timeouts on 10tb SAS Drives.
(Doc ID 2786574.1)
Last updated on JUNE 22, 2021
Applies to:Oracle MiniCluster Specific Software - Version 1.x and later
Information in this document applies to any platform.
The kernel zone ASM grid provides ASFS file system shares mounted via NFS in the global zones and DM and AP Vm's on each node.
The mount points are /sharedstore and /commonfs, and are sourced in the SHARED Disk Group in the Kernel Zone ASM Grid.
The SHARED DG is made up of six 8tb or 10tb disks, depending on when system was ordered.
When 10tb drives are used in the ASM instance, we see command timeouts in the Global Zone resulting in ASM IO delays resulting in ASM dropping disks and having to do a re-balance act.
Any subsequent IO timeouts during re-balance cause more disk drops, which can lead to the SHARED DG being taken offline, which results in stale NFS handles, hung process, etc. which have open file descriptors on /sharedstore or /commonfs mounts.
This documented generated from a customer SR where this problem was first discovered.
-Case was opened for I/O Errors on Shared Device as seen in the ASM alert log:
None required. A new system with 10tb SAS drives will experience this issue under load
To view full details, sign in with your My Oracle Support account.
Don't have a My Oracle Support account? Click to get started!
In this Document