PowerStore: Unexpected Node reboot can occur after an excessive number of Host ABORT TASK commands
Summary: If Hosts generate large numbers of ABORT TASK requests to the PowerStore appliance, it can lead to an unexpected Node reboot.
Symptoms
An unexpected node reboot may be observed when the appliance receives many ABORT TASK requests from hosts.
The node should reboot and automatically recover.
Cause
There is a software issue in the SCSI Target Subsystem (SCST) on the PowerStore node that may fail to properly release queued commands after processing many ABORT TASK requests, causing the kernel task to stop responding and may result in a node reboot.
This symptom tends to occur when one or more connected host initiators send many ABORT TASKs due to environmental reasons.
Resolution
The issue is fixed in PowerStoreOS 4.0.0.
Workaround
Upgrade to PowerStoreOS 4.0.0 and above. If NDU is not possible, resolving excessive ABORT TASKs from host initiators may help prevent recurrence.
If the issue persists, contact your Service Provider for assistance.