Síntomas
A node may unexpectedly reboot and recover on the target appliance, if the I/O pattern of the replicated data has a repeating pattern of IO fluctuating of very high I/Os and then very low I/O.
This issue only affects PowerStore appliances installed with PowerStoreOS 3.0.x where TCP Replication is used between both sites.
Causa
There is a software issue when a host on the source is writing to a replicated volume and the I/O pattern is bursty (very high IOs for short period and then very little and repeats itself).
Resolución
Workaround:
- A node affected by this issue will reboot and replication should recover.
- If you are affected by this issue, upgrade to PowerStoreOS 3.2 or contact Technical Support for assistance.
Resolution:
- This issue is fixed in PowerStoreOS 3.2.