Flashes (Alerts)
Abstract
An issue was introduced in V6.1.0.4 of the SAN Volume Controller code, which can result in host write operations failing to be written to disk, while the host that issued the write is informed that the write has completed successfully.
Content
Under high host workload conditions, if a node shutdown or reboot occurs, or a FlashCopy prepare command is issued, a situation can occur in which host writes that have entered the SAN Volume Controller node caches are not successfully written to backend physical disks.
When this occurs, the hosts that have submitted writes will have been informed that the writes had been successfully written at the point at which they had entered the node caches. The hosts will therefore not be aware that the writes have in fact ultimately failed to be committed to backend disks.
This failure to complete writes to physical disks is likely to result in an unreported condition where data is incompletely written to Volumes.
Fix
This issue is resolved in the V6.1.0.6 PTF release, available for download from the following URL:
http://www.ibm.com/support/docview.wss?uid=ssg1S4000954
Was this topic helpful?
Document Information
Modified date:
25 September 2022
UID
ssg1S1003756