Flashes (Alerts)
Abstract
If a SAN Volume Controller node's internal HDD fails, this may lead to VIOS hosts running with the SDDPCM multipathing driver experiencing I/O hangs or timeouts
Content
If a SAN Volume Controller node's internal HDD fails in such a way that it causes one or more of its filesystems to become read-only, this can result in the node entering a state in which its fibre channel ports remain online, but it is no longer participating in the cluster servicing I/O requests.
Under this condition, VIOS hosts running with the SDDPCM multipathing driver have been observed to experience I/O hangs or timeouts.
The following flash describes the VIOS/SDDPCM behaviour in more detail:
http://www-01.ibm.com/support/docview.wss?uid=ssg1S1003753
Fix
This issue has been addressed from a SAN Volume Controller perspective by an improved HDD failure recovery mechanism, which will prevent the failing node from leaving its fibre channel ports online when encountering a HDD failure of this type.
This new mechanism was introduced by APAR IC74194 in the V5.1.0.9 PTF release, available from the following URL:
http://www-01.ibm.com/support/docview.wss?uid=ssg1S4000955
This APAR will also be included in a future V6.1.0.x PTF release.
Was this topic helpful?
Document Information
Modified date:
25 September 2022
UID
ssg1S1003757