This is a follow-on to flash S1003579, noting additional changes made to prevent permanent application failures from occurring, when a AIX or VIOS server receives a SCSI error condition from a SVC cluster of SCSI Check Condition, Not Ready, Logical Unit not accessible, Target port in Unavailable State. These errors will be logged as a SC_DISK_ERR2 error in the AIX or VIOS server error log. The scope of this flash is expanded to all AIX and VIOS SDDPCM servers and clients (not just NPIV clients).
There is a follow-on flash available, recommended SDDPCM level is 126.96.36.199 or higher
Contents of this flash:
Permanent application failures may occur when an AIX based partition (running AIX, or VIOS) that is a host to a SVC cluster, receives a SCSI error condition (SCSI Check Condition, Not Ready, Logical Unit not accessible, Target port in Unavailable State) from the SVC cluster, if the following AIX APARS are installed along with any level of SDDPCM older then 188.8.131.52:
This exposure, although rare, exists in any AIX or VIOS partition, using SDDPCM and with the AIX APARS listed in the following table installed.
This issue has been primarily encountered by VIOS client partitions that are using NPIV. However it can also occur with any AIX based partition (stand alone AIX, VIOS and VIOS client using VSCSI).
The changes made for VIOS client hangs in technote #ssg1S1003579 require additional AIX driver and SDDPCM code updates for the following condition:
An SVC error is logged as an SC_DISK_ERR2 in the VIOS error log and translated:
SCSI Status 0x2 - Check Condition
Sense Key 0x2 - Not Ready
Code/Qualifier 0x4/0xC - Logical Unit Not Accessible, Target Port In Unavailable State
This condition may be present for a short time when performing a firmware upgrade of the SVC cluster, or when a SVC cluster node resets (due to a node assert, node failure or intentional restart of a node), and other potential causes not mentioned here.
The recoverable errors may be logged as SC_DISK_ERR2 and SC_DISK_ERR5 logged on the VIOS, and SC_DISK_ERR1, SC_DISK_ERR2, and SC_DISK_ERR7's logged on the AIX VIO client. These normally recoverable errors may cause a permanent LVDD or application failure without the matching SDDPCM code change.
The following AIX APARs will expose this issue. If any of these APARs (or their subsequent supersedes) are installed on a partition running SDDPCM, then SDDPCM must be at 184.108.40.206 or higher to eliminate the exposure.
53TL11 IZ62309 or IZ63808 (IZ63808 replaces IZ62309)
53TL12 (IZ63809 included in base)
61TL4 IZ62301 or IZ63813 (IZ63813 replaces IZ62301)
61TL5 (IZ63814 included in base)
VIOS 2.1 FP22 or higher
VIOS 1.5 FP11.1 SP01 or higher
Internal Use Only
cmvc sw 4748
NPIV clients of SDDPCM hosts may experience permanent application errors during SVC concurrent code upgrade or node reset with certain APARS and SDDPCM versions. The risk, although rare, exists in any AIX SDDPCM host or client.
17 June 2018