Pillar Axiom: Boot From SAN failing

Asset ID:	1-72-1908817.1
Update Date:	2018-01-08
Keywords:

Solution Type Problem Resolution Sure

Solution 1908817.1 : Pillar Axiom: Boot From SAN failing

Applies to:

Pillar Axiom 600 Storage System - Version All Versions to All Versions [Release All Releases]
Information in this document applies to any platform.

Symptoms

Host boot from Axiom failing with 'Error loading /s.v00 Fatal error: 8 (Device error) ' observed on host

Cause

Two different causes identified.:
1: I/O is received down non (slammer CU) owning the boot volume. This causes a path failover between slammer CUs during the host boot which the host cannot recover from.
2: SAN boot volume located on over-driven brick resulting in Task_Set_Full (flow control) to the host from which it cannot recover from.

1: In slammer logs, slammer ownership change recorded at time of host boot.
* scc_SlunUpdateOwnerMsgParse 640 "slunTid 0x160016 give ownership to (0x2009000b08012345)"

2: In slammer logs, Task_Set_Full returned to host during host boot.
scc_CommonBsHandleReqError SAN SAN_SCC "job 0x3afce800 slunTid 0xfec00db BS req 0xab2ef18 returned result code 1"
scc_IoTraceCmdTaskComplete SAN SAN_IO "OP job 0x3afce800 lunChanNexus=0x0000070d statusAscAscqOp=0x28000028 slunTid 0xfec00db"

Solution

1: Strictly follow the host SAN boot operating system guidelines and configuration to ensure that I/O only follows the preferred/owning path.
If necessary, strip back the configuration to bare minimum (single HBA, path) to get working setup then add back SAN/path elements to identify cause.

2: Set QoS on the boot volume to "Premium" to ensure it resides on fastest internal storage on array.
Set host Qdepth settings per firmware README/release notes.

Attachments

This solution has no attachment