IBM Support

Known Issue - False Reports of Memory Errors on 2145-8F2 and 2145-8F4 SVC Hardware

Flash (Alert)


Abstract

A problem exists in some versions of SAN Volume Controller software causing spurious reporting of memory faults on 2145-8F2 and 2145-8F4 hardware platforms. This occurs infrequently and results in the node displaying Node Error 511, 512, 513 or 514

Content

During startup and normal running, SVC nodes constantly test the integrity of the installed memory. A problem has been noted in which certain combinations of SVC hardware and software can falsely indicate a memory fault.

This causes the affected node to halt normal operation, cease functioning as a member of the SVC cluster and display one of the following messages on the front panel:

Node Error: 511

Node Error: 512

Node Error: 513

Node Error: 514

The remaining node(s) in the SVC cluster log a Cluster Error, code 1195.

The problem affects the following models of SVC hardware:

2145-8F2 and 2145-8F4

running software at any of the following levels:

V4.2.1.x (x less than 10)

V4.3.0.x (all versions)

V4.3.1.x (x less than 6)

If an error is encountered on a hardware and software combination which matches the details specified above, the customer is advised that the fault indication is very likely to be spurious and the node should be added back into the cluster as quickly as reasonably possible, using this procedure:

1. Power down the affected node.

2. Power up the affected node with a monitor and keyboard attached and boot into the BIOS.

3. Check the status of the 4 memory banks.

4. If any bank is disabled, enable it.

5. Power down the node.

6. Power up the node and allow time for it to boot.

7. If the node shows either "Node Error: 578" or "Cluster:" on the front panel, proceed to step 8. Otherwise, call IBM support and report a real memory fault.

8. Remove the old instance of the node from the SVC cluster, by issuing the command:

svctask rmnode <node_id>

where <node_id> is the ID of the offline node in the output of svcinfo lsnode

9. Add the node back into the cluster by issuing the command:

svctask addnode -panelname <panel_name> -name <name_of_node> -iogrp <io_group>

This issue has been addressed by APAR IC60803 in the V4.2.1.10 and V4.3.1.6 PTFs


Cross reference information
Segment Product Component Platform Version Edition
Storage Virtualization SAN Volume Controller V4.2.x Platform Independent V4.2.1.x

Document information

More support for: SAN Volume Controller
V4.3.x

Version: 4.2.1.x, 4.3.x

Operating system(s): Platform Independent

Software edition: Enterprise

Reference #: S1003514

Modified date: 01 May 2009


Translate this page: