EVA 4100 controller mastership

[Rik-Bel]

Hi,

We have in our EVA4100 the continuous change of controller mastership, day in day out, every X hours; is it caused by the battery?
see logs, at the end battery logging

Rik

07:34:38:865
8-Aug-2008
Controller A 0310001f #3629
A Storage System Virtual Disk has changed controller mastership.
Corrective action code: 00 More details

07:34:38:391
8-Aug-2008
Controller A 4207001b #3628
Indicated Virtual Disk has transitioned to ownership by the other HSV200 controller.
Corrective action code: 00 More details

04:34:24:987
8-Aug-2008
Controller A 04130003 #3627
Storage System event reporting is still active.
Corrective action code: 00 More details

16:34:24:564
7-Aug-2008
Controller A 04130003 #3626
Storage System event reporting is still active.
Corrective action code: 00 More details

04:34:24:069
7-Aug-2008
Controller A 0310001f #3625
A Storage System Virtual Disk has changed controller mastership.
Corrective action code: 00 More details

04:34:23:455
7-Aug-2008
Controller A 4207001b #3624
Indicated Virtual Disk has transitioned to ownership by the other HSV200 controller.
Corrective action code: 00 More details

23:34:18:920
6-Aug-2008
Controller A 0310001f #3623
A Storage System Virtual Disk has changed controller mastership.
Corrective action code: 00 More details

23:34:05:939
6-Aug-2008
Controller B 4207001b #3622
Indicated Virtual Disk has transitioned to ownership by the other HSV200 controller.
Corrective action code: 00 More details

20:34:14:839
6-Aug-2008
Controller A 0310001f #3621
A Storage System Virtual Disk has changed controller mastership.
Corrective action code: 00 More details

20:34:01:850
6-Aug-2008
Controller B 4207001b #3620
Indicated Virtual Disk has transitioned to ownership by the other HSV200 controller.
Corrective action code: 00 More details

12:33:55:803
6-Aug-2008
Controller A 04130003 #3619
Storage System event reporting is still active.
Corrective action code: 00 More details

00:33:55:338
6-Aug-2008
Controller A 0310001f #3618
A Storage System Virtual Disk has changed controller mastership.
Corrective action code: 00 More details

00:33:54:475
6-Aug-2008
Controller A 4207001b #3617
Indicated Virtual Disk has transitioned to ownership by the other HSV200 controller.
Corrective action code: 00 More details

13:33:42:706
5-Aug-2008
Controller A 0310001f #3616
A Storage System Virtual Disk has changed controller mastership.
Corrective action code: 00 More details

13:33:42:098
5-Aug-2008
Controller A 4207001b #3615
Indicated Virtual Disk has transitioned to ownership by the other HSV200 controller.
Corrective action code: 00 More details

03:33:33:433
5-Aug-2008
Controller A 0310001f #3614
A Storage System Virtual Disk has changed controller mastership.
Corrective action code: 00 More details

03:33:20:616
5-Aug-2008
Controller B 4207001b #3613
Indicated Virtual Disk has transitioned to ownership by the other HSV200 controller.
Corrective action code: 00 More details

17:33:10:007
4-Aug-2008
Controller A 04130003 #3612
Storage System event reporting is still active.
Corrective action code: 00 More details

05:33:09:574
4-Aug-2008
Controller A 0310001f #3611
A Storage System Virtual Disk has changed controller mastership.
Corrective action code: 00 More details

05:32:56:943
4-Aug-2008
Controller B 4207001b #3610
Indicated Virtual Disk has transitioned to ownership by the other HSV200 controller.
Corrective action code: 00 More details

02:52:31:526
4-Aug-2008
Controller A 0e030019 #3609
The status of battery assembly '0' has changed.
Corrective action code: 00 More details

02:47:12:187
4-Aug-2008
Controller A 0e030019 #3608
The status of battery assembly '0' has changed.
Corrective action code: 00 More details

02:33:06:635
4-Aug-2008
Controller A 0310001f #3607
A Storage System Virtual Disk has changed controller mastership.
Corrective action code: 00 More details

[Uwe Zessin]

What operating systems are involved? It sounds like you have a non-consistent I/O access - maybe a different cabeling and/or non-optimized multipath setting.

[Rik-Bel]

Hi,

We have 2 W2k3 and 4 ESX 3.0x, how start troubleshoot this ? The status of the battery assy ?

Rik

[Uwe Zessin]

> The status of the battery assy ?

A CAC value of 0 means that no action is necessary. It is just an informational message.


You will have to do a traffic analysis to find out which server(s) are causing the ownership transfers.

[Rik-Bel]

How can this be done ?

[Uwe Zessin]

You need full documentation about the infrastructure (e.g. which adapter port goes to which switch port and has connectivity to which controller ports).

From there you need to find out which hosts are causing traffic to the virtual disks which results in ownership transfers. The tools are the Fibre Channel switch performance monitor and the EVAperf utility.

[Rik-Bel]

Thanks, I'll do this.

[Rik-Bel]

Hi,

The eva perf let me see the usage and statistics but why it will change mastership is not obvious, the usage i divided over the 2 contollers because i divided the lun presentation over the 2 controllers manually.

Rik

[Paul .]

More than likely, your VMFS pathing is not configured correctly.

With EVA, the default is FIXED, with a single ACTIVE path.

Make sure the PREFERRED ACTIVE path of each VMFS volume, ON EACH HOST, is configured to use the same storage controller!

Multiple ESX hosts access the same VMFS volumes. So, if you have multiple VMs on a VMFS volume, and the VMs are running on a different ESX hosts, and the VMFS volume is configured with different paths on your hosts, each time a read/write operation is performed the Vdisk will transfer ownership.

For each VMFS on each host, choose the same preferred path.

[georgemad]

Very late Comment:

Could this be a scheduled battery load test log ?