EVA Controller Reboot Instructions

[Kites BM]

Hi,

We had a problem during firmware upgrade of EVA 4000/6000 system.

I'm looking for controller reboot instructions (hopefully without downtime).

Any documents or referals will do.

Adv Thanks ,
Kites

[Torsten.]

>> We had a problem during firmware upgrade ...

Details?

[P Muralidhar Kini]

Hi Kites,

>> We had a problem during firmware upgrade of EVA 4000/6000 system.
What is the problem ?

To restart the EVA controller you need to use the SSSU utility.
HP StorageWorks Storage System Scripting Utility is a command line interface
that allows you to configure and control EVA arrays.

Check the attachment for SSSU commands
"HP StorageWorks Storage System Scripting Utility reference"

To download SSSU,
http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=us&prodTypeId=12169&prodSeriesId=471497&prodNameId=471498&swEnvOID=181&swLang=13&mode=2&taskId=135&swItem=co-30955-1

Hope this helps.

Regards,
Murali

[Kites BM]

HP mistakenly informed us that a recent EVA firmware upgrade could be done on-line which caused an extensive outage. I think that it was the controller restart portion of the upgrade that caused the problem.

Now they suggest that master controller needs to be restarted to fix the problem(after I sent all the logs to them)

We want to avoid the downtime as all the hosts connected boot from SAN.

whether the reboot is possible without outage??? I'm looking for guides to build a plan for this activity (need client approval for downtime if in case):(

Kites

[Víctor Cespón]

If both controllers are online, you can reboot one and all the vdisks will move to the other.

In Command View select the EVA - shutdown - restart - Controller A/B.

HP support should be able to tell you which one is the master controller.

[Jan Soska]

Hello,
what firmware do you try upgrade? We have Eva6000, updated firmware ~3 times, no problem, no downtime.

Jan

[Kites BM]

I understand that controller reboot requires min IO and that its onlline activity.

We have an ITIL process to follow and present procedural steps to the client to seek approvals for change management.

Any docs or guide for reboot of controllers would really help.

Thanks again,
Kites

[Víctor Cespón]

After firmware update the controllers perfom a resync. This is a fast reboot and should not take more than one minute. It's proportional to the number of disks and vdisks on the EVA.

Check the section:

Managing host I/O timeouts for an online upgrade

On the firmware update guide.

http://bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c02104123/c02104123.pdf

We have had problems with Solaris hosts, they timeout sooner than others. In any case you can change the disk timeout from 60 to 120 seconds on your hosts.

[gregersenj]

You must understand what on-line mean, at least in HP terms.

Off-line = System must be shutdown and upgraded.

On-line = You upgrade the system running, but a reboot is required to activate new FW.

Rolling = You upgrade on the system running, "no reboot" required.

I read that somewhere in an EVA manual, I think it was the upgrade manual.

Beware on the EVA. The EVA will reboot when the FW has been flashed. No futher warnings no way to delay the reboot. When you have entered YES for the second time, You're no longer in control.

BR
/jag

[Jeff_Traigle]

Also be aware that there is an advisory out on XCS 6.220 from October and November 2009. Although the alert makes the problem seem fairly benign, the problem can cause an outage if a disk fails in a certain way. Supposedly, HP has only seen an outage occur three times world-wide due to this. Given that both EVA 6000s we upgraded were diagnosed as having this issue, I suspect the actual number of outages is much greater and the incidents either 1.) weren't reported to HP, 2.) HP did not properly diagnose the root cause, or 3.) HP isn't being very forthcoming in acknowledging the number of times customers have really faced outages resulting from this issue. The "sure way" to correct the problem is to power cycle the EVA. So we've been told. Given problems I've seen the past several months and seeing other people's similar issues of having a single disk failure cause a supposedly "one in a million" outage going back several years on different models, HP's credibility with the EVA line is nearly non-existent in my book these days.

http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=c01914371&dimid=1006920700&dicid=alr_oct09&jumpid=em_alerts/us/oct09/all/xbu/emailsubid/mrm/mcc/loc/rbu_category/alerts

[Jan Soska]

Hello,
we had the same logs a year ago (after firmware upgrade).
I if remeber, they disapeared after controller reboot, so nothing disastrous.

I noticed much bigger firmware issues in newer 4400/6400/8400 line requiring power off whole array, stats bad.. But as I know, it is solved now.

Jan