[cisco-voip] Cisco Revoking Drive Pull Support!

Tue Feb 27 09:31:25 EST 2007

In practice we have found that if a "CHKDSK /F" is completed prior to
breaking a mirror the failure rate has dropped to zero. I have posted my
procedure for the HP servers before but I will do it again for anyone
that wishes to have a successful mirror for failback/disaster recovery.:

ALWAYS START WITH THE PUBLISHER AND 
COMPLETE ONE SERVER AT A TIME

For MCS-7845 CallManager Servers

On MCS-7845 servers there are two separate RAID arrays: two disks in
slots 0 and 1 (RAID Array A) and two disks in slots 2 and 3 (RAID Array
B).

Starting with the Publisher and then repeating the steps on the
Subscribers carry out the following steps.  With these steps you check
the Fall Back and the upgrade drives before continuing, this is to
ensure your fallback scenario is good.

*	Label up each drive with the server name AND drive slot number
they are mounted in.
*	"CHKDSK /F" command on all logical
*	Shut down the server cleanly.
*	Partially remove one drive from each RAID Array (We will refer
to these drives as 0 and 2), leaving them half in their drive slots.
*	Start up server and ensure when prompted for HDD recovery F2 is
selected.
*	Confirm that windows boots fully.
*	Shut down the server cleanly.
*	Reinsert the original drives 0 and 2. Do not use replacement
disks at this time.
*	Completely remove from each RAID Array, the drives that were
just tested (We will refer to these as drives as 1 and 3) ensuring you
have label each disk with the slot number it was just removed from
(slots 1 or 3) AND the server name.
*	Start up server and ensure when prompted for HDD recovery F2 is
selected.
*	Confirm windows boots fully.
*	Insert the new replacement disks into drive slot 1 and drive
slot 3.
*	Ship the two removed disks to the address given below using the
packaging the new disks were shipped in.

OBTAINING GOOD MIRROR DISKS

For MCS-7835 CallManager Servers

On MCS-7835 servers there is only one RAID array: two disks in slots 0
and 1 (RAID Array A).

Starting with the Publisher and then repeating the steps on the
Subscribers carry out the following steps.  With these steps you check
the Fall Back and the upgrade drives before continuing, this is to
ensure your fallback scenario is good.

*	Label up each drive with the server name AND drive slot number
they are mounted in.
*	"CHKDSK /F" command on all logical
*	Shut down the server cleanly.
*	Partially remove the lower drive (we will refer to this drive as
drive 0) from the RAID Array, leaving it half in its drive slot.
*	Start up server and ensure when prompted for HDD recovery F2 is
selected.
*	Confirm that windows boots fully.
*	Shut down the server cleanly.
*	Reinsert the original drive 0. Do not use replacement disks at
this time.
*	Completely remove from the RAID Array, the drive that was just
tested (We will refer to this drive as drive 1) ensuring you have label
the disk with the slot number it was just removed from (slots 1) AND the
server name.
*	Start up server and ensure when prompted for HDD recovery F2 is
selected.
*	Confirm windows boots fully.
*	Insert the new replacement disks into drive slot 1.
*	Ship the two removed disks to the address given below using the
packaging the new disks were shipped in.

Thanks,
Mike Lay 

-----Original Message-----
From: cisco-voip-bounces at puck.nether.net
[mailto:cisco-voip-bounces at puck.nether.net] On Behalf Of Ryan Ratliff
(rratliff)
Sent: Tuesday, February 27, 2007 8:15 AM
To: Ed Leatherman
Cc: cisco-voip at puck.nether.net; Matt Slaga (US); Chris Young (chriyoun)
Subject: Re: [cisco-voip] Cisco Revoking Drive Pull Support!

HP has told us for a long time that they did not support breaking
mirrors for backup purposes.  The assumption that the drive you pulled
was already bad so if data is lost from it then it's not a big deal.  I
thought they had recently changed their support policy on this but
perhaps I was wrong.

-Ryan

On Feb 27, 2007, at 8:47 AM, Ed Leatherman wrote:

So if you pull a drive and the array doesnt think its bad then you're
screwed if it doesnt re-build? That seems pretty ridiculous for
something that supposed to be built for fault tolerance etc.

Is this just HP arbitrarily deciding not to support something because
their RAID controller might not work as expected? That gives me warm and
fuzzies...

On 2/27/07, Matt Slaga (US) <Matt.Slaga at us.didata.com> wrote:
For those of you who swap hard drives before upgrades, beware!  Cisco
has changed their stance and no longer supports this.

Internal Cisco conversation:

Hi,

I have a client that was upgrading their CallManager server this
weekend.  They powered down the server and pulled a drive.  They then
powered the server back up and tried to mirror a new drive - pretty
standard practice for backup and recovery on CallManager servers.

Unfortunately there was a problem and the drive wouldn't mirror.

Cisco's TAC engineer got an HP engineer on the phone to help
troubleshoot and the HP engineer told the customer and TAC that pulling
a drive and mirroring a new drive was not supported.  And that the array
only allows you to pull a drive it the array thinks the drive has
failed.  Again this was done on a server that had been powered down.

Needless to say our client was not happy since this is a pretty standard
practice and one that Cisco has even documented.  I need to understand
why the HP engineer said we should not be doing this.  As an ex partner
- I did this on a regular basis and never had issues.

If this is something that should not be done then we have a bigger
problem since a lot of our documentation references pulling a drive as a
backup procedure.  Any help and guidance would be appreciated since we
have to respond back to the client today.

Anita Dulude

Consulting Systems Engineer

Cisco Systems, Inc

Response:

Anita,

It is not supported as a method to backup a system.  Please provide the
links where it is documented and I will get someone from CallManager
product management to revise the document.

Dave English

Product Manager - MCS Platforms

Cisco

IP Communications Business Unit

Response:

I wanted to follow up from the links I sent you to see if the
documentation is correct or not.  This is critical since there are a lot
of engineers and clients following this procedure and if this is wrong -
a lot of documents will need to be updated.  I found Meetingplace
documents that document pulling drives also.  So this is definitely
beyond just CallManager.

Anita Dulude

Consulting Systems Engineer

Cisco Systems, Inc

Response:

Anita,

It is not supported.  The documents are wrong and should be corrected.
I have asked Wade Hamblin of CallManager Product Management to have the
docs revised.

Dave English

Product Manager - MCS Platforms

Cisco

IP Communications Business Unit

Disclaimer: This e-mail communication and any attachments may contain
confidential and privileged information and is for use by the designated
addressee(s) named above only. If you are not the intended addressee,
you are hereby notified that you have received this communication in
error and that any use or reproduction of this email or its contents is
strictly prohibited and may be unlawful. If you have received this
communication in error, please notify us immediately by replying to this
message and deleting it from your computer. Thank you.

_______________________________________________
cisco-voip mailing list
cisco-voip at puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-voip

--
Ed Leatherman
Senior Voice Engineer
West Virginia University
Telecommunications and Network Operations
_______________________________________________
cisco-voip mailing list
cisco-voip at puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-voip
_______________________________________________
cisco-voip mailing list
cisco-voip at puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-voip