disaster recovery

[khilari]

Hi guys, is there any indepth disaster recover white paper available...best practices etc

[James R. Ferguson]

Hi:

There is no one white paper or particular collection that I know about.

From a technical perspective, you can begin to see the scope of good practice (and the decisions and trade-offs necessary to implement it) in some of the HP documentation at:

http://www.docs.hp.com/en/ha.html

For example:

http://www.docs.hp.com/en/T1906-90024/index.html

Regards!

...JRF...

[Bill Hassell]

You need to define disaster recovery. Do you mean that your computer room is smoke and rubble? Or that your boot disk has crashed and the computer no longer works? Or that your primary network router or firewall has failed? Or that the database or application you are using no longer works and you have no support contract?

You'll need to define the disasters, and then be prepared for expensive solutions associated with providing the recovery tools and labor to maintain the backup systems.

[VK2COT]

Hello,

To add to other valuable comments:

a) Disaster recovery should be based on your
company's business continuity plan.

b) Answer one single question and then create
DR solutions:

How much does it cost your company if
servers go down for, say, one hour?

If the answer is in millions of dollars,
then you need very resilient, robust DR solutions.

If the answer is along the lines "oh,
if server goes down, we can live without it for 24 hours"... then you can build simple
and cheap DR solutions.

c) Frankly, after 23 years in this business,
I am still amazed how often the companies do
not know how to calculate monetary value of
downtime of servers.

So, talk to your business and get their business continuity plans firstly. IT admins
must not assume they know how the business
runs!

Cheers,

VK2COT

[Steven E. Protter]

Shalom,

The second in a series of give me a single document on disaster recovery.

Lots of examples:
http://docs.hp.com/en/hpux11iv3.html

http://docs.hp.com/en/oshpux11i.html

There are many strategies. The two links above involve using a continental cluster or high availability solution to maintain instant fail over to a new site. They are very expensive, but do provide near instant fail over.

The first step in DR is to determine how long the business will tolerate being down. Most businesses, will tolerate some down time.

Having a DR site in that instance is enough, along with backup tapes as close to current as possible. Most important in a normal scenario is having instructions on how to bring the DR site up, because in the event of a DR scenario, there is a likelihood the original system administrators are going to be the opposite of alive.

SEP

[khilari]

Lets say I have a system RP7400 and I make an ignite image of it. Now, I make an ignite image and include only vg00 in it, whereas there was only 1 external vg, vg01. When I make the ignite image the ignite puts vg02 configuration into /tmp so that when I do a vgimport later, its configuration comes back..
Now, I do a re-ignite⠦⠦ I get vg00 back and all the configuration back.. Then I do a vgimport and get vg02 back⠦⠦. Lets say vg02 was 100 gig with 5 disks *c3t2d1 to d5).
Where I am recovering the system, lets say sungard they just give me 1 disk worth 100 gigs say c10t2d2⠦⠦ Although the space is 100 gigs that is given to me to restore but the device files are different⠦. Will that matter in the restore⠦⠦⠦. How do I reconfigure vg02 disk device files from 5 disks *c3t2d1 to d5). To 1 disk of 100 gigs named c10t2d2.

[Bill Hassell]

Well, at your Sungard site, the vgimport command is useless. The disks are blank so you have to vgcreate the volume, lvcreate the lvols and then restore the data from your backup tapes. You are correct to use Ignite for vg00, then restore vg02 from tape.

NOTE: Unless these disks are used as raw database volumes (not mounted with files), the names for the disk special files (c10t2d2) is unimportant.

[Prashanth Waugh]

For DR you can consider the following points
1) CA Replication across the two DC
2) manual DR or Automatic DR
3) Backup policies and restoration time if something happen wrong
4) both side are Active/Active or Active/passive. If the side are Active/Active then you have to consider the DR side local failover also.sometimes what will happen if you are in a situation of DR and similtenoulsy something wrong happen on primary node of DR site then the applications means packges will move to Local HA.This will happen if the Local HA will act as a DR Node.


Regards
Prashant

[Sivakumar MJ._1]

Hi Khilari,

My points on DR

1. Return on Investment for DR
2. How much the Business Impact without DR.
3. The Potential impacts of each type of disaster or event
4. Do a Risk Analysis.
5. The appropriate solution/Technology/Vendor/Policy for DR.