Disaster Recovery (DR) setup?

Doctortt,

 

Everyone here is happy to share

 

Example of our DR setup for EV

 

Running EV 2007 SP2

Running DA 7 SP3

 

7 EV servers doing a mix of mailbox journaling and FSA

OS on servers Windows 2003 SP2

 

2 SQL servers clustered

OS Windows 2003 SP2 64Bit

SQL 2005 Enterprise with two instances

 

 

All index and databases are stored on an EMC SAN

 

 

all archived data stored on a Centers

 

DR

 

Index servers are backed up fully once a day

 

SQL databases are backed up to disk once a day, SQL transation logs are backed up to disk every 15 minutes

Both SQL log backups and database backups are backed up to tape daily and then deleted (to conserve space)

SQL Databases are backed up to tape every night

 

DA server export directory we run a full backup friday and differential weekdays (data does not change much daily)

 

Centera

 

Our Centera DR is we have a duplicate in a different building and we do realtime one way replication from our live Centera (Source) to our DR Centera (Destination)

 

 

I know most may think this is overkill but we have learned from bitter experience that because of the 30TB of data we have in our archive if we needed to rebuild it from scratch it would almost certainly kill us so we protect as much as we can

 

FYI we have so much storage because our present retention is keep forever

 

 

Doctortt,

 

In one of my customer we have clustered 3 node of EV with Veritas Cluster in Global Cluster manner.

2 nodes are located at HQ, 1 node is located at DR site.

HQ data is being replicated to DR site over WAN by Veritas Volume Replicator.

 

I hope this would give you an idea

 

Regards,

 

Koray

 

Just to give the very far other end of the spectrum we are a small shop and we do a backup of the databases nightly (for db maintenance as much as anything) we vault every friday backup to tape and then rerun the vault to place placeholders and keep our friday tapes offsite for one week. We  do a full backup of the vault and sql servers once a month which leave the house for a month.

 

A wopping 30 GB (yes mr 30 TB I said GB) vaulted from file and exchange server, woot woot

 

(hey I said the very far other end of the spectrum)

:)

I wish i could do weekly or monthly backups

 

And only had 30GB vaults

 

Life would be so much easier

Do you have your EV servers clustered? I didn't see any failover mechanism in your DR Setup

sleddog,

 

Thats my next step

 

I was on EV6 SP5 not so long ago. Now that I'm on the latest and greatest my next plan is to cluster the EV servers. I'm planning on creating two clusters half of my servers will go to one and half to the other. I will have one passive server per cluster and each EV node will be configured to fail over to the passive node incase of failure

 

I think thats my last area to cover. After that i think i'll be covered but always cautious