After the first lab, and a bit of reflection, here’s
another idea, presented in a different way. This time I’m considering snapshot
count (I have a real-world scenario in mind where one of the clusters is stuck
on ONTAP 9.1 for a bit so need to keep below 255).
DAG Layout
The table image below shows the DAG layout. 5 databases (DB1
to DB5), across 3 mailbox servers (MB1, MB2, MB3). MB1 and MB2 in the primary
site share activation preference 1 & 2 (odds on 1, evens on 2). MB3 in the
secondary site has activation preference 3.
Image: Database Activation Preference on Mailbox
Servers in our DAG
Backup Strategy
We want to keep 30 days retention, but only on MB3. MB1
and MB2 have 7 days retention.
To keep well within the 255-snapshot limitation of ONTAP
9.3 and below, our backup strategy is:
10 resource groups:
- 5 resource groups per site (1 for each database)
3 backup policies:
- 3 days hourly full & logs (used for site 1&2)
- 7 days daily full & logs (user for site 1)
- 30 days daily full & logs (used for site 2)
There is a script the runs on the SnapCenter server, to
trim the logs backups from site 2’s 30 days full, to keep maximum 7 days logs retention.
Image: SCE Backup Strategy
Exchange Server
Backup Policies
Details of the 3 Exchange Server Backup Policies:
Policy
name: 3 days hourly FL
Backup
type: Full backup and log
backup
Availability
group settings: Backup on
Selected DAG Servers
Schedule
Type: Hourly
UTM
retention: Delete Snapshot
copies older than : 3 days
On
demand Full backup retention: Delete Snapshot copies older than : 3 days
Hourly
Full backup retention: Delete
Snapshot copies older than : 3 days
Policy
name: 7 days daily FL
Backup
type: Full backup and log
backup
Availability
group settings: Backup on
Selected DAG Servers
Schedule
Type: Daily
UTM
retention: Delete Snapshot
copies older than : 7 days
On
demand Full backup retention: Delete Snapshot copies older than : 7 days
Hourly
Full backup retention: Delete
Snapshot copies older than : 7 days
Policy
name: 30 days hourly FL
Backup
type: Full backup and log
backup
Availability
group settings: Backup on
Selected DAG Servers
Schedule
Type: Daily
UTM
retention: Delete Snapshot
copies older than : 7 days
On
demand Full backup retention: Delete Snapshot copies older than : 7 days
Hourly
Full backup retention: Delete
Snapshot copies older than : 30 days
Image: New Exchange Server Backup Policy
Resource Groups
Just doing the resource groups here for DB1. Rinse and
repeat for DB2,3,4,5.
Resource Group
Name: DB1 MB1 MB2
Selected Resources:
DB1
Backup Policies: 3 days hourly FL, 7 days daily
FL
Backup Servers: MB1, MB2
3 days hourly
FL - applied schedule = Hourly
7 days daily FL
- applied schedule = Daily
Resource Group
Name: DB1 MB3
Selected Resources:
DB1
Backup Policies: 3 days hourly FL, 30 days daily
FL
Backup Servers: MB3
3 days hourly
FL - applied schedule = Hourly
30 days daily
FL - applied schedule = Daily
Image: Selecting Resources to Protect
Image: Configuring polices and schedules
Image: Microsoft Exchange Server > Resource Groups
for DB1
How Much Should You
Backup in one Go?
To avoid hitting Windows 10 seconds VSS snapshot timeout,
you might need to consider the size of your resource group. If it is a small
DAG, one resource group is fine. If it is a large DAG, multiple resource groups
might be required.
Resource Group ideas from large to small:
- All databases on all servers in a DAG
- All databases on all servers in a DAG per site
...
- Backup per server per site (all databases on that
server)
- Backup per database per site (all servers in the DAG
with that database)
THE END
Comments
Post a Comment