Veritas Checkpoints for content store backups

Sharon

Our current backingstore size is over 180 GB. Our current backup process is taking over 6 hours to complete. This is not acceptable because we are a global company.

We are trying to move to Veritas Checkpoints. We implemented the following tasks in our QA environment:

Freeze Teamsite
Unmount previous day's checkpoint
Remove previous day's checkpoint
Create Checkpoint
Mount Checkpoint
Unfreeze Teamsite
Create tar file of checkpoint
Backup tar file to another storage management system

After TeamSite is thawed, we get I/O errors when users attempt to update content in TeamSite (errors below). We replaced the Freeze/Unfreeze with TeamSite stop_all/start processes but still the same I/O errors occur.

Does anyone have experience using Veritas checkpoints? And if so, do you have any suggestions for what is causing the errors?


iwtrace.log entries:

[Wed Jul 28 00:55:21 2010] TEnvironment::Flush flushed 0, unlinked 0
[Wed Jul 28 00:55:21 2010] TEnvironment::Freeze freeze until Sun Aug 8 14:42:00 2010
[Wed Jul 28 00:55:45 2010] TEnvironment::Thaw
[Wed Jul 28 01:12:22 2010] TFileHandle: open(/web/data/iw-store/default/d1/dcc/d88/ffe-1,O_CREAT...) fails, errno 5: I/O error
[Wed Jul 28 01:12:22 2010] TFileHandle: Retrying open(/web/data/iw-store/default/d1/dcc/d88/ffe-1,O_CREAT...)
[Wed Jul 28 01:12:23 2010] TFileHandle: open(/web/data/iw-store/default/d1/dcc/d88/ffe-1,O_CREAT...) fails, errno 5: I/O error

Crsb

is there a valid explanation why it is 180GB?
Do you use iwfsshrink and other contentstore CLTs on periodic bases? these tools help maintain the smallest contentstore size possible and improve teamsite performance overtime.
Consider downsizing your contentstore or splitting it into multiple stores or at least to an archive store and a active store.
another approach is to run TeamSite on VM where you can take snapshots without having to freeze the application but there are disadvantages to this approach.