Surgient Success

I build machines where I have to make continuous backups of the harddisk during frequest intervals. How much size do these snapshots take? Does it depend on the size of app config?

Moderator edit: This post previously within thread http://success.surgient.com/forums/t/304.aspx

Snapshots are captures of VMware disk writing differences.  Think of them as a writing journal on a block (sector) level where blocks can be overwritten.  If your disk writes are minimal the snapshots are small.  If you overwrite most of your disk(s), then the snapshots can rival the base disk image in size.  The total size of an app config snapshot amounts to the sum of the write activity for each server.

Richard,

Is there a threshold of snapshots to base image that one needs to be aware of, or can unlimited number of snapshots be made (depending on library size of course) since they all reference a base image? 

The only limits are imposed by either organizational quotas or backing storage.  We have customers with terabytes of snapshots unconstrained by which base images they refer to.

In our current environment I need to take a snapshot before running a test on the VM. If the test is a failure, I will need to go back to the snapshot and reuse the same image for the next test. If the test is a success, is there a way I can commit the snapshot to the base disk image? If I can get everything in to one single disk, it would be easy for the backups.

Thanks,

Praveen. 

Richard Cardona:

The only limits are imposed by either organizational quotas or backing storage.  We have customers with terabytes of snapshots unconstrained by which base images they refer to.

I do not know from where I have heard or read about this, but isn't it true that the more the number of delta disks (snapshots), the lesser will be the performance of the machine using the disk? Is it different in the VQMS environment?

The Surgient platform doesn't keep chaining snapshots, one on top of another.  We only go "one deep" with our snapshots, to eliminate the need to have to be cognizant of the snapshot chain depth.  This means you avoid the management headaches, performance constraints, and virtualization vendor length limitations.

When you deploy a snapshot taken via the Surgient platform, you actually get a writeable copy of the snapshot.  "Snapshotting the snapshot" is simply the act of saving that writeable copy back to the library, this time including the blocks changed during the most recent deployment.

That answers my question of whether keeping the snapshots with the base disk. In my case, I would just delete the old snapshot if my test is successful and create a new snapshot, where as if my test is a failure, all I need to do is roll back to the saved snapshot.

Thanks,

Praveen.
 

Or you can promote your snapshot and move it into the library.