Skip to main content

I am wanting to increase my level of understanding of the IntelliSnap Big Data backup and recovery process. I have searched the Community, Partner University, and knowledgebases, and documentation without success.

 

Can anyone help with the following

  • Provide list of links or explanation that explains how this process works for backup and recovery and the differences between VM, San based Snaps, and COW(Linux Copy On Write).
  • COW (Linux Copy on Write) Snapshots vs hardware snapshots, pros and cons
  • Are there scalability limitations with COW snaps.
  • A break down on what some of the options actually do when configuring the Big Data VMware snaps. The documentation either ignores or glosses over this.
  • When using SAN Snapshots from a VM using Big Data (say mongo Idata agent), will I need to present the LUN as an RDM using NPIV. Is there a detailed method for doing this?

Hello @DaveK 
I had a look around for a intelli-snap white paper but could not find one as that seems to be what you are looking for but could not find it. 

I would say your questions require a solutions architect or some level of PS engagement to provide the education you are searching for. Sorry i cannot be much more help and good luck in your search for more knowledge.

 

Kind regards

Albert Williams


Thanks for the link, it has shown what SANS support Snapshots for various databases.

Would it be possible to answer the following

  1. Am I correct in assuming that when using Native Snapshot engine, no SAN Hardware Snap is being taken and that the Snapshot is leveraging VSS in Windows or COW in Unix.
  2. Given that Native SNAP is an “in OS snap” if you wanted to recover from the SNAP copy the operating system would need to be intact on the protected server. 
  3. What if any is the performance penalty in using a Native Snapshot, vs a Hardware based SAN snapshot in both backup and recovery and are there any performance tuning changes that could be made to mitigate these penalties.
  4. When using VM IntelliSnap's, is there a performance penalty in a File system or Database Recovery if the protected volume is made up of multiple extents split over multiple VMDK’s. and does this allow to proxy and non-proxy restores alike

Reply