In the background, NimbleOS finds all of the blocks that don't have any references to them, and deletes them. When you delete a snapshot, you delete the map (metadata). Instead of a large file for your volume or VM, that volume is made up of millions of small data blocks. When a snapshot is taken, a copy of that map of all of the blocks is made, and the live volume continues to use that existing ever changing map. The way I think of it, a Nimble volume is a map (metadata) that says all of the blocks that contain the data for the volume. Nimble snapshots and their volumes are very different than how VMware and Hyper-V do snapshots. So, if deleting the snapshot deletes the changes made to a disk and deleting the parent disk deletes all of the data the snapshot doesn't contain, then I'm a little befuddled as to how to manage snapshots on a Nimble, unless HPE's wording is very unclear on the topic. Only the data on the snapshot is lost." If I understand this correctly and if we use my previous example, wouldn't this mean that deleting a snapshot would remove the recorded instance of the word "dog" being written to the file? If that's true and if there be two versions of the file, one in the parent disk with the word "cat" and one in the child disk with the word "dog", how does one remove the redundancy in the storage without also losing the data stored in the snapshot? If the snapshots only track differences in blocks changed since the snapshot was taken, then the snapshot does not contain all of the blocks of information that the parent disk has. Taking this line of reasoning into Nimble, I'm a little confused because the Nimble documentation Opens a new window states that "deleting a snapshot has no impact on the original volume. So, the child's (Snapshot2) contents were saved and the parent's (Snapshot1) were discarded. The file's contents were in conflict when trying to merge the two disks. If you boot back to Snapshot2 and then delete Snapshot2, the file will now say "dog" in Snapshot1 because you've deleted your ability to go back to a time when the file said "cat" in a snapshot. If the file needs to say "cat" again, the VM can be restored to Snapshot1 and it's as if Snapshot2 never happened. The file's contents are changed to say "dog" and Snapshot2 is taken. Hypothetically a text file with the word "cat" is created and Snapshot1 is taken. A snapshot creates a child disk that begins to store changes to the data that is on the parent disk from the point of inception of that child disk, and when the child disk is deleted any changes held by that child disk are merged into the parent disk with conflicts erring on the side of the child disk's copy of the data.įor example, consider a VM. My experience with snapshots is based around the way they are implemented in Hyper-V. So, instead of paying for bigger drives to offer up more space until the new unit is in-place, we are looking to get rid of some of the snapshots to reduce space usage. This SAN is running out of space and is soon to be replaced. In short, if I delete a Nimble SAN snapshot, is the data lost or does it get merged back into either the snapshot before it or the base disk? So, hopefully someone here has experience with this. It's been a while, but I can't find any information about this at all online.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |