We have such problem : When an IS disk is being backed up, a snapshot is being created on the nodes where the stripes reside. I don't know if it is possible, but It would be great to create the snapshots elsewhere and not to use space from the IS nodes . Here are our arguments:
- Currently if there is not enough space for creating a snapshot for some disk, the backup task fails and there is no way to prevent this.
- It is not feasible to monitor all nodes and to consider how much space should be left unused (the size of the larger vdisk). Even if we could handle that, we can not choose the nodes where an IS disk will be created. So we have very little control on the space allocation on the nodes.
- We have backup servers and dedicated backup storage space. It doesn't seem ok to require backup space on the IS nodes. These are usually SSD disks and are expensive.
- We have been told that the snapshot that is created when an IS disk is being backed up takes as much space as the declared size of the disk. Don't we have thin provisioning for IS disks ? It would be much better if the snapshot didn't needed to be as large as the disk itself - this is lots of space wasted.
- If we have a disk that is larger than 50 % the size of the stripe, than there is not enough space to create a snapshot for this vdisk and so there will be no backups for this disk.
- It would be better not to create the snapshots for IS disks on the node where the disk is stored.
- If possible let the IS disks be thin provisioned and the backup snapshots too.
- Remove the necessity to constantly monitor and check the nodes space allocation and the everyday movement of stripes in order to be able to backup a disk.
Please sign in to leave a comment.