When you create a deduplication data store, you specify whether the hash destination is on a Solid State Drive (SSD mode) or the hard disk drive (RAM mode). If you configured hard disk as the hash destination, you need more memory to process hash keys. As a result when your backup size grows, all your memory may get exhausted. In that case, you can add an SSD to back up more data. Similarly if you had configured an SSD as the hash destination, you need less memory to process hash keys. However, if you are moving to a higher memory machine, you might want to switch to the RAM mode for a faster hash processing.
To switch the hash destination from a RAM to SSD or SSD to a RAM, Arcserve UDP lets you modify an existing data store and change the mode as required.
You can modify an existing data store even when it is running but the data store restarts after you save the change.
When you switch from the RAM to SSD mode, you would need less memory. So, Arcserve UDP automatically decreases the minimum value of "Hash Memory Allocation". However, you can manually change Hash Memory Allocation. For this case, you change the hash destination folders to SSD. When you change the hash destination, Arcserve UDP automatically copies the hash files to the new location on SSD.
When you switch from the SSD to RAM mode, the RAM should be large enough to accommodate the current hash database. For example, before the change, the data store created 30 GB of hash files on SSD. Now after the change, you should allocate at least a 30 GB memory for hash files. If the RAM is not enough, the switch fails. In this case, Arcserve UDP automatically increases the following two parameters:
This ensures that data store starts after the modification.
For this case, you change the hash destination folders to the hard disk drive. When you change the hash destination, Arcserve UDP automatically copies the hash files to the new location on hard disk drive.
|
Copyright © 2016
|
|