I am having an issue where i have a really low ratio on deduplication. I already checked the proportionalCapacity to be in 0, but my ratio is still in 1.64x.
In my environment i have two different policies:
If i have these two policies applied, will the deduplication and compression apply to both groups of VMs and between different policies?
What other factors do i have to got in mind to increase the ratio?
Thanks in advance.
Hello Lalegre,
"but my ratio is still in 1.64x."
I have seen worse and I have seen better in general mixed workloads (lower generally being with relatively low % capacity utilised so hard to tell).
What is your current % used of this vsanDatastore and what % of that is a few (relatively) massive VMs?
"If i have these two policies applied, will the deduplication and compression apply to both groups of VMs and between different policies?"
As it is applied at the 4K block level I don't see why RAID5/RAID1 should matter (other than potentially the parity data being uniquely unique).
"What other factors do i have to got in mind to increase the ratio?"
Many other factors:
The most important ones being the uniqueness of your data (e.g. 100% unique = zero dedupe savings, while 0% unique means +1.00x for every copy on that Disk-Group); and the size of the Disk-Groups - bigger Disk-Groups mean more data mean better odds of commonality and thus increased savings, this is obviously going to be a trade-off though as the bigger the Disk-Group the more that has to rebuild from a single disk failing, lower potential (usable) cache:capacity ratios and the fact that 2 Disk-Groups performs better than 1 (with the equivalent size cache+capacity split over 2 Disk-Groups).
More info to add to this here:
https://storagehub.vmware.com/export_to_pdf/vsan-space-efficiency-technologies
Bob
Hello Lalegre,
"but my ratio is still in 1.64x."
I have seen worse and I have seen better in general mixed workloads (lower generally being with relatively low % capacity utilised so hard to tell).
What is your current % used of this vsanDatastore and what % of that is a few (relatively) massive VMs?
"If i have these two policies applied, will the deduplication and compression apply to both groups of VMs and between different policies?"
As it is applied at the 4K block level I don't see why RAID5/RAID1 should matter (other than potentially the parity data being uniquely unique).
"What other factors do i have to got in mind to increase the ratio?"
Many other factors:
The most important ones being the uniqueness of your data (e.g. 100% unique = zero dedupe savings, while 0% unique means +1.00x for every copy on that Disk-Group); and the size of the Disk-Groups - bigger Disk-Groups mean more data mean better odds of commonality and thus increased savings, this is obviously going to be a trade-off though as the bigger the Disk-Group the more that has to rebuild from a single disk failing, lower potential (usable) cache:capacity ratios and the fact that 2 Disk-Groups performs better than 1 (with the equivalent size cache+capacity split over 2 Disk-Groups).
More info to add to this here:
https://storagehub.vmware.com/export_to_pdf/vsan-space-efficiency-technologies
Bob
Sorry for the delay.
The current usage of the vSAN cluster is 72% and i have a total of 153 VMs and there are more to go in the next few weeks.
Another question:
Is there a way to see if a VM is currently been deduplicated? I mean to see the information of deduplication and compression individually.
Thanks in advance.
Hi Lalegre,
Another question:
Is there a way to see if a VM is currently been deduplicated? I mean to see the information of deduplication and compression individually.
No, you are not able to see this on VM level. Deduplication and compression is a cluster wide setting.
However, you can see a used capacity breakdown for certain objects. See figure below.
Rick
