Hi all,
I'm testing a virtual machine setup with Server2016 as host and replica and Server2016 and SQL-Server 2016 in the guest-VM.My setup is 1TB Raid 1 for the host-system and 1TB Raid 1 for the VM. Hyper-V Replica is enabled and synchronizing every 5 minutes. There are 5 DB's with together about 60GB in the VM.
Despite the fact that i should probably use a different harddisk-setup/Raid for the VM, I see poor write and read performance in the VM. I used Perfmon to messure the "avereage disk sec/write" and "average disk sec/read" performance.
These are measurements over 4 days:
- Host AverageDiskSecWrite of the harddisk the VM resides on = 0,027
- Host AverageDiskSecRead of the harddisk the VM resides on = 0,015
- in the VM AverageDiskSecWrite of the harddisk = 0,361
- in the VM AverageDiskSecRead of the harddisk = 0,016
I see that the AverageDiskSecWrite in the VM is very poor. I also tried a DB-restore in the VM with hyperV replica enabled and disabled. With replication enabled it took about 2 and a half hours to restore a 42GB Database. With replication disabled it took about 40min to restore the same 42GB Database.
Now to my actual questions and hopefully someone can help me please :).
Why is it, that with replication enabled it takes more than 4 times longer to restore a DB?
Are there any HyperV-settings that I could use to get better write and read Performance in the VM?
I already googled a lot but despite the fact that i should use a different harddisks-setup for the VM i didn't find anything useful.
The things simplest setup suggested is a raid 1 for the VM-OS and a different raid 1 for the VM .mdf and .ldf files.
Kind Regards
Chris