Hi,
We've started deploying Server 2016 VMs to migrate away from Server 2008, but they seem to be sooo painfully sluggish and slow. I've read about disabling the UNMAP feature with the following commands:
fsutil behavior set DisableDeleteNotify NTFS 1
fsutil behavior set DisableDeleteNotify ReFS 1
These commands have been run a few weeks ago, but still the performance of the VM is no different, they are almost unusable. Any one got any other ideas or suggestions?
Currently running VCSA 6.5 with ESX 6.0 hosts. Both at the latest build levels.
We don't have any of these issues on 2008 servers.
Seeing as how vSphere 6.0 is EoGS in ten days, you might should start with upgrading your hosts first.
Not sure thats going to happen as the hardware is not certified for ESX 6.5 and with a plan in the future to potentially migrate this environment to the cloud, then that also will likely mean it wont happen. So do you have a reasonable response to the issue now?
If it's only a "potential" at this point, then your environment will be unsupported in 9 days. That said, unless you can provide more details on how you've configured your VMs and your underlying storage, there's not much to go on.
Its been configured this way for the past few years, many of. There are no issues with the performance of other VMs running 2008, CentOS, even 2012. But 2016 just seems to be painfully slow.
The forums are seemingly littered with people with the same issues, they cant all be mis-configured surely!!!
Hi
could you share the performance graphs from vsphere? CPU usage, RAM Usage, disk usage/queue, LUN latency....
The disable of the UNMAP feature looks like conflicts 6.5 version of ESX.
Regards