Extremely High IOstat Figures, help diagnose?

[lastb0isct]

Hi All,

I have recently put my plex server as a VM and am curious if i can get some help to figure out some issues i'm experiencing. When playing any file through the plex VM i'm getting extremely high IOstat figures:



As you can see this is not going to work well, but it is the same system i was using previously for plex. The VM is help locally on a mac mini and is ONLY hosting that one VM. Nothing else is running on it.

It plays files from a NFS mount to my Freenas Box. Never been an issue in the past either.

Can anyone point me in the right direction on what might be causing this?

 

[sigxcpu]

Is your hard drive dying?

 

[mir]

The high IOstat is caused by transcoding I guess.
What disk type and controller do you use for the VM (Proxmox setting)?
How many CPU's have you assigned for the VM? (Very basic testing have shown me that to have smooth transcoding 4 CPU's were required)

 

[lastb0isct]

I have 8 CPUs dedicated to this. Disk type is local IDE0 and qcow2. I'm attempting from a different mac mini and am getting much better results, so might be that the HDD is on its way out of something. I will do a bit more testing and let you know what i find.

Even running it from my Shared storage over NFS is much faster and less waits.

 

[mir]

Disk type is local IDE0 and qcow2

Don't use IDE if you want god disk performance. Either use virtio or SCSI combined with virtio scsi controller.
To squeeze more performance out of your virtual disks change qcow2 to raw.

 

[lastb0isct]

Is it possible to migrate/move the disk to a virtio disk type?

 

[mir]

1) Shutdown the server
2) Remove the disk
3) Readd it and choose virtio

If this is the boot disk remember to adjust boot order in the options tab.

 

[lastb0isct]

Looks as though it might have just been a bad VM, maybe got corrupt somehow. I migrated over a snapshot version of the VM to the host in question and it is working without a hitch. Sorry for the false alarm and thanks for the help.