Hello!
I have a problem with the combination RAID 1 and Workstation (9). When resuming on the RAID 1 array, it takes twice as long (two minutes) as from a disk without RAID. Anyone having a clue why that is?
System and disks works as normal for all other software. Average I/O performance is 85MByte/sec (SATA 2.0)
Host:
Guest:
Parkdale HDD speed test on RAID 1 and single HDD drive is almost the same, around 85MByte/sec Read/write.
Anyone understanding why resume using VMWare Worksation (9) takes the double amount of time on the RAID 1 array compared to a non RAID disk?
Thanks.
Basic concept called Mirroring which runs on backend of the Raid 1. Whenever you try to resume the VM it starts but made sure that same copy of data is being copied after the vm is resumed and verifies the data consistency by matching its data.
in Single HD there is no need to verify the consistency and its just runs normally without verifying any mirroring or other capabilites.
So Raid is the reason of the slow resume of vms.
Hello Anjani!
I'm not sure if I'm understand your reply correctly - are you saying that WS detects the RAID array and therefore performs some extra checks to maintain consistency among HDs in the RAID array? Isn't the RAID array supposed to be transparent for programs accessing the drive? I'e all needed checking s are handled by the RAID controller, not the calling program?
If WS are doing some extra checks when detecting RAID 1 (mirror), then I can't see the gain or purpose as the RAID controller handles consistency and other "administrative" tasks, right? Can this "feature" be turned off in WS?
No other software I put on the RAID array works this way. Other VMs with the same amount of RAM and disk resume in 15 sec.
This is really strange. Hope that you can clarify what you mean.
Thanks