Attached are system logs from both servers.  Running ESXI 5.1 with latest updates and unRAID 5.0.5.  Had configured the first system with no issues (ran pre-clear script on both) and when performed parity test got approximately 80MB/s and completed within 400 minutes of so. 

 

Setup the identical system (syslogB) and am getting 4.44MB/sec.  And now when I try and run a parity test on the first system (syslogA), it's even slower. Don't know what is wrong.  Please help.

SyslogAandB.zip

Bump - please help

Boot bare metal and retest. Otherwise ask for help in the virtualization area of the forum.

SyslogA:

Looks like you have > 3GB of mem installed on your unRAID VM.  I would reduce it to 3GB since that is the maximum memory that ESXi will allow before a 32Bit OS needs to use PAE.  ESXi reserves 1GB for devices so the top 1GB of 4GB you have will only be used by PAE and I think you might see better performance by dropping it to 3GB and not using PAE on a 32Bit VM.  Also it looks like your unRAID flash drive might need to have the file system checked based on this:

Mar 23 00:06:16 Tower kernel: FAT-fs (sdb1): Volume was not properly unmounted. Some data may be corrupt. Please run fsck.

The memory comment is based on this line from your log:

Mar 23 00:06:16 Tower kernel: 4228MB HIGHMEM available.

which looks like > 3GB installed in the VM to me.  You shouldn't need more than 3GB with a VM install of unRAID since most plugins that would require memory would be in separate VMs.  At least if following the same pattern I'm using buy keeping my unRAID clear of as many plugins as I can - anyway.  Now will these changes fix your problem - maybe not - but it is what I would try first.

 

SyslogB:

I see one thing that looks the same in the log from SyslogA:

Mar 22 01:47:12 Tower2 kernel: FAT-fs (sda1): Volume was not properly unmounted. Some data may be corrupt. Please run fsck.

Which sounds like it wasn't shut down properly the last time you shutdown your unRAID VMs.  So I would suggest you run a check disk on a Windows box/VM for each of your unRAID flash drives.  The unclean shutdown certainly look like what happened based on this:

Mar 22 01:47:35 Tower2 kernel: md: correcting parity, sector=128

So it did a parity check when it started back up.  Looks like from SyslogB that this VM has much lower memory on it so no need to reduce it further you might even want to increase it some but stay below 3GB if you do.

 

So fix your flash drive problems and do clean shutdowns and see what your speeds are.

 

Note I would get a second opinion on your logs as others will be able to read them better than myself.  But doing as I suggested will not hurt anything either.