Slow local copy of vm between datastores

[oaklodge]

I'm new to ESXi and recently had to upgrade the raid in an HP ProLiant ML330 G6

I installed a temporary 1TB SATA disk connected to the mother board

I created a second datastore on the new drive and then copied the VMs to the new datastore.

All worked fine but it took 2 full days to copy 600G of data (3 virtual disks).

I'm now trying to copy the data back to the new raid (4 1TB disks as Raid 10) and it is EXTREAMLY slow.

I'm doing the copy using putty to connect to the shell (having tried the vSphere client).

Is this poor performance a configuration issue for the raid or the server?

(may laptop performs 50% faster!)

Thanks.

[idle-jam]

doe your smart array raid controller has Flash Backed Write Cache? you will need that for an acceptable write performance else it will be very very slow.

[AndreTheGiant]

In any case the copy will be slow (if you use datastore browser or command line or SCP).

But in your case seems a big problem with the write cache policy.

As suggested check if your controller is able to perform a write back policy.

Andre

[oaklodge]

Hmm, that looks like the problem - there is a slot sitting empty.

I have a choice between 512MB and 1GB - any idea how to choose which to use?

Thanks.

[idle-jam]

the price between 512MB and 1GB is not much in my country, i would buy 1GB and forgetting about write cache issue once and for all ..

[oaklodge]

Thanks