Hello community!
I have one more problem with a Hypervisor that runs VMware ESXi 5.1.0 (Build 799733). Please allow me to explain a little the history in order to let you understand what I am facing.
This host runs several machines which I had to migrate in order to upgrade the hard drives on an RAID-5 array with larger ones. So I installed temporarily a 3TB SATA drive connected directly to the motherboard, I migrated the most VMs there and then I destroyed the RAID-5 array I had (it was built by 5+1 hot spare SATA 7.2k HDDs of 1TB each, which was giving a total of 4TB of capacity). Then I built two new RAID-5 arrays with 3 HDDs each. The first one was consisted by 3x SATA 7.2k 3TB HDDs and the second one was also a RAID-5 array consisted by 3 of the old 1TB SATA 7.2k HDDs I had. No-one array has a hot-spare now. Some VMs were migrated to a D-Link NAS connected to the hypervisor via NFS. Really slow, but it was just temporary. As soon as I migrated the files (I used Veeam Backup & replication 6.1), I destroyed the array and built the two new ones as explained above.
Then I started migrating the VMs back to the normal Datastore (meaning the internal Hot-spare HDDs of the server, connected to a RAID controller). I used again the same software (Veeam Backup & Replication 6.1). I realized that the virtual disks were now thick provisioned, and I am sure that I had no choice of that, while I was preparing to copy the files with veeam. But that not the real problem.
Between others, I had a Windows 2008 server which was my storage server. Huge capacity (2TB) with lots of shared folders, different sharing permissions and hundreds of gigs of data. All these managed by a Windows 2008 R2 std server. The server itself (meaning the VM of course) was a virtual disk of about 30GB. That is just for the OS and the applications. All the data mentioned above are on a second virtual disk, which is stored on another directory (outside of the VMs directory) on the datastore, just to be clear that it's a separate virtual disk. Of course it is an independent virtual disk and it is possible to be attached on any Windows VM, in case something is wrong with the Srv-Storage (the Windows 2008 R2 machine). Unfortunately that virtual disk of 2TB is just one vmdk file and as soon as I am trying to put it back on the new datastore I created with the new HDDs, it fails. It starts normaly, but at the end it fails.
I can;t understand how I managed to migrate it on that temporary SATA drive and it was migrated without any issue. It is working properly now. The problem is that I can't put it back.
So I need your lights on two things, regarding that project. How to migrate that huge virtual disk of 2TB without issues back to the new array? Ideally I would like to have it thin provisioned. And finally please let me know if there is a way to make all the VMs as thin provisioned again, as I don't like to take capacity from the datastore that is not needed.
Once again, thank you so much in advance. I am sure that community will help me again and I hope I will be able to help others pretty soon.
Kind regards,
Angelos Pitsos
