ESX 6.5 Raid 5 Rebuild no Datastore

[gscalone82]

Hello to all

i have this issue after a raid rebuild.

Vmware Esxi 6.5 start and no problem.

But i don't see any vmfs datastore. when i try to mount i get this error

either the selected disk already has a vmfs datastore or the host cannot perform a partition table

i try to use vmfs recovery from diskinternals with a boot iso and i see after 48h the data.

is possible restore the vmfs datastore? i think that i only lost the partition.

Can someone help me?

Thanks

Giorgio

[devakumar]

Hi Giorgio,

May I know how this happened? Did you replace any failed drive?

Check the vmkernel.log and see if the datastore is detected as snapshot.

If they are not check the partition table using partedUtil getptbl command

You can recreate the partition table using VOMA - VMware Knowledge Base  . I am also including the other KB for recreating the missing VMFS datastore parition - VMware Knowledge Base

Thanks

Devakumar

[gscalone82]

Hi Devakumar

thanks for your reply

no disk replace now i will try your link and i hope to solve the issue

thanks

giorgio

[continuum]

> i try to use vmfs recovery from diskinternals with a boot iso and i see after 48h the data.
Unfortunately the fact that Diskinternals is able to display your missing content after days (- or UFSexplorer shows the content after minutes)
does not mean that the VMFS-volume is in such a good state that rebuilding the partition table can be expected to fix your problems.
Do I guess right ? - do you use RAID 5 ???
Raid rebuils with RAID 5 and VMFS content are hit or miss: in my experiences that results can be:
- no serious issues other than a wiped partitiontable
- VMFS volume is not good enough to mount it again with ESXi buildin tools - but good enough to extract the VMs via manually created dd-commands
- VMFS volume lost VMFS metadata for directories and files - this means that only thick provisioned and unfragmented VMDKs can be recovered manually
- RAID rebuild gone so wrong that nothing but very small user files can be recovered
Anyway - to analyse which option you have I would need a VMFS header dump.
See Create a VMFS-Header-dump using an ESXi-Host in production | VM-Sickbay
If you are lucky recreating the partitiontable is all you need to do.
Do not know if using VOMA for this task will do anything good - in my daily work with recoveries I have not seen any reason to even try VOMA ...
Be careful - if rebuilding the partitiontable with partedUtil goes wrong - you probably make fixing the issue harder than it would be if you had done nothing at all.
The VMFS header dump hopefully provides all info to:
- be able to correctly build the partitiointable
- be able to create dd-scripts to manually extract your VMs (in case the VMFS is too damaged to mount via ESXi but still good enough to read from Linux
If you want me to check the VMFS I would suggest to call me via Skype.
In most cases I need less than an hour to offer a solution or to be able to give you some really bad news.
Ulli
skype: sanbarrow