VMFS volume error after powercut

[RaidJan]

Hello,

I have a ESXi 5.5 installed on Dell PowerEdge T110 2  who was running well. I decided to migrated to 6.0 but in case of any problem I made a clone of my disks. During the clone I got a long powercut and the UPS couldn't support it and now I cannot start the systen and cannot update.

If I try to start ESXi 5.5 I got an error:

error loading /misc_dri.v00

If I start the update with the ISO burned I got:

The selected disk has a VMFS volume which cannot be scanned. It may be corrupted or disconnected. Please select a different disk or correct the issue.

So I was thinking to start with the ISO of ESXi 5.5 to repair the system but I got this error:

The system has found a problem on your machine and cannot continue.

Could not load multiboot module : Bad parameter.

I tried also to burn 2 times each ISO to be sure they're not broken but doesn't change anything. 

I didn't see anything on the forum to help me, maybe I miss something?

Also I migrated before from 5.1 to 5.5 and worked well. If I rollback the system to 5.1, do I loss my VMs?

Thanks for some helps

[cykVM · Accepted solution]

I would try a fresh install to a 16GB+ USB key or SD-Card first and afterwards try to add the datastore. Follow this KB: VMware KB: Installing ESXi 5.x on a supported USB flash drive or SD flash card Looks like your storage controller is at least partner/OEM supported (DELL) and you need the drivers for it.

If the datastore is mountable with that temporary system I would first do a backup of all your VMs, if not you may try to repair it.

[RaidJan]

I didn't install with the ISO designed for DELL, at this time I don't remember if it's was available or maybe I didn't saw it.

Ok I will try this, and give you the feedback,

Thanks for your help

[cykVM]

At least for 5.5 Update 2 version the DELL customized ISO is found here: VMware ESXi 5.5 Update 2 Driver Details | Dell US

[RaidJan]

I will download this one also thanks. I will anyway try with the ISO I used to install the system before, maybe better same one. If doesn't work I try with this one !

[cykVM]

Which storage controller do you have there?

I at least won't do the downgrade until you know the datastore is working/mountable. The alternate bootbank might be corrupted, too.

[RaidJan]

1 HD SATA 500GB

2* HD SATA 1TO / RAID

[cykVM]

Ok, that's your disk configuration. I guess that VMWare is installed on the 500GB (single) disk and the datastore is on the RAID volume?

But still not clear which storage controller is in use...

Or don't you have any separate controller and this is onbaord (BIOS) RAID?

[RaidJan]

Yes I don't have any storage controller, I mean it's the basic one from the motherboard I believe.

And I said mistake about my HD, this one is where I do some tests but after check I have 3 datastores.

- System installed on the 500GO with 1 datastore

- And one datastore for each 1TO hard drive.

Was not in RAID because I remember now I had some issue with one hard drive.

Why it's important to know the storage controller?

Also after check I installed with the basic ESXi 5.5, not the custom image for Dell, the funny thing when I try to use your ISO just for boot (in case I wanted to see if this ISO I could repair) I got direct an error "Multiboot header is not found".

I didn't have time today to test with the USB media, tomorrow I will have a look, thanks for your time cykVM !

[cykVM]

It's important to know the storage controller because sometimes you need a vendor driver for the controller as it's not supported by (native) VMWare.

That's why the larger vendors usually provide customized installation images for VMWare including those drivers.

It's no wonder you had issues with the RAID, because it's basically software (BIOS) RAID function which is not supported by VMWare.

I would probably try to install a working VMWare to USB media or if you have another (empty) disk handy use that and connect only 1 of that 3 disks used before (test them one by one).

Looks like - maybe - "only" the 500GB disk is corrupted and the 2 1TB ones might be OK, so disconnect the 500GB one first and test one of the 1TB ones.

[RaidJan]

Sorry to answer late, time to test everything!

So I follow your advice and install VMware 6.0 on one external USB key (16GB). I tried to install the 5.5 version but I got the boot error during install... With 6.0 version it's worked fine.

But to perform this installation I had to remove the 500GB hard drive who is corrupted has you was thinking (if I keep it VMware is refusing installation even on the external usb key, I mean I cannot reach the page where you choose your hard drive).

So when install done, I just exported all the VMDK. I will format the first drive (500GB) and will set my both 1TO hard drives in RAID. It makes sense for you to keep the system install on an external USB key? (on this case if I had a copy of the system install on external USB key I could probably restore it very easily)

[cykVM]

As you are using the onboard controller I assume that RAID won't work at all. As stated above this is basically a software RAID setup in your BIOS and not supported by VMWare at all. VMWare will just present the separate 1TB disks and not the RAID volume.

There is no disadvantage in running ESXi from a USB key. You may even clone the key after host is configured to your needs. Maybe even consider to take a host configuration backup from time to time, see VMware KB: Backing up and restoring ESXi configuration using the vSphere Command-Line Interface and ... which you are able to restore to a fresh installation if anything goes wrong.

I guess there were no VMs on that corrupted 500GB disk which you need to restore?

[RaidJan]

Yeah just got lucky on this case, no VM on this hard drive... It was the datastore1 but we used for saved ISO of OS.

As you are using the onboard controller I assume that RAID won't work at all. As stated above this is basically a software RAID setup in your BIOS and not supported by VMWare at all. VMWare will just present the separate 1TB disks and not the RAID volume.

--> But if I used the Dell VMware I should be able to use with RAID no?

[cykVM]

--> But if I used the Dell VMware I should be able to use with RAID no?

Well it depends on the actual controller. As an example: With HP's SmartArray B120i/B320i onboard controller there's only vendor (HP) supported driver.

But this is not a software RAID. Your controller should have at least a battery backup (a rechargeable battery supporting the controller's cache in case of a power failure). Without a battery and cache memory it should be a very basic software controller which is not supported at all by VMWare not even with the DELL customized version.

With DELL boxes usually you need at least some DELL Perc xxxx controller for RAID volumes to be supported.

[RaidJan]

As you thought my controller (percs100) isn't compatible. So I just installed on the USB key and I don't used my both 1TO in raid. So I format all my hard drive and used as before 3 datastore (one per disk).

I will have a look about your procedure for backup my configuration. Thanks for your time cykVM

[cykVM]

It's quite easy to distinguish the newer Perc controllers, the software ones are called S... and the hardware ones H...

DELL has a nice table with basic specs on the Percs down to Series 6/7: Dell PowerEdge RAID Controller (PERC) | Dell

You can even see from this list which models support a BBU or flash backed cache.