I run a OMV server for more than 6 months without issues until i accidentally turned off my UPS. After turning the OMV box on I've noticed connection to the server has been severed. After connecting a keyboard and monitor and booting knoppix I noticed the one of the data drives is inaccessible. Gparted reports the following errors Invalid argument during seek for read on /dev/sda and The backup GPT table is corrupt, but the primary OK, so that will be used The gdisk output in the imagehttps://ibb.co/Lt74p0p mentions something about the partition being larger than the drive, not sure if the info is useful but the 4Tb drive in question had been cloned from a 2Tb one with partitions grown to new size. Is there a way to fix this issue without reformating?

The usually recommended tool for recovering a partition table is testdisk. It's sometimes not the easiest tool to use, but it's quite good at discovering where partitions should be. It's included in the SystemRescue live CD, which is my personal favorite for doing system recovery, disk image backups, etc.

1 members found this post helpful.

I suggest testdisk(terminal application) too. You will have first to understand how it works and be very careful, but it does good job, for partitions reappear and also for saving to other disk the content of corrupted disk (photorec also saves content, same package with testdisk I think). Testdisk is in most Linux repositories also, for if you install your disk in another computer.
I am not sure if you sould select MBR or EFI GPT in testdisk, regarding what I read in your photo, maybe try both.
Visit all testdisk's capabilities, then write to disk!
I wish you luck!

1 members found this post helpful.

I'm not sure if this is a new feature or not. I use parted to create my tables/partitions. Recently, while doing some bit-flip testing, when the partition table got corrupted, parted offered to restore a backup of the partition table.

1 members found this post helpful.

testdisk is the usual answer, but given it's a data drive, and probably a single partition, I'd just do as gdisk states. Delete the partition from say parted, then fire gdisk back up and let it do its stuff. When all is clean, allocate a partition over the entirety of the remainder of the disk.
Simple.

1 members found this post helpful.

Thanks to everybody for your replies, I used testdisk to recover data and configs from data drive but it turned out the system drive, although seemed to be ok (it mounted without errors and seemed to be readable at first glance) was also corrupted, i could not access some folders from a live media and testdisk showed errors when recovering files.

But luckily i have recovered the important data on these drives - docker config folder and compose stacks, the most important data i have are my photo/video gallery and my documents, but they are stored on my pc, on two drives on the server that suffered the outage, and a secondary mini pc that is used mainly for backups and data sync.

P.S. - I have added a separate physical drive to store docker configs so i can make drive image backups for the worst case scenario like i have experienced.

I think Ranish has updated the program maybe look for it.