Hi,

I've got a mythbuntu server running at home and this morning I noticed that I couldn't log in. I turned on the screen and it was blank.

I tried to reboot but it wouldn't get past the BIOS POST screen as there was no MBR to read.

I have two SATA drives in this installation one for the OS (ext4) the other for the Mythtv recordings and movies (jfs).

I attached the drives on two seperate systems but only one partition (/home) was readable on one drive and only for a short time the device then disappears from /dev.

Running fsck on the ext4 drive comes up with:
"zero length partition?"

The jfs drive when running jfs_fsck says:
"Invalid magic number in the superblock (S).
Superblock is corrupt and cannot be repaired
since both primary and secondary copies are corrupt"

I can't seem to recover either of the drives.

Now I'd love to know what happened as I've never seen two drives fail simultaneously. I have had an entire system fail before because the PSU fried all the hardware, but never this.

The system boots fine with a rescue CD and I'm running memtest at the moment and all looks good.

So I'm scratching my head, do I need a new PSU, motherboard, hard drives or just continue with all the current hardware and do a fresh install?

Was I hacked and the intruder wanted to damage rather than steal and did dd commands on my disks? I don't know.. I only had ssh on port 2222 running (I never get probed on this port) with good passwords and apache on 443 for the mythweb interface.

I just don't know how to continue now as I don't have any indication what failed and if it will fail again. If only I could find the smoking gun so I'd now how to rebuild. I couldn't get to the /var partition to see the logs. :-(

I've had 2 SATA drives fail in a very short interval. Both were extremely hot and had been crammed close together - my fault, an old case. Was a while ago, when SATA disks were reasonably new on the consumer market.
Put (E)IDE disks in and an extra case fan, no repetition. Circumstantial only, but removed any suspicion on the disk controller chip/motherboard. Are your two disks on the same controller ?.

Hi Syg00 thanks for your reply. I don't think it was heat as the drives are spaced well apart and the case is very cool I also have lm_sensor alarms on it, and as you know it's still winter down here in God's Own Country. :-D

However! Some good news!

The JFS hard drive was recoverable. Oddly jfs_fsck would not work correctly when using a sata to usb controller and mounted on my laptop. When I put the drive back in the original machine and using a rescue disk ran jfs_fsck it recovered the journal and all the files are still there!

The other drive is semi cactus. I can recover almost all partitions except for '/' which is a shame as that has my myth mysql database on it. I did have a nightly backup in place but 6 months ago I rebuilt the machine and kept putting off writing some new backup scripts.

I'm guessing the drive controller on the actual hard drive is bust? Whenever I do anything with the first partition I get I/O errors and lose the device but oddly the other partitions are happy to play along.

Oh well at least I have all my recordings, although without the database it's a huge mess and difficult to import into a new database.

Hi

I did computer assistance that was installed windows xp. Presented data corruption. I tested the memory and then realized that the problem was the HDD. I changed the HDD of the machine ...

The HDD that was defective, it was relatively new, had 3 years of use. PSU used was common, not quality. I thought then that those who caused the problem in HDD was the poor quality of the psu.

I decided, out of curiosity and intuition, remove the logic board of the hdd. I checked that there are contacts between the pins and the logic board inside the hdd. I noticed that the contacts on the logic board, were oxidized. I used a soft rubber to clean the contacts, in order to remove the oxidation.
Reassemble the logic board of the HDD instead.

Hdd works perfectly for some time ...

The hdd was defective Samsung brand, but I noticed that the brand Seagate HDDs also use the same type of contacts ....

good luck

We usually loose stuff in a power drop or spike. I have lost stuff in lightning strikes too.

I'd guess a power issue.