Kernel panic and automatic rebooting.

[sossego]

I'm not able to post the message due to the fact that the FreeBSD installation constantly reboots.

"Ufs_dir" along with "bad inode" and "mangled entry."

I would like to try to fix the problem.
Do I use the "Safe" along with "Single" and "verbose" options during boot?
If it doesn't help, there will be an edit to this post ....
Going to try now.

Edit: Did not work.
Remember that I am not able to post the exact message.

"Sector 32901 at ino 512"

Mounting /var cause a kernel panic..

What do I need to do?

 

[sossego]

The problem has stopped for the moment; but, please do not have this marked as solved. It may reoccur.

Thanks muchly.

 

[xtaz]

I had exactly this same issue and posted here thinking it was a problem with FreeBSD. You can see my posts here: http://forums.freebsd.org/showthread.php?t=35438. It actually turned out that my hard drive appeared to have gone bad, bought new hard drive, reinstalled, problem has gone away.

I suggest that you install sysutils/smartmontools from the ports and run this command:

# smartctl -a /dev/ada0

Change the device name for whatever your hard drive actually is. You'll see some counters at the bottom. In my case the ones for Current_Pending_Sector, Offline_Uncorrectable, and Multi_Zone_Error_Rate had increased above zero showing that there were physical issues so I returned it for a warranty replacement.

You can also make it test the whole drive by running this:

# smartctl -t long /dev/ada0

And then check the output of the first command again after a while to see the results. In my case it showed an LBA read error after about 90% had completed.

 

[wblock@]

The standard fix is to boot in single user mode and run
# fsck -y -tufs

However: the problem going away is bad. Intermittent problems are almost always hardware. If it's memory or a power supply going bad, running fsck(8) could make things much worse. Last week I had a machine start acting weird, errors when trying to compile a port, always different. One DIMM had just suddenly failed.

So disconnect the drive and test the hardware. Do the easy tests first, like taking out some of the memory, or swapping with known-good parts. Move the drive into another machine and make a backup of it first, then run the SMART tests xtaz suggests.