CAPM repository server got re-booted and now we are not able to start the data base.
I tried to rollback the data base but unable to find valid epoch number.
When you use admintools to start the db, it opens the command screen and it immediately closes.
Release : 3.x
In /var/log/messages we see these errors:
Apr 14 08:30:20 gaalpltapp0015b kernel: EXT4-fs error (device dm-4): ext4_mb_generate_buddy:758: group 592, block bitmap and bg descriptor inconsistent: 14073 vs 14121 free clusters
Apr 14 08:30:20 gaalpltapp0015b kernel: JBD2: Spotted dirty metadata buffer (dev = dm-4, blocknr = 0). There's a risk of filesystem corruption in case of system crash.
Apr 14 08:30:21 gaalpltapp0015b kernel: EXT4-fs error (device dm-4): ext4_iget:4589: inode #4719393: comm vertica: bogus i_mode (0)
There is corruption on the file system.
Have the system admin verify the file system integrity.
Then use admintools to restart the db.
Note: if the file system corruption has damaged the /catalog or /data files, you may have to restore the db from a backup.