This is a discussion on failed primary and mirrored root disks: BAD SUPER BLOCK within the Sun Solaris Hardware forums, part of the Solaris Operating System category; --> I have a Sun Enterprise 3500 running Solaris 2.6 and Veritas Volume Manager 3.0.4. I have two boot disks ...
| |||||||
| FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
| ||||
| I have a Sun Enterprise 3500 running Solaris 2.6 and Veritas Volume Manager 3.0.4. I have two boot disks that are mirrored via Veritas. The first boot disk failed a few days ago. The error was "rootdisk can't boot Please boot from one of the following disks: Disk: rootmirror Device:c0t1d0s2". I tried several ways to get it to boot from the second rootmirror disk, but nothing worked. It would report error "boot: cannot open c0t1d0s2" or "boot: cannot open rootmirror". Appeared to be a hard failure, so I removed the failed primary boot disk out of slot 0, pulled the mirrored rootdisk from slot 1 and put into slot 0. Luckily the system started fine from the roomirror disk at this point...at least for a while. About two days later, while I was trying to locate a spare disk to replace the failed primary boot disk, I started getting write errors to the rootmirror disk dumped onto the console. I shut the system down, tried to reboot, but when it came up I received error "Boot disk failed. The file just loaded does not appear to be executable." Some googling showed that this error occurs when the default bootblock has been corrupted, so I started to follow the procedure to restore it. I booted to singleuser mode from the solaris CD, and tried to run fsck on the rootmirror disk, but then received error "BAD SUPER BLOCK: MAGIC NUMBER WRONG USE AN ALTERNATE SUPER-BLOCK TO SUPPLY NEEDED INFORMATION". I tried alternate b=32, same error. I ran "newfs -N..." to get the rest of the superblocks, tried at least 20 different values from the beginning, middle, and end of the list, but still get the same BAD SUPER BLOCK error. I can run format and see that the label and partition information are correct. I also rewrote the partition and label information to the disk, which works without any reported errors. But subsequent fsck still fails with same error. I tried fsck against the entire disk (s2) and the individual partitions (root on s0, swap on s1, usr on s6), but all get the same error, even with alternate super block values. Why would I be able to label and partition the disk without error, but not be able to fsck? I guess its possible that both disks have hard failures, but withing a few days of each other seems improbable. Any ideas on what else I can do from here short of building up a new boot disk from scratch? Sadly, there is no backup of the root disk. Yeah, I know about backups... TIA for any assistance. |