Disk Cache issue?

[warmongr]

Note the question mark at the end of the subject line. Here is the situation. I have several machines booting from a storage area network device. Ever once in awhile we come in to find a dead machine that either complains about inodes being out of whack or with a corrupt superblock. I have to admit not having alot of experience troubleshooting the ext3 filesystem and all of my efforts to do so with debugfs or dumpe2fs result in Bad magic number in super-block while trying to open /dev/hda3. What are the tools to troubleshoot ext3?

Thanks,
War.

 

[zeland]

Have you tried scanning the filesystem fsck?


--== Anything can go wrong. It's just a matter of how far wrong it will go till people think its right. ==--

 

[Annihilannic]

I have several machines booting from a storage area network device.

You have several machines booting from the same device? If they are all accessing that device read-write then it is no wonder you are getting filesystem corruption; if you want to do that you need to use a type of filesystem that is specifically designed for concurrent use by multiple systems. I'm not sure if there are any which will also allow you to boot from them. Red Hat's GFS is one example of them, and I think Veritas do a cluster version of VXFS.

Annihilannic.

 

[warmongr]

Anni:

The SAN is set up as a RAID V (16 disks) with (2) Virtual Volumes (6 disks and 2 hot spares per vv). Each VV is then assigned via falconstor target/initiator pairing. In other words each machine is assigned it's own boot drive, they are not all booting from the same virtual volume/disk.

Zeland, I have tried fsck but what I'm after are tools comparable to the ext2 tools that allow you to dumpe2fs and look for magic numbers etc. Any advice there?

Thanks,
W

 

[Annihilannic]

The ext2 tools *should* work fine with ext3, ext3 is just ext2 with journalling.

Surely they should be appearing as scsi devices, i.e. /dev/sda3 instead of /dev/hda3?

Do the tools report "bad magic number" when you try to use them on a healthy filesystem?

Are you booted from another device when attempting to do this (e.g. CD-ROM)?

Annihilannic.

 

[warmongr]

Your correct that they are showing up as scsi devices. I just used hda as an example. I have resolved my issue which was stupid. Ok really stupid. I was ssh'd into another machine and trying to access /dev/sda3 when the box I was ssh'd into was direct attached at /dev/hda3. Forgive and thanks for the input. But since we're on the subject if I do get a bad superblock error again, how do I boot from an alternate and copy the good superblock to the bad one. dumpe2fs /dev/sda3 |grep superblock...

... now what?

N

 

[Annihilannic]

See the man page for fsck.ext2, specifically the -b option.

Annihilannic.

 

[warmongr]

Thanks for your help.

War