Very odd messages in /var/log/messages

[redsevens]

Does anybody know what this means?
My message log (/var/log/messages) stopped logging files last week. I found out and was able to simply run syslogd (it wasn't listed when I did a ps -aux) and get it running again. However, after running it, I noticed that every 20 minutes after that, there was an entry in the log labled -- MARK -- and every day at about 4 am there would be a sequence of messages like this:

04:02:00 kernel:  failed
04:02:14 kernel: attempt to access beyond end of device
04:02:14 kernel: 02:00: rw=0, want=14, limit=4
04:02:14 kernel: dev 02:00 blksize=512 blocknr=27 sector=27 size=512 count=1
04:02:14 kernel: Directory sread (sector 0x1b) failed

And that goes on for about 30 seconds, logging on for many pages. I decided to reboot and run an fsck, but was told at bootup that the filesystems are all clean, and since the reboot all has been ok (granted it's only been a few hours, but there are no "MARK" lines yet).

Any guesses as to what happened, and if I need to do anything to prevent it from happening again? I am guessing the kernel messages are related to a cron job, but the only cron.daily's I have are the typical ones, logrotate, anacron, tmpwatch, etc. whatever is the default items for RH6.2. This caused no damage that I have found, all web and email services are still running. But I like to know when funky stuff is happening on my servers.

-Joe

 

[chenn]

Things like that happen sometimes with devices, like floppies and hd's. An sread error is a sector READ error - indicating crashed disks. You should try to find out which device is dev 2:00 or consider trying another harddisk.
I suggest backing up everything on your disks immidiately!
I'm not sure if fsck can fix such errors as they're not filesystem but hardware errors... Good luck!
regards
chenn

 

[redsevens]

Well, since the reboot, there have been no such other messages like that. However, for future reference, how would I figure out what drive is 02:00? And does anybody know what the

-- MARK --

lines mean? They were happening every 20 minutes, and they have also since disappeared after the reboot. Taking the slightly paranoid approach, I installed a few security things, namely portsentry and hostsentry, and changed the root password.

 

[ahab]

Weird, I got the same sread error at the same time - 4:02 am. Same device too. Hmmm. Device 02:00 seems to be floppy. Further up in the log it says:

Feb 26 04:02:15 6WFDC01 kernel: floppy0: disk absent or changed during operation
Feb 26 04:02:15 6WFDC01 kernel: end_request: I/O error, dev 02:00 (floppy), sector 19
Feb 26 04:02:15 6WFDC01 kernel: Directory sread (sector 0x13) failed

Some sort of automatic disk checking thing is looking for a floppy that isn't there. That's what I think.