Help to understand syslog messages

[Laszlo Csizmadia]

I have found lot of new entries in my syslog which I can't really interpret:
Aug 27 10:49:28 geleclf1 vmunix: SCSI: bp: 0000000049562400
Aug 27 10:49:28 geleclf1 vmunix: dev: cd0d3640
Aug 27 10:49:28 geleclf1 vmunix: cdb: 4d 00 6e 00 00 00 00 01 44 00
Aug 27 10:49:28 geleclf1 vmunix: residual: 144
Aug 27 10:49:28 geleclf1 vmunix: status: (400) None -- Incomplete
Aug 27 11:00:26 geleclf1 vmunix:

Could somebody explain me what are these? How to find out which device is generating these?
Thanks in advance.

[Bruno Vidal · Accepted solution]

Hi,
it seams that one of your SCSI device has some trouble. The "dev" is like this:
two first digit are the major of the driver
the other the minor.
So in your case: cd -> 205 -> stape
0d3640 is the minor so it is:
c13t3d6 with option 40 -> for stape it is no rewind option -> so look at the tape that has his ext_bus with id 13, at scsi address 3 and lun 6.

Cheers.

[Tim Adamson_1]

dev: cd0d3640

cd is hex for decimal 205

lsdev | grep 205 ---> tape

So its a tape drive

0d3640 ---> c13t3d6

Do you have a tape drive in /dev/rmt named
c13t3d6BEST or something like that?

An ioscan would be handy too.

Anyway, thats my analysis!


Tim

[Cheryl Griffin]

>>vmunix: status: (400) None -- Incomplete >>
Makes me think you have device files but the hardware is no longer responding.

cd0d3640 --> c13t3d6

# lssf /dev/rmt/c13t3d6
# diskinfo -v /dev/rmt/c13t3d6

Did you have hardware at this location but removed it? If so remove the devices with rmsf.

If you have hardware at this location that was not removed, check that it is repsonding with:
# ioscan -fnC tape

And physically check that the drive is still connected, operational, etc.

[Tim D Fulford]

Hi

You might also try using "mstm", this may give you some additional clues as to why it did not work.

Usually SCSI resets are indicative of some hardware problem!!

Regards

Tim

[Steven E. Protter]

In addition:

I'd look at ioscan | more for possible unclaimed drives.

dmesg | grep lbolt

for eventual scsi timeouts.

If you have X-Windows access to the box xstm is point and click user friendly(instead of mstm).

You might also do this:

strings /etc/lvmtab

with the output of that:

pvdisplay

If a disk is gone, you'll get a nasty error message.

Be agressive and spot this before it brings yoru system down. Depending on configuration, this could be quite serious.

SEP