EMC DRIVE

[Nobody's Hero]

I am getting this error in syslog about one of my symmetrix disks. How do I convert this device number to something I can relate too?

vmunix: LVM: vg[0]: pvnum=1 (dev_t=0x1f015000) is POWERFAILED

[Paula J Frazer-Campbell · Accepted solution]

Hi

Look at the device numbers, e.g. 0x1f031100

The first two hex digits (1f) refer is the major device number. 1f = 31 (dec).

Do as lsdev and look for the matching driver. You will find that block major device 31 is "sdisk" - SCSI disk.

The next two hex digits (03) refer to the bus instance number. The next hex digit (1) is the SCSI ID and the next hex digit (1) is the LUN.

The following 2 hex digits are driver specific.

1f031100 decodes to /dev/dsk/c3t1d1.

Do an ioscan -fn and find the host bus adapter that corresponds to c3.


Paula

[Elena Leontieva]

The disk_device can be determined by the using the dev_t
value. For example:

A dev_t value of 0x1c045000 is associated with c4t5d0

04 = instance number
5 = SCSI address number
0 = LUN number

[Ian Dennison_1]

Quick and dirty way,...

vgdisplay -v /dev/vg00 |more

Look for the first PV or notice which disk it howls about.

More professional way,...

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x244768c57f64d4118fee0090279cd0f9,00.html
(Down the bottom of the posting, the answer by Ramesh is good)

Share and Enjoy! ian

[Dario_1]

Hi!

Another way to find out is:

ll /dev/dsk

and look for the number 0x1f0150000

Regards,

Dario

[Stefan Farrelly]

These messages are often very misleading. The only way to be SURE you find the device is to check the diagnostic hardware logs.

1. run xstm (GUI)
2. go to TOOLS -> UTILITY -> LOGTOOL
3. select raw log
4. go to bottom of log, go backwards and check all hardware errors for time when you got the error.

Often you may see a single error for one device first and then errors following for other devices. Usually its the first, or the most common (ie. the largest) number of errors against a device which is the culprit.

[Zeev Schultz]

btw how many pvlinks do you have from the server to symmetrix?it could be either a link (scsi,fc) or a LUN problem (it's a lun 0...).
I'd check powerpath (thats how EMC call it?:))
and if one link has failed but another (or more) are ok,would go for cables/terminations.

Zeev

[Rita C Workman]

Well the folks have shown you how to convert the dev info to a more familiar c_t_d_ number....

Now my next thing would be to ask what is the timeout value set for. My guess is that it is still at the default value, so may I suggest that you may need to simply increase it from the default value (what is that 30seconds or 60 seconds)....anyway I set mine to 180
pvchange -t 180 /dev/dsk/c_t_d_

Now to change all your disks to a timeout of 180 you could:

for d in `ls /dev/dsk/*`
do
pvchange -t 180 $d
done

Rgrds,
Rita

[Chris Vail]

Here's a quick-and-dirty shell script to convert the hex code to the cXtXdX format. It does a little error checking, but not much:

#!/bin/ksh
Z=$1
I=`echo $Z|tr '[a-z]' '[A-Z]'`
LENGTH_I=`echo $I|wc -m`
if test "$LENGTH_I" -eq 9
then
A=`echo $I|sed s/..//`
B=`echo $A|sed s/....$//`
C=`echo "ibase=16;$B"|bc`

R=`echo $I|sed s/....//`
S=`echo $R|sed s/...$//`
T=`echo "ibase=16;$S"|bc`

F=`echo $I|sed s/.....//`
G=`echo $F|sed s/..$//`
D=`echo "ibase=16;$G"|bc`

echo "This translates to c"$C"t"$T"d"$D" "
echo

else
echo "The input string must be exactly 8 bytes. "
fi