getting an err when running script

[Shehan]

Dear All

I am getting below error when running a script. My server is DS10 running tru64 5.1B OS.

/dev/rrz1c: character special (8/1026) EIDE #0 "WD400BB-60DGA0" disk #8 (SCSI ID #1) (SCSI LUN #0) errors = 0/2162

This disk was newly replaced because of restarting issue in the server.It was ok after replace this disk. But we are still getting this error.

Please anybody advice me.

Regards
Nirukshitha

[Rob Leadbeater]

Hi,

> I am getting below error when running a script.

Sharing the script with us would be a good start...

Cheers,

Rob

[Pieter 't Hart]

/dev/rz1c is "legacy" notation.
For Tru64 V5.x you should use the notation /dev/disk/dsk1c

maybe the "/dev/rz1c" special file does not map any more to the replaced disk?

In Tru64 4.x the devicename only depends on the bus/target/lun so rz1 remains rz1 after replacement.

wIn Tru64 %.x when replacing a disk the disk initially gets a new devicename
e.g you got dsk1, dsk2, dsk, dsk4 and replace dsk2, then the OS calls it dsk5.
and you must give some extra commands to put in the "original" place
"dsfmgr -e dsk5 dsk2".

try "hwmrg -view devices" to get the current mapping

Pieter

[Shehan]

Hi

Please find the attached command output.According to this, disk notation is correct.Anyway I already replaced this same disk due to server rebooting when some application is started up. Then we replaced the disk and did the same as you mentioned in your mail.But, we are getting disk error when running health check script.


Regards
Nirukshitha

[Pieter 't Hart]

Like Rob allready asked : what's happening in the script.
is it a custom health-check?


yes we are looking at the same disk /dev/rz1 is pointing to (SCSI ID #1) (SCSI LUN #0) .

# hwmgr -v d
46: /dev/disk/dsk0c Maxtor 6E040L0 bus-0-targ-0-lun-0
47: /dev/disk/dsk1c WDC WD400BB-60DGA0 bus-0-targ-1-lun-0
#
# hwmgr -show scsi
46: 0 STATION_A disk none 2 1 dsk0 [0/0/0]
47: 1 STATION_A disk none 2 1 dsk1 [0/1/0]

seems to me these are IDE disks? and not "real" SCSI disks

at this moment the dik dsk1 is not mounted!
all volumes mounted are on dsk0 or on a remote host (nfm).

maybe the script fails because no disklabel (partitiontable) is present on this disk ?

try "disklabel -r dsk1"
if this returns "no disklabel" do "disklabel -rw dsk1"

[Shehan]

Hi

This is health checking script. Also we had received some SCSI CAM events in the binary error log. Can this be a hard disk issue?

Regards
Nirukshitha

[Pieter 't Hart]

it's still not clear to me what script you are running.
what command do you issue, where is the script located, (maybe post the content of the script) and at what step comes the error?

If you mean the sys_check script, this gives an error on the CD-rom drive when no CD is present/mounted; this also logs CAM/SCSI errors that may be ignored.
it may also occur when a disk has no label.
did you check the disklabel on the new disk as i suggested?

did you view the information in the page that sys_check generates?

[Rob Leadbeater]

Hi,

> This is health checking script

Nobody is going to be able to help you, if you cannot show us the contents of the script !

Cheers,

Rob

[Shehan]

Hi
This is not a our own script. This is one of the customer implemented.So they don't like to share with us. Anyway I will check with them possibility of getting this.

BTW, when we run the disklable command on the same disk, it shows bad sector status. So it can be the issue for this.

Regards
Nirukshitha