scsimgr replace wwid

[joseph51]

Hi

I have seen the following error in the syslog . I know this happened when the new disk which we added and removed later which had the device files

The OS is 11.31

Nov 18 10:24:14 prod1 vmunix: Evpd inquiry page 83h/80h failed or the current page 83h/80h data do not match the previous known page 83h/80h data on LUN id 0x0 probed beneath the target path (class = tgtpath, instance = 38) The lun path is (class = lunpath, instance 601).Run 'scsimgr replace_wwid' command to validate the change

its repeating with instance 602,603,604

I removed the device files with the following steps

root@prod1:/usr/local/bin : ioscan -fnC disk | grep -i no_hw
disk    704  2/0/12/1/0/4/0.1.18.0.0.1.6   sdisk   NO_HW       DEVICE       IBM     2145
disk    702  2/0/12/1/0/4/0.1.50.0.0.1.6   sdisk   NO_HW       DEVICE       IBM     2145
disk    703  3/0/12/1/0/4/0.2.2.0.0.1.6    sdisk   NO_HW       DEVICE       IBM     2145
disk    700  3/0/12/1/0/4/0.2.34.0.0.1.6   sdisk   NO_HW       DEVICE       IBM     2145
root@prod1:/usr/local/bin : lssf /dev/rdisk/disk701
esdisk section 0 at address 64000/0xfa00/0xc8 /dev/rdisk/disk701
root@prod1:/usr/local/bin :

root@prod1:/usr/local/bin : rmsf -H 64000/0xfa00/0xc8
root@prod1:/usr/local/bin : ioscan -fnC disk | grep -i no_hw

IS this enough to clear the error and correct , I have seen in some other servers some scsimgr command issued before removing the device files

Like this

scsimgr -f replace_wwid -C lunpath -I 600

What exactly this does , and what is the use of this command and what is happening in background

[Torsten.]

I'm sure you just removed the NO_HW items, but you probably did not get them back. The scsimgr command allows the system to access a device after a replacement or change. For example a disk or tape drive will come with a new WWN, so the systems prevents access for safety reasons. scsimgr tells the system to not do this check and access it anyway.

[joseph51]

when I checked the status of instance again , I could see one is as AUTH Failed
root@prod1:/ : scsimgr get_info -C lunpath -I 600

STATUS INFORMATION FOR LUN PATH : lunpath600

Generic Status Information

SCSI services internal state = UNOPEN
Open close state = AUTH_FAILED
Protocol = fibre_channel
EVPD page 0x83 description code = 1



root@prod1:/ : scsimgr get_info -C lunpath -I 603

STATUS INFORMATION FOR LUN PATH : lunpath603

Generic Status Information

SCSI services internal state = UNOPEN
Open close state = STANDBY
Protocol = fibre_channel



root@prod1:/ : scsimgr get_info -C lunpath -I 602

STATUS INFORMATION FOR LUN PATH : lunpath602

Generic Status Information

SCSI services internal state = UNOPEN
Open close state = ACTIVE
Protocol = fibre_channel

So I need to issue the replace_wwid command only for 600 instance right ?

root@prod1:/ : scsimgr -p get_attr -C lunpath -I 600 -a hw_path
3/0/10/1/0/4/0.0x5005076801205aaf.0x400e000000000000
root@prod1:/ : lssf 3/0/10/1/0/4/0
lssf: 3/0/10/1/0/4/0: No such file or directory
root@prod1:/ : lssf 3/0/10/1/0/4/0.0x5005076801205aaf.0x400e000000000000
lssf: 3/0/10/1/0/4/0.0x5005076801205aaf.0x400e000000000000: No such file or directory
root@prod1:/ :



To be issued

scsimgr -f replace_wwid -H 64000/0xfa00/0xc2 right ?

required for all other , or only this AUTH Failed

[Torsten.]

This is a disk array, right? So what did you change on the array side? Physical disk changes inside an array are usually not seen by a server...

[Torsten.]

If you for example present a LUN to the server, then remove this presentation, delete and re-create the LUN anyhow and present it again to the server, you will see exactly what you currently see. To solve this google "SCSI Management and Diagnostics utility on HP-UX 11i v3" and read page 29 of the PDF.

[binu_mathew]

Thank you torsten