Replacing a failed root disk

Rikki hinn Ogurlegi — Tue, 13 Nov 2012 04:07:25 GMT

I'm having problems replacing a dead root/boot disk in a 11iV3 system. The machine in question is a BL860c.

[root@blesi ~]# model
ia64 hp server BL860c
[root@blesi ~]# uname -a
HP-UX blesi B.11.31 U ia64 3117009088 unlimited-user license

[root@blesi ~]# strings /etc/lvmtab | head -3
/dev/vg00
/dev/disk/disk2_p2
/dev/disk/disk3_p2

disk2 is OK, disk3 has failed. Thanks to mirrordisk the box is still running without issue. I removed the failed disk (Hotswap SAS) and put in a new one. But here 11iV3 bit me in the behind. In previous releases of HP-UX the OS just detected a new disk with the same HW path and all that was needed was to just restore partitions and VG conf and you where done. However, this machine detected the disk as a new entity, with a new HW path, even tho it was inserted into the same physical slot on the machine as the failed disk was. So I consulted Google and found out that this is normal for 11iV3 and found documentation to replace the disk. But I cant say it has helped me much.

More info:

ioscan -m lun:

disk      3 64000/0xfa00/0x1    esdisk NO_HW       DEVICE       offline HP      DG146BB976
             0/2/1/0.0x5000c5000ab2a89d.0x0
                      /dev/disk/disk3      /dev/rdisk/disk3
                      /dev/disk/disk3_p1   /dev/rdisk/disk3_p1
                      /dev/disk/disk3_p2   /dev/rdisk/disk3_p2
                      /dev/disk/disk3_p3   /dev/rdisk/disk3_p3

disk    472 64000/0xfa00/0x78   esdisk CLAIMED     DEVICE       online   HP      DG146BB976
             0/2/1/0.0x5000c5000ab287a9.0x0
                      /dev/disk/disk472   /dev/rdisk/disk472

old disk is disk3, new disk is disk472.

[root@blesi ~]# diskinfo /dev/rdisk/disk472
SCSI describe of /dev/rdisk/disk472:
             vendor: HP
         product id: DG146BB976
               type: direct access
               size: 143374744 Kbytes
   bytes per sector: 512

Looks right. Now, according to the docs I need to let the storage subsystem know about the change:

[root@blesi ~]# scsimgr replace_wwid -D /dev/rdisk/disk3
scsimgr:WARNING: Performing replace_wwid on the resource may have some impact on system operation.
Do you really want to replace? (y/[n])? y
scsimgr: Successfully validated binding of LUN paths with new LUN.

[root@blesi ~]# io_redirect_dsf -d /dev/disk/disk3 -n /dev/disk/disk472
Number of old DSFs=8.
Number of new DSFs=2.
The number of old and new DSFs must be the same.

This confuses me. It's almost like the OS wants me to partition the disk as disk472 and then do the redirect but thats nowhere to be seen in the docs.

How should I proceed here?

Thanks in advance.

P.S. This thread has been moved from Integrity Servers to HP-UX > sysadmin - HP Forums Moderator

Re: Replacing a failed root disk

Todd McDaniel_1 — Fri, 16 Nov 2012 19:03:11 GMT

If the issue is that it has a new hardware device name, first I would blow away the old disk3 devices.

Then, also blow away the new ones it created when you inserted the new disk, then run ioscan again.

if that won't work, consult your HP Support.

Regards,

Todd McDaniel

topic Re: Replacing a failed root disk in Operating System - HP-UX

Replacing a failed root disk

Re: Replacing a failed root disk