Anyone seeing issues with BCV on HPUX Itanium?

No. It does not mount using mount command.

Tried running fsck to address lvol and got following error:

vxfs fsck: V-3-25290: could not red from block offset devid/vlknum 0/88309758. Device may be missing in vset.

vxfs fsck: V-3-20694: cannot initialize aggregate

file system check failure, aborting ...

What I did was to manually increase the volume group's lvol on the BCV side and run vgcfgbackup.  Then I did a resync (full) of all the disks associated to that lvol, after which I ran fsck on the lvol (fsadm -F vxfs -y -o full /dev/vg/lvol) and it ran at that point.  I was able to mount up the BCV side of the file system.

But.......

When it goes to run again the full BCV refresh cycle - once again it fails on same lvol.  When you re-display the volume group - the settings are back to the old values for this lvol.  As though I had never manually reset them.

/rcw

I am going to close this discussion, since I found a resolution.

Basically, from an HPUX perspective, I forced it back to basics.

The disk increase came from a disk already in the volume group, so it should have been a simple thing to increase the size on the primary/host side.  Since there really was not change that would require a mapfile update, the increase should have been reflected at the next refresh.  But it wasn't quite working out right.

So I took the BCV side fully back to ground zero.  I dismounted the whole group of disks for the whole disk group.  I made all the affected volume groups inactive.  I them pvcreated every last BCV disk involved and rebuilt the BCV side (using the same disk(s) ) over again, as if doing it for the first time.  Ran fsck on the lvols and everyone ran clean, including the one give me headaches.  Mount file systems manually and they all mounted up clean.

I then did  manual run of the script we use for automation, that will dismount again, de-activate volume groups, run an incremental sync and remount the BCV's again.  It ran clean.  Ran it again - it ran clean.

So it appears that 'maybe', since I'm mixing sync'ing up from PARisc primary side to an Itanium BCV side, something is just not handling simple increases at the LVM level.  I know there were changes to the LVM structure when it hit Itanium, but didn't think it would affect it like this.

I have no issues if I create a new lvol and attach disk & make the required adjustements.  The next refresh cycle automation script runs fine.

It is only with simple increases to an lvol (even with existing available disk) that I'm seeing this.

The resolution > redo the entire BCV creation process as if you were setting it up for the very first time.

Regards,

Rita

Maybe it has to do with vxfs filesystem version mis-match of some sort?

Can you tell the vxfs versions on both sides i.e. standard and BCV hosts?

Actually the file systems mount at the same version level.

Rita,

I'm not a HP-UX expert but maybe the filesystem information is still being cached on the host that mounts the standard device; Issuing sync command to flush the buffers then doing incremental establish/split again to pick up the change might help.