AIX

Yes, either that or migratelp or - as i said before - simply remove the LVs copy to empty the disk, then remove the disk, add a new disk to the VG and then remirror.

Actually the migratelp process can read the LP in question - just not its bad copy. And again, you do not need to have this copy at all! Just remove it, then recreate it again on the new disk.

Furthermore, lspv does not show the "online" disks, it just shows the relation of disks and VGs. Do a lsdev -Cc disk to see all disk devices and their status. "Available" is good, "Defined" means missing. It means the system has some device definition that once described an existing device, but this device is not there at the moment. If this does not reveal anything do a rmdev on all hdisk devices, then run cfgmgr to re-add the disk devices again. This is, as bad as it may sound, actually non-disruptive.

I think that, as the disk is alread in the status "missing" in the VGs display, you may have to rely on the procedure to get a missing disk out of the system. Again, this involves some odmget and odmdelete gymnastics which are non-trivial and possibly disruptive too. Do that only when the system is down AND after having made absolutely sure you have a working backup!

I hope this helps.

bakunin

/PS: you may also want to read this old thread for further information

Well, thankfully the system came up, after a remote admin rebooted the box (yes I admit it was my fault for not informing the IT group). IBM support is looking to make sure the hd2 is truly mirrored.
the LVM has somehow corrected itself, or finally bypassed the bad spot for the hd2 now shows sync'd. Thank you so much for your timely help and responses on this. If no one else ever appreciates your help, you have one person who does.

Code:

 rootvg:
LV NAME             TYPE       LPs     PPs     PVs  LV STATE      MOUNT POINT
hd5                 boot       1       1       1    closed/syncd  N/A
hd6                 paging     4       4       1    open/syncd    N/A
hd8                 jfs2log    1       1       1    open/syncd    N/A
hd4                 jfs2       60      60      1    open/syncd    /
0516-1147 : Warning - logical volume hd2 may be partially mirrored.
hd2                 jfs2       40      40      2    open/syncd    /usr
hd9var              jfs2       16      16      1    open/syncd    /var
hd3                 jfs2       20      20      1    open/syncd    /tmp
hd1                 jfs2       40      40      1    open/syncd    /home
hd10opt             jfs2       40      40      1    open/syncd    /opt
hd11admin           jfs2       1       1       1    open/syncd    /admin
livedump            jfs2       1       1       1    open/syncd    /var/adm/ras/livedump
lvol1               jfs2       60      60      1    open/syncd    /usr/sys/inst.images

Just for completeness, did you manage to get hd2 properly mirrored? If so, can you post your command for anyone hitting finding this thread with a similar issue.



Thanks, in advance,
Robin

Initially the mirror was successfully initiated and all the other volumes became mirrored and sync'd. it attached hd2 and was in process of syncing then during the sync it hit a bad sector (sector sparing failed on the hard disk) on the original disk. IBM is saying the VGDA is messed up on the original disk for whatever reason. They are rebuilding the hdisk1 vgda and hopefully this will get the hd2 in sync. One thing that surprises me is hd2 is now reporting open/sync whereas before it is was open/stale before IBM is rebuilding the vgda. Can lvm lie to me? IBM did a vgda map of hdisk1 with dd.

There is a little-known tool called readvgda which does what its name suggests - read the VGDA and print its content out to screen. I use it regularly when working on (to me) unknown systems when i need to find out about the VG type (classic, big or scalable) because otherwise there is no direct way to get that information elsewhere. You can use it for any other purposes too, though. The format is:

Code:

readvgda hdiskN [ | more]

where hdiskN is any hdisk device in the VG. Here is a longer article about it but knowing what to search for you find many others too.

The VGDA is stored in (at least) two different locations of the disk and if one is broken you may still read the other one and use that. You can easily do that with dd so you won't need IBM to do that for you. Of course this is a potentially risky operation so my suggestion is to test it first on a throwaway system/disk to get acquainted.

I hope this helps.

bakunin