VXVM to LVM conversion

[Michael Steele_2]

Lots out there for lvm to vxvm but I need a vxvm to lvm procedure.

[IT_2007]

Do you mean to LVM to VxVM conversion or otherway round? I saw your another post asking about disk addition after LVM to VxVM convert.

[Mridul Shrivastava · Accepted solution]

Note: It is always recommended that prior to making any changes to the system to create an Ignite backup, fbackup and/or backup your system in the event you need to recover your system back to a usable stable state. You may obtain the latest Ignite product and associated PAX patch at the following URL:
http://www.docs.hp.com/en/IUX/

1) - Create another Ignite tape using make_tape_recovery â x inc_entire=vg00 â p. This will be the tape you will use to migrate from VxVM to LVM.

2) - Edit the /var/opt/ignite/recovery/latest/system_cfg file to remove the vxvm Disk/Filesystem Layout and replace it with LVM Disk/ Filesystem Layout.

3) - Continue creating the Ignite tape using make_tape_recovery â r.

4) - Recover the system using the new Ignite tape with the LVM Disk/Filesystem Layout.


Editing the system_cfg file:
The only differences in the /var/opt/ignite/recovery/latest/system_cfg files between VxVM and LVM Disk/ Filesystem Layouts is the portion titled â Disk and Filesystemsâ . You will also notice that the order of the logical volumes in the VxVM system_cfg file do not reflect the numerical order you would see in the LVM system_cfg file. There is no need to reorder the lvols. Just make sure to use the correct minor number for each lvol especially lvol1, lvol2 and lvol3.

1) make_tape_recovery â x inc_entire=vg00 -p
2) cd /var/opt/ignite/recovery/latest
3) cp system_cfg to system_cfg.orig
4) vi system_cfg
5) remove the line VxVM_hostid=â loopbackâ
6) change volume_group=â rootdgâ to volume_group=â vg00â
7) change the usage=â VXVMâ to usage=â LVMâ
8) change logical_volume â standvolâ to logical_volume â lvol1â
9) add line bad_block_relocate=false just after largefiles=false line
10) add line minor_number=0x01 just after stripe_size=0
11) Follow the same steps for swap and root lvols.


On the remaining lvols;
12) Edit the lvolname to reflect the default logical volume number.
13) Add line bad_block_relocate=true just after largefiles=false line â ¦ note that the bad_block_relocate line is false for lvol1 through lvol3
14) Change contiguous_allocation=true to contiguous_allocation=false
15) Add the minor_number=0x0# to reflect the proper minor number for that logical volume.


Defaults:
Mountpoint Logical Volume Minor Number
/stand lvol1 0x01
swap lvol2 0x02
/ lvol3 0x03
/tmp lvol4 0x04
/home lvol5 0x05
/opt lvol6 0x06
/usr lvol7 0x07
/var lvol8 0x08

[Michael Steele_2]

Yes, this is the 2nd half of the test in order to get around LVM's max_pe limitations. First convert to vxvm inorder to add in the 50gb lun and then go back to lvm from vxvm. The first half looks good with vxvmconvert but the second half looks like a data migration from tape or cp'ing between file systems.

The problem with ignite, which is a minor one, is the pax backup and restore of the data. I'd just like to migrate over without copying or restoring like vxvmconvert does when going from lvm to vxvm. But there doesn't seem to be an equivalent.

[Kevin Wright]

why not just re-create the VG with the right max_pe size from Ignite (if it's vg00) or destroy and recreate it if it's not vg00?

[Michael Steele_2]

,