Moving from HP-UX 11i v2 to HP-UX 11i v3 |
 |
If you are migrating a system from HP-UX 11i v2
to HP-UX 11i v3, please refer to the LVM migration white paper described
in the “Related Documentation” section.
It contains information on migrating an LVM configuration from the
legacy naming model to the agile naming model.
Existing LVM configurations created on HP-UX 11i
v2 continue to work on HP-UX 11i v3 under the legacy naming model.
However, there is a change in behavior for Alternate
Links (PVLinks):
In HP-UX 11i v3, management of multi-pathed devices
is available outside of LVM using the next generation mass storage
stack. By default, the next generation mass storage stack distributes
I/O requests across all available paths to a multi-pathed disk, even
when using legacy device special files. Using LVM with persistent
or legacy device special files may cause I/O requests to be sent across
alternate links, even if the links are not configured as PVLinks;
this should not introduce any errors, but it does differ from PVLink
behavior in previous releases.
HP recommends converting volume groups with multi-pathed
disks to persistent device special files and native multi-pathing,
as described in the migration white paper in “Related Documentation”.
However, if you want backward-compatible PVLink
behavior, you must use legacy device special files for physical volumes,
and disable the mass storage stack multi-pathing for those physical
volumes. To disable multi-pathing on legacy device special files,
use the scsimgr command to configure a global device
tunable called leg_mpath_enable.
For each multi-pathed disk, enter the following
command:
# scsimgr save_attr -D /dev/rdisk/diskn -a leg_mpath_enable=false |
Alternatively, you can disable multi-pathing for
all legacy device files with this command:
# scsimgr save_attr -a leg_mpath_enable=false |
Note that this has no effect on multi-pathing
through persistent device special files. For more information, see scsimgr(1M).
Printable version
