topic Re: Logical Volumes and Service Guard in Operating System - HP-UX

Logical Volumes and Service Guard

Dario_1 — Mon, 22 Apr 2002 17:53:34 GMT

If I have a RAID5 LUN created with 3 different disks in a SG environment. Can I create a Logical Volume and assign it to one of the disk within the LUN? or the system is looking at the LUN as one big disk.
Additional Info:
Our FC60 disk array consitsts of two primary components, the FC60 controller enclosure and the HP SureStore E Disk System SC10 enxlosure.

Re: Logical Volumes and Service Guard

hpuxrox — Mon, 22 Apr 2002 17:57:04 GMT

I am not completly sure what your asking,

They systems will need to see all three disks to be useable.

Re: Logical Volumes and Service Guard

A. Clay Stephenson — Mon, 22 Apr 2002 18:01:15 GMT

If I understand your question, MC/SG (as does the underlying LVM that it relies upon), sees this as one device. A volume group can be made up of any number of LUN's (or in the case of simple disks - any number of disks) but for your purpose each LUN is a disk to the host regardless of the actual number of physical disks within the array that make up the LUN.

Re: Logical Volumes and Service Guard

Jeff Schussele — Mon, 22 Apr 2002 18:07:59 GMT

Hi Dario,

If I understand your qestions correctly - I'll answer the 2nd one first.
Yes the systems sees the LUN as one disk.
Then that would imply that you can't allocate a new LV to JUST one of the disks in the existing LUN.
BUT depending on how you've allocated the disks to the LUN. IF just part of the disks (all three) have been allocated to the LUN it may be possible to create another LUN & allocate that to the new LV.
Believe it will have to be RAID5 also.
All Service Guard deals with are Volume Groups & Logical Volumes - it doesn't really know how they are created or configured. It just needs to be able to mount/unmount the FS & activate/deactivate the VG.

Rgds,
Jeff

Re: Logical Volumes and Service Guard

Dario_1 — Mon, 22 Apr 2002 18:39:15 GMT

Thank you for your help. Points have been assigned.

Regards,
Dario