SOS - 3310s not Partioning the same - partitions, logical drives,

From: Sammy (mesammy@gmail.com)
Date: Tue Sep 27 2005 - 20:56:16 EDT


Hello all,

I posted yesterday about deleting metasets and have been working
non-stop since (hence lack of feedback). We have 2 times 2 node Sun
Cluster 3.1's - two 3310s both to two v490s.

We eventually stored the failing cluster from scratch.

On the working cluster, the 3310s are Raid 5'd and then 3 partitions
made on the logical drive, appearing as 3 (x2) disks to the v490s.
The partition tables on the working cluster look like this, and are
created automatically by the 3310s:

Total disk cylinders available: 34873 + 2 (reserved cylinders)

Part Tag Flag Cylinders Size Blocks
  0 usr wm 3 - 34872 68.11GB (34870/0/0) 142827520
  1 unassigned wm 0 0 (0/0/0) 0
  2 unassigned wm 0 0 (0/0/0) 0
  3 unassigned wm 0 0 (0/0/0) 0
  4 unassigned wm 0 0 (0/0/0) 0
  5 unassigned wm 0 0 (0/0/0) 0
  6 unassigned wm 0 0 (0/0/0) 0
  7 usr wu 0 - 2 6.00MB (3/0/0) 12288

Whilst the problem cluster looks like this:

Total disk cylinders available: 34873 + 2 (reserved cylinders)

Part Tag Flag Cylinders Size Blocks
  0 root wm 0 - 63 128.00MB (64/0/0) 262144
  1 swap wu 64 - 127 128.00MB (64/0/0) 262144
  2 backup wu 0 - 34872 68.11GB (34873/0/0) 142839808
  3 unassigned wm 0 0 (0/0/0) 0
  4 unassigned wm 0 0 (0/0/0) 0
  5 unassigned wm 0 0 (0/0/0) 0
  6 usr wm 128 - 34872 67.86GB (34745/0/0) 142315520
  7 unassigned wm 0 0 (0/0/0) 0

Why the difference???
We have tried redoing the 2nd cluster a number of times, using cli and
ssconsole to make the partitions on the logical drive, but it always
comes up the same, with swap and root partitions being made....
How did the first cluster manage to end up with such a simple partition???
Is there something wrong with the first cluster??? - it has no backup
partition - this is like a soft partition. But all we did was split
the logical drive on the 3310 firmware in the only way that it offers.

Yours,
completely baffled and sleep deprived,

Sammy and Hugo.

p.s. we could maybe fmthard the 2nd cluster disk to be like the first
- but the systems have to be identical, and they are not identical if
one has to be fmtharded and the other not - aaarrrgghhh!
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers



This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:31:50 EDT