SUMMARY: SUN SE 3510 FC does not present backup slice

From: Qing Chang <qchang_at_sten.sunnybrook.utoronto.ca>
Date: Mon Dec 01 2003 - 15:44:07 EST
Hi Managers,

sorry for the long delay of this summary. We tried
different options and ended up with the solution
presented here.

First thanks to Joe Moore and Darren Dunham for their
valuable tips on the Tbyte limit on Solaris below
Solaris 9 8/03.

The problem of not being able to add a quarum device for
Sun Cluster 3.0 is directly related to how Sun deal with
disk bigger than 1 Tbyte: They have introduced
an Extensible Firmware Intaerface (EFI) to label disk
begger than 1 TB. Compare to VTOC scheme
of labeling disk, the difference is slice 2 is regarded
as a NORMAL slice. It DOES NOT represent the whole disk.

On the other hand, SunCluster 3.0 insists to look for a
VTOC style of slice two to determine if the disk device
can be used as a quorum device.

We did not get a definite anwser if SC 3.1 support EFI
from SUN. At last we decided to make smaller RAID sets
to satisfy SC 3.0.

Note with disk size smaller than 1 TB Solaris uses VTOC
automatically.

About multiterabytes support, Solaris 9 earlier than
release 8/03 supports multiterabytes disk and filessystem
as long as the system is patched with MU4.

Qing Chang

====== original question ========
> Hi Managers,
> 
> I have a SUN SE 3510 FC Array with dual RAID controllers
> connected to two V880 servers with identical configuration
> to build a HA solution with Sun Cluster 3.0 5/02:
> SunOS server1 5.9 Generic_112233-08 sun4u sparc SUNW,Sun-Fire-880
> SunOS server2 5.9 Generic_112233-08 sun4u sparc SUNW,Sun-Fire-880
> 
> The firmware vesion of the Array is 3.27K.
> 
> Two RAID 5 sets are created with 1.33TB and 1.47TB in size
> on the Array.
> 
> The servers can see the drives no problem, but there is
> no backup slice (slice 2) created by default, see prtvtoc
> output for the RAID set with 1.33TB below:
> 
> ========= begin prtvtoc =========
> * /dev/rdsk/c3t44d0s2 partition map
> *
> * Dimensions:
> *     512 bytes/sector
> * 2862141440 sectors
> * 2862141373 accessible sectors
> *
> * Flags:
> *   1: unmountable
> *  10: read-only
> *
> *                          First     Sector    Last
> * Partition  Tag  Flags    Sector     Count    Sector  Mount Directory
>        0      2    00         34    262144    262177
>        1      3    01     262178    262144    524321
>        6      4    00     524322 2861600700 2862125021
>        8     11    00  2862125022     16384 2862141405
> ========== end prtvtoc ========
> 
> The real problem is Sun Cluster insist to use slice 2 to
> add a quorum device, following is the error when adding
> a quorum device with above RAID set:
> ========== error begin ========
>     Which global device do you want to use (d<N>)?  d17
>     Is it okay to proceed with the update (yes/no) [yes]?
> scconf -a -q globaldev=d17
> Sep 26 14:14:34 server1 cl_runtime: WARNING: CMM: Open failed with error 
> '(No such device or address)' and errno = 6 for quorum device 
> '/dev/did/rdsk/d17s2'
> scconf:  Failed to add quorum device (d17) - unable to scrub the device.
> 
> Sep 26 14:14:34 server1 . Unable to scrub device.
> Command failed.
> =========== error end ========
> 
> I tried to rebuild partition table and create a backup slice
> manually, but Sun Cluster does not recognize it. In fact even
> format does not recognize it, it is considered a regular
> slice ocuppying the whole disk, although I set the Tag(5)
> and Flag (01) right for a backup slice.
> 
> Can anybody help me out on this? I'll summarize.
> 
> Thanks,
> 
> Qing Chang
> _______________________________________________
> sunmanagers mailing list
> sunmanagers@sunmanagers.org
> http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
Received on Mon Dec 1 15:49:29 2003

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:43:23 EST