Hi,

at first I would like to thank you giving me ability to play with Solaris 
Cluster :)

Brief intro: I'm at process of playing with Solaris Cluster Express 07/07(nv68) 
on x86_32. I have dual node cluster with third node (also x86_32 Solaris , but 
nv67) as a iSCSI target. SCSI-2 reservations are used on quoum builded from 
iSCSI LUN(iscsitadm supports SCSI-2 reservations, not the PGR = SCSI-3 ones for 
>= 3 nodes ).

I've almost had working Oracle RAC cluster on 10U3 eariler with Sun Cluster 3.2 
with 32-bit addon sent by Sun engineer, but due to ISCSI/VHCI bug that caused 
panic during reconfiguration I've switched to Solaris Cluster Express(the bug 
was fixed in nv67, there is no patch for this one for 10u3)..... as this is 
proof-of-concept, it is not going to be used in production env.

On 10U3 there were no problems running metainit after having multiowner metaset 
configured; I'm getting "unit not found" bug as described in FAQ and while 
trying to use 
workaround(http://opensolaris.org/os/community/ha-clusters/ohac/Documentation/relnotes/;jsessionid=A2C75F76D9693A4F7080D600C61E9B5A,
 sic!) kernel segfaulted...

> ::status
debugging crash dump vmcore.0 (32-bit) from solrac2
operating system: 5.11 snv_68 (i86pc)
panic message: forced crash dump initiated at user request
dump content: kernel pages only
> $C
e246ff18 vpanic(fea41b34)
e246ff4c kadmin+0xcb(5, 1, 0, e5c7ea98)
e246ff84 uadmin+0x89()
e246ffac sys_sysenter+0x106()

System is an 2xCPU SMP on VMware, total 4 CPUs(2 for each VM). Are there any 
chances for successful configuration of metainit or should I just setup Oracle 
RAC using pure /dev/did ?

Also http://bugs.opensolaris.org/view_bug.do?bug_id=6569237 is not published, 
right?

Help, appreciated...
--

This message posted from opensolaris.org


Reply via email to