Hi,

When the administrator adds an iSCSI device to the system, that device looks
like a SCSI device (at least the Quorum subsystem sees the device as
a SCSI device).

Sun Cluster supports an iSCSI device as a Quorum device when the
iSCSI device is on the same subnet as the cluster nodes.
We do not support an iSCSI device as a Quorum device when
the iSCSI device is on a different subnet. The limitation is due
to the behavior of iSCSI in Solaris 10.

The iSCSI device could be configured for Quorum using
any of the following:

   SCSI2
   SCSI3
   Software Quorum

The iSCSI device that will be used as a quorum device cannot
be a disk that is locally connected to only one cluster node
and exported via iSCSI. Each cluster node must have a direct
path to the device that is not dependent on another node being up.

I am not aware of any limitations caused in this area when
Sun Cluster runs on LDoms.

Regards,
Ellard

On 07/20/09 12:34, Octave Orgeron wrote:
> Thanks for the suggestion. Interested in testing out SC or OpenHA on LDoms 
> and use iSCSI since I don't have a SAN at home. Thanks!
> 
>  *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
> Octave J. Orgeron
> Solaris Virtualization Architect and Consultant
> Web: http://unixconsole.blogspot.com
> E-Mail: unixconsole at yahoo.com
> *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
> 
> 
> 
> ----- Original Message ----
> From: Ashutosh Tripathi <Ashutosh.Tripathi at Sun.COM>
> To: Octave Orgeron <unixconsole at yahoo.com>
> Cc: Paolo Merisio <merisiop at gmail.com>; ldoms-discuss at opensolaris.org; 
> clusters <ha-clusters-discuss at opensolaris.org>
> Sent: Monday, July 20, 2009 2:27:24 PM
> Subject: iSCSI & Quorum [was: Re: [ldoms-discuss] LDoms 1.2]
> 
> [Cross posting to ha-clusters-discuss]
> Hi Octave,
> 
>     Given that iSCSI targets are purely a network thing, i doubt
> very much if using them with or without LDoms has any difference.
> 
>     That leaves the question of how SC/OHAC supports iSCSI disks.
> 
> I believe that we (SC) have focused on OpenSolaris/OHAC for iSCSI support
> given the rather large number of issues they have on S10. Perhaps
> someone on ha-clusters-discuss can shed more light on this question
> (both general iSCSI support as well as supporting those as Quorum disks).
> 
> Regards,
> -ashu
> 
> 
> Octave Orgeron wrote:
>> Quick question, are iSCSI disks supported for quorum and in SC or Open HA 
>> with LDoms?
>>
>>  
>> *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
>> Octave J. Orgeron
>> Solaris Virtualization Architect and Consultant
>> Web: http://unixconsole.blogspot.com
>> E-Mail: unixconsole at yahoo.com
>> *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
>>
>>
>>
>> ----- Original Message ----
>> From: Ashutosh Tripathi <Ashutosh.Tripathi at Sun.COM>
>> To: Paolo Merisio <merisiop at gmail.com>
>> Cc: ldoms-discuss at opensolaris.org
>> Sent: Monday, July 20, 2009 1:16:40 PM
>> Subject: Re: [ldoms-discuss] LDoms 1.2
>>
>> Hi Paolo,
>>
>>     Tell us a bit more about the kind of device
>> d5 is. The only kind on which Quorum is supported is
>> ones backed by full LUNs in the I/O domain.
>>
>>     d5 looks suspiciously small to be a full disk, but
>> can you confirm?
>>
>> -ashu
>>
>> Paolo Merisio wrote:
>>> Ok, here more information about new test.
>>>
>>> root at n1 # [b]metaset -s 2ds[/b]
>>>
>>> Set name = 2ds, Set number = 2
>>>
>>> Host                Owner
>>>   n1                 Yes
>>>   n2                   n3                Driv Dbase
>>>
>>> d5   Yes [b]Import/export of metaset, mount of fs in metaset are all ok on 
>>> all 3 nodes[/b] root at n1 # [b]cldev show d5[/b]
>>>
>>> === DID Device Instances ===                  DID Device Name:              
>>>                   /dev/did/rdsk/d5
>>>   Full Device Path:                                n1:/dev/rdsk/c0d2
>>>   Full Device Path:                                n2:/dev/rdsk/c0d2
>>>   Full Device Path:                                n3:/dev/rdsk/c0d2
>>>   Replication:                                     none
>>>   default_fencing:                                 nofencing
>>>
>>> root at n1 # [b]prtvtoc /dev/rdsk/c0d2s0[/b]
>>> * /dev/rdsk/c0d2s0 partition map
>>> *
>>> * Dimensions:
>>> *     512 bytes/sector
>>> *      16 sectors/track
>>> *       4 tracks/cylinder
>>> *      64 sectors/cylinder
>>> *   32768 cylinders
>>> *   32766 accessible cylinders
>>> *
>>> * Flags:
>>> *   1: unmountable
>>> *  10: read-only
>>> *
>>> *                          First     Sector    Last
>>> * Partition  Tag  Flags    Sector     Count    Sector  Mount Directory
>>>        0      4    00       8256   2088768   2097023
>>>        7      4    01          0      8256      8255
>>>
>>>
>>> root at n1 # tail -f /var/adm/messages&
>>> [1] 5597
>>>
>>> Jul 17 09:49:41 n1 Cluster.Framework: [ID 801593 daemon.notice] stdout: 
>>> becoming primary for 2ds
>>>
>>> [b]root at n1 # clq add d5
>>> clq:  (C192716) I/O error.
>>> root at n1 # [/b]
>>>
>>> No other errors on console of other nodes or messages of other nodes.
>>> I hope this helps you to troubleshooting this problem.
>>>
>>> Thanks
>> _______________________________________________
>> ldoms-discuss mailing list
>> ldoms-discuss at opensolaris.org
>> http://mail.opensolaris.org/mailman/listinfo/ldoms-discuss
>>
>>
>>
>>      
> 
> 
>       
> _______________________________________________
> ha-clusters-discuss mailing list
> ha-clusters-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/ha-clusters-discuss

Reply via email to