Re: Cluster configuration - request for input.



Brian,

Thanks for your input. When in production, each node will be Gigabit
connected directly to a router so thoughput shouldn't be an issue in the
event of a failure. As for the 1TB LUN, only about 300GB will be in use at
the outset, the rest is for growth. Your concerns, however, have not fallen
on deaf ears and such considerations will be discussed internally.

The LUNs are RAID-5 on a brand new EMC SAN.

--
Tim.

"Brian Desmond [MVP]" <brian@xxxxxxxxxxxxxxxx> wrote in message
news:Oo96nL$0FHA.3756@xxxxxxxxxxxxxxxxxxxxxxx
> Tim,
>
> At a high level, this is fine. Running active/active here is a good use of
> available resources. Just make sure you have enough network throughput to
> each machine to support both of them if need be.
>
> My only concern with those LUN sizes is SLA. What if something happens to
> that 1T lun. How are you going to restore that from tape in any reasonable
> amount of time?
>
> --
> Thanks,
> Brian Desmond
> Windows Server MVP
>
> www.briandesmond.com
>
>
> "Tim Munro" <Excelsior@xxxxxxxxxxx> wrote in message
> news:uDneWX%230FHA.2428@xxxxxxxxxxxxxxxxxxxxxxx
>> Hello all,
>>
>> I am about to start configuring a W2k3 (Active directory) cluster for
>> file services offerings (and a possible DFS root).
>>
>> Hardware:
>> HP DL380 G4, 4GB ram, RAID-1 36GB system drive
>> SAN attached shared disk resources (0.5TB, and 1TB plus 1GB
>> quorum)
>> Each disk has one partition, the full size of the disk.
>>
>> I would like to create one virtual node accessing the 0.5TB disk, and
>> one virtual node accessing the 1TB disk. I would like to run each of
>> these virtual nodes one on each of the physical nodes.
>>
>> Why?
>> 1) political "perception" that student traffic and faculty/staff
>> traffic are on separate machines.
>> 2) spread the load out a bit. (both are fairly heavily used)
>>
>> Each would act as a failover for the other if necessary.
>>
>> Questions:
>> Is the disk partitioning acceptable? Optimal? (I may need to
>> dynamically expand the disks in the future)
>> Is this an acceptable cluster setup? Common? Supported? recommended?
>>
>> Any ideas, comments, do's and don'ts are greatly appreciated.
>>
>> Thanks very much
>>
>> --
>> Tim.
>>
>
>


.



Relevant Pages

  • Re: How to make disk device "Available"?
    ... >> PROM Revision ... > I cannot tell from that output if that is the LUN you are ... Description: 1.0 GB SCSI Disk ...
    (comp.unix.aix)
  • Re: Cluster configuration - request for input.
    ... This cluster is NOT for AD DCs, ... All are Fibre Channel 143 GB disks. ... As for the 1TB LUN, only about 300GB will be in use ... >> Tim. ...
    (microsoft.public.windows.server.clustering)
  • Re: Adding more iSCSI LUNs to existing MSCS 2000
    ... MVP - Windows Server - Clustering ... > If I add an iSCSI or fibre channel disk to 2 machines at once, ... > Adding a LUN, no problem. ... > both disks are not reading/writing to the LUN on the SAN. ...
    (microsoft.public.windows.server.clustering)
  • [HPADM] Summary: Nike Model 10 Question
    ... Disk vendor is not an issue. ... Craig ... >and not something pulled out of another LUN Thanks! ... different capacities for 4G drives. ...
    (HP-UX-Admin)
  • Re: Disk Performance issues?
    ... The only way to get more I/O per user is to either add more disks to the LUN ... It sounds like you have all your databases on the single RAID 10 ... I am just not sure if there is anything I can do to speed up disk access at this point? ...
    (microsoft.public.exchange.admin)