On 7 March 2011 22:57, Christopher J. Walker <[log in to unmask]> wrote:
> On 07/03/11 18:35, Peter Grandi wrote:
>>
>> On 07/03/11 15:18, Sam Skipsey wrote:
>>
>> [ ... ]
>>>
>>> I also feel I should note that Tier-2 sites are not expected to have
>>> Tier-1 levels of resiliency or redundancy, which appears to be what
>>> you're trying to approach. ;)
>
> One can aspire to this of course - and there are some things that are easy
> to do.
>
>>
>> More resilience and redundancy for our grid systems would mean that
>> I can delay responding to what would otherwise be urgent grid issues
>> that would impact non-grid matters where my management requires near-24x7
>> availability from both systems and myself. Also, as to grid systems
>> themselves, as someone pointed out, if one of our essential grid
>> servers fails during Friday night and I don't realize or otherwise
>> I am not available until Monday morning, that's our stats for the month
>> gone.
>
> Atom: Why buy new machines when you have old worker nodes that are
> presumably sufficiently powerful - and if you have several, then you don't
> need to worry about warranty.
>
> VMs: I'd like to do this - at least in part to To be honest, I think this is
> probably the way to go. If you have shared storage, Vserver is presumably
> Whether
>
The thing with VMs is balancing your I/O load and all that. It's
certainly more than possible to stick a Site-BDII and an APEL service
on a VM (in fact, you can probably stick them in the *same* VM...),
but CE and SE services are a little more tricky - they both have a
certain amount of I/O they need, especially at their peak loads.
(It may be the case that lighter weight implementations would be
happier in a VM - I suspect StoRM SEs would be better on a VM than
DPM, for example.)
Sam
|