Well, it makes it slightly easier to guarantee that you've not missed
publishing anything.
In general, though, already-published records are basically
irrelevant, and just use up space in the DB.
(One of the strange database implementation details in APEL, along
with the use of MyISAM db, rather than InnoDB.)
Sam
On 9 December 2010 11:24, Matt Doidge <[log in to unmask]> wrote:
> Thanks for all the replies everyone, it seems like VMs are fine for the
> glite-APEL box - that makes me a lot happier. I've also been intrigued by
> Winnie's question - is there a strong argument for or against importing the
> old database?
>
> Cheers,
> Matt
>
> Winnie Lacesso wrote:
>>
>> This discussion's of interest as Bristol's about to do this too. I was
>> going to follow Birmingham's route and not import the old db:
>> http://www.ep.ph.bham.ac.uk/twiki/bin/view/Computing/ApelUpgrade
>> Minimize work, why not. Anyone can say why it should / has to be imported?
>>
>> Also, in the Nagios monitoring, all the glite-APEL are
>> "No matching services"
>>
>> https://gridppnagios.physics.ox.ac.uk/nagios/cgi-bin/status.cgi?hostgroup=node-APEL&style=overview
>> Naively one might think: . software development of glite-APEL
>> . test functionality till success
>> . test monitoring till success
>> . release for production deployment
>>
>> Step 3 an afterthought?
>>
>> "Everybody's a critic" / sorry
>
|