Having gone through two power failures in two seperate "world class
datacenters", i've learnt that redundancy should exist in all levels,
sometimes even to the network uplink (In one setup, we had a seperate leased
circuit running into our servers). Never neglect even the finer 'simple'
points of failures - cheap hdd cables, too little inodes (system monitor
monitored everyting from bytes free to swap space usage, but we forgot that
we had far too many small files), admin swapping a working raid1 hdd out for
a blank (happens soooo often)

'Artificial Intelligence is no match for Human Stupidity'

Elvin
-



On 7/26/07, Maximilian Jackson <[EMAIL PROTECTED]> wrote:
>
> hehe yes indeed.
>
> actually, i remember one occassion, the SPOF of this cluster set up is the
> heartbeat line itself.
>
> *ubergoonz <[EMAIL PROTECTED]>* wrote:
>
>
>
>
>
> On 7/25/07, Maximilian Jackson <[EMAIL PROTECTED]> wrote:
> >
> > Secondary system is just a cluster setup isn't it? Personally, I think
> > having 2 systems is rather good. You have that sense of redundancy.
> >
>
> yeah, given that if money is of no concern, N+N cluster will eliminate
> all(almost, don't forget act of god) SPOF.
>
>
> ------------------------------
> Try Yahoo! 
> Mail<http://us.rd.yahoo.com/mailuk/taglines/gmail/see_difference/*http://uk.rd.yahoo.com/evt=48525/*http://uk.mail.yahoo.com>now
>  with Unlimited Storage and see the difference.
>
>
> _______________________________________________
> Slugnet mailing list
> [email protected]
> http://www.lugs.org.sg/mailman/listinfo/slugnet
>
>
_______________________________________________
Slugnet mailing list
[email protected]
http://www.lugs.org.sg/mailman/listinfo/slugnet

Reply via email to