[H-GEN] Best FS for use with multiple Unix variants

[Robert Brockway]

On Wed, 19 Aug 2009, Arjen Lentz wrote:

> Good, but needs updating.

I disagree that the information is out of date.

> As I wrote, in some cases these days offline no longer works, the time 
> required to restore is longer than the time the business is allowed to 
> be not functional. it's not just time required to get data back on-site, 
> but actually mainly the time just restoring.

The offsite backup still has value, even if a DR site is in use.

> - being in multiple locations/datacentres is quite easy/cheap now. this

That depends on the volume of data & transactions involved.  A DR site is 
still prohibitively expensive for a lot of organisations.  Small 
organisations often run on very tight margins but may actually be ok with 
an outage lasting a few days (eg, a corner shop).  Many large sites have 
trouble actually pushing the amount of data required to maintain full 
sync.

> way a backup is effectively on-site; it's just not the same site ;-)

Replication is subject to numerous failure modes.  Eg, a compromise that 
is not detected until after replication has occured.  You may find all 
your copies compromised or broken, depending on what happened.  This is 
where the offsite backup comes in.

Aside from anything else you have in place, not having backups that are 
offsite, offline & tested is sheer folly.

I do have quite a bit of information on DR which need to be added to the 
site yet (see the banner on the wiki).

Rob

-- 
I tried to change the world but they had a no-return policy
Projected IPv4 exhaustion: http://www.potaroo.net/tools/ipv4/index.html