[volunteergrid2-l] Problems

[Peter Secor]

Turn it up to 11!

One failure case that I ran into a couple times is what happened to 
Allmydata (the company). This is the case where there are lots of 
storage nodes (100+) and plenty of space on the grid, and using 
something like 3/7/10 as the parameters for needed/healthy/total. What 
happened was that we lost access to a large percentage of our machines 
at any one time (power problem), but, on average if you had a particular 
filecap, you could retrieve it within some number of iterations through 
powering up different combinations of the servers. The big problem was 
that different caps were stored across different sets of servers, so 
let's say that you had a nice big directory structure, then each dircap 
would be on a different set of servers, so traversing it to find your 
data would be very difficult because though the data was still extant in 
the grid, it was always a different set of servers that contained it.

Ps

On 10/4/12 12:25 AM, Ted Rolle, Jr. wrote:
> Thanks so much.  I had no idea what to use.  Those numbers seemed to
> work OK, but I'll change mine to yours.
> Would 11 instead of 10 be OK? :-)
>