[erlang-questions] auto-syncing mnesia after a network split

[mats cronqvist]

Claes Wikström <klacke@REDACTED> writes:

> Jim McCoy wrote:
>
>> would be to add some complexity on the mechanics of the join mechanism
>> (e.g. a quorum system like paxos to decide membership and agreement
>
>
> A lot of the mnesia clusters only have 2 nodes - thus making
> quorums a no-goer. typical HA telcoms chassis systems have exactly
> two management blades.

  i find it quite humorous that the telecoms industry is paying
  thousands of people to work on solving this problem (how to reliably
  rejoin a partitioned network consisting of two nodes); a problem
  that has been proven (over 30 years ago) to have no solution.

  30,000 man-years... that a chunk of change(*). certainly a lot more
  expensive than adding a third node.

  it's less humourous (at least to me) that i personally spent a few
  of those man-years, when all I had to do was google on "leslie
  lamport."

  mats

(*) yes, i made that figure up. but i do believe it's the right
magnitude.