0 is a perfectly valid id.
node - 1 is modulo the maximum token value. that token range is 0 - 2**127
so node - 1 in this case is 2**127
has anybody written up anything related to recovery for fails in EC2?
this morning i woke up to find 1 (of 4) nodes marked as unreachable. i used the datastax (1.0.7) ami to set up my clu ster and the node that fail had the token id of 0 (this is the seed node - right?). the docs says to replace a failed node to set the token to failed node - 1, which i don't think would work. luckily rebooting the machine
but i'm left with the question of what should i do if node with the token ID of 0 fails and should i even have a node with that ID.
On 2/23/2012 11:21 AM, aaron morton wrote:
General EC2 setup.