Re: Understanding gossip and seeds

[Jeff Jirsa]

On 2017-07-22 07:08 (-0700), Daniel Hölbling-Inzko 
 wrote: 
> Seeds are there to bootstrap a node for the very first time when it's has
> zero knowledge about the ring.
> 
> I think I also read somewhere that seed nodes are periodically queried for
> some sanity checks and therefore one should not include too many nodes in
> the seed list.

Seeds have 2 purposes:

1) To find the ring the very first time,
2) To speed up convergence.

Each second, each live node will gossip with one other live node at random, and 
"maybe" gossip with one of the seed nodes.

If all of your seed nodes die or get replaced, you'll still have a working 
cluster, no big deal, it just may take longer for new changes (new nodes, 
leaving nodes, schema changes) to fully propagate..


-
To unsubscribe, e-mail: user-unsubscr...@cassandra.apache.org
For additional commands, e-mail: user-h...@cassandra.apache.org

Re: Understanding gossip and seeds

[Daniel Hölbling-Inzko]

Seeds are there to bootstrap a node for the very first time when it's has
zero knowledge about the ring.

I think I also read somewhere that seed nodes are periodically queried for
some sanity checks and therefore one should not include too many nodes in
the seed list.
kurt greaves  schrieb am Sa. 22. Juli 2017 um 01:48:

> Haven't checked the code but pretty sure it's because it will always use
> the known state stored in the system tables. the seeds in the yaml are
> mostly for initial set up, used to discover the rest of the nodes in the
> ring.
>
> Once that's done there is little reason to refer to them again, unless
> forced.
>

Re: Understanding gossip and seeds

[kurt greaves]

Haven't checked the code but pretty sure it's because it will always use
the known state stored in the system tables. the seeds in the yaml are
mostly for initial set up, used to discover the rest of the nodes in the
ring.

Once that's done there is little reason to refer to them again, unless
forced.