[
https://issues.apache.org/jira/browse/CASSANDRA-5836?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16382084#comment-16382084
]
Jeremiah Jordan commented on CASSANDRA-5836:
--------------------------------------------
So why don’t seed nodes bootstrap? The idea is that they are the first nodes
added to a cluster, so they can’t bootstrap, there is nothing to bootstrap
from. If we let seed nodes bootstrap, what is the logic for the new cluster
startup case? Have to be careful we don’t just start failing everyone’s new
cluster creations because bootstrap isn’t possible.
Some thoughts on what could be done:
Do we only allow seed nodes with only themselves in the list to start without
bootstrapping? So if you want to keep the list the same on all nodes you need
to modify the first nodes yaml?
Do we require setting autobootstrap to false for the first nodes in a cluster
and get rid of the special casing around seed nodes all together? Changing the
burden from modifying a nodes seedlist to re-add it, to making sure
autobootstrap is true?
If we want to mess with this let’s make sure to think through exactly how a
change here will affects cluster ops for initial creation and also for
replacing/adding nodes later.
I’m not sure what the right answer is, but let’s make sure any change here is a
clear win for all cases rather than just moving around the burdensome thing.
Also making sure to kee p clear logging around what is happening, as we do at
least currently have a clear clear log message that a node did not bootstrap
because it was a seed, that was a nice addition to help in tracking down why
your new node had no data.
> Seed nodes should be able to bootstrap without manual intervention
> ------------------------------------------------------------------
>
> Key: CASSANDRA-5836
> URL: https://issues.apache.org/jira/browse/CASSANDRA-5836
> Project: Cassandra
> Issue Type: Bug
> Reporter: Bill Hathaway
> Priority: Minor
>
> The current logic doesn't allow a seed node to be bootstrapped. If a user
> wants to bootstrap a node configured as a seed (for example to replace a seed
> node via replace_token), they first need to remove the node's own IP from the
> seed list, and then start the bootstrap process. This seems like an
> unnecessary step since a node never uses itself as a seed.
> I think it would be a better experience if the logic was changed to allow a
> seed node to bootstrap without manual intervention when there are other seed
> nodes up in a ring.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
