[
https://issues.apache.org/jira/browse/CASSANDRA-644?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12801117#action_12801117
]
Jaakko Laine edited comment on CASSANDRA-644 at 1/16/10 7:23 AM:
-----------------------------------------------------------------
patch attached:
- removeToken now removes node from gossip
- fat clients (nodes without token) are removed from gossip after 1 hour of
inactivity
- added justRemovedEndPoints to gossip to prevent removed nodes from
reappearing immediately. Removed nodes are kept here for RING_DELAY period,
during which time new joins from them are ignored. It takes a while for remove
token gossip to propagate to all nodes. During this time some nodes will
continue to gossip about the just-now-being-removed-node, while others have
already removed it.
- fixed a bug related to removeToken command which allowed node's own token
being removed.
was (Author: jaakko):
patch attached:
- removeToken now removes node from gossip
- fat clients (nodes without token) are removed from gossip after 1 hour of
inactivity
- added justRemovedEndPoints to gossip to prevent removed nodes from
reappearing immediately. Removed nodes are kept here for RING_DELAY period,
during which time new joins from them are ignored. It takes a while for remove
token gossip to propagate to all nodes. During this time some nodes will
continue to gossip about the just-now-being-removed-node, while others have
already removed it.
> Provide way to remove nodes from gossip entirely
> ------------------------------------------------
>
> Key: CASSANDRA-644
> URL: https://issues.apache.org/jira/browse/CASSANDRA-644
> Project: Cassandra
> Issue Type: Bug
> Components: Tools
> Affects Versions: 0.5
> Reporter: Jonathan Ellis
> Assignee: Jaakko Laine
> Priority: Minor
> Fix For: 0.9
>
> Attachments: 644.patch
>
>
> As reported in CASSANDRA-634, "Now that we're gossiping about dead nodes as
> well, gossip digest continues to grow without boundary when nodes come and
> go. This information will never disappear as it will be propagated to new
> nodes no matter how old and obsolete it is. To counter this, we need some
> mechanism to (1) either remove dead node from endpointstateinfo or (2) at
> some point stop to gossip about it, or both."
> This is also seen when using "fat clients" that participate in the gossip
> ring; if a client leaves and does not come back it stays in the gossip
> forever. (This can be confusing if the client does start up again,
> connecting to a _different_ cluster, but the old one notices it is back and
> starts gossiping to it again!)
> I would prefer to leave management of these things explicit; 3 days is long
> enough that the fat client problem in particular needs another solution, and
> if it needs another solution then that can become the only solution. :)
> So I would be in favor of removeToken clearing out gossip entries, and also
> adding a command to remove an endpoint from the gossip ring that does not
> have a token associated with it (like fat clients). A command to ask "what
> are all the known gossip hosts" would also be useful, since nodeprobe ring
> only includes nodes w/ tokens.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
