cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Richard Low (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-7307) New nodes mark dead nodes as up for 10 minutes
Date Thu, 05 Jun 2014 08:26:01 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-7307?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14018575#comment-14018575
] 

Richard Low commented on CASSANDRA-7307:
----------------------------------------

I don't know about the FD params so can't comment on that, but my test case shows it when
growing a 2 node cluster to 3 nodes. We also saw it in prod in a larger cluster when doing
node replacement. So it shows in both small and large clusters.

bq. but isn't adding capacity when things are up, more common than when things are down?

True, but replacing nodes when nodes are down is common. That's why we really care about it
- currently if there are 2 nodes down from a replica set you can't use -Dcassandra.replace_address
because the replacement node tries to stream from the dead one.

bq. What do you get for 10s / 5s, Richard Low

I will test locally but harder to test on a large cluster. Brandon suggested 1 second when
doing replacement.

NB on 1.1 we had no problems with this so the initial conditions used to be better, at least
when doing replacement.

> New nodes mark dead nodes as up for 10 minutes
> ----------------------------------------------
>
>                 Key: CASSANDRA-7307
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-7307
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Richard Low
>            Assignee: Brandon Williams
>             Fix For: 1.2.17
>
>
> When doing a node replacement when other nodes are down we see the down nodes marked
as up for about 10 minutes. This means requests are routed to the dead nodes causing timeouts.
It also means replacing a node when multiple nodes from a replica set is extremely difficult
- the node usually tries to stream from a dead node and the replacement fails.
> This isn't limited to host replacement. I did a simple test:
> 1. Create a 2 node cluster
> 2. Kill node 2
> 3. Start a 3rd node with a unique token (I used auto_bootstrap=false but I don't think
this is significant)
> The 3rd node lists node 2 (127.0.0.2) as up for almost 10 minutes:
> {code}
> INFO [main] 2014-05-27 14:28:24,753 CassandraDaemon.java (line 119) Logging initialized
> INFO [GossipStage:1] 2014-05-27 14:28:31,492 Gossiper.java (line 843) Node /127.0.0.2
is now part of the cluster
> INFO [GossipStage:1] 2014-05-27 14:28:31,495 Gossiper.java (line 809) InetAddress /127.0.0.2
is now UP
> INFO [GossipTasks:1] 2014-05-27 14:37:44,526 Gossiper.java (line 823) InetAddress /127.0.0.2
is now DOWN
> {code}
> I reproduced on 1.2.15 and 1.2.16.



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Mime
View raw message