Skip to content

redis cluster doesn't failover automatically #1047

@frafra

Description

@frafra

This is what I get when I stop the master.

[8465] 10 Apr 22:05:43.433 * Caching the disconnected master state.
[8465] 10 Apr 22:05:43.769 * Connecting to MASTER...
[8465] 10 Apr 22:05:43.769 * MASTER <-> SLAVE sync started
[8465] 10 Apr 22:05:43.769 # Error condition on socket for SYNC: Connection refused
[8465] 10 Apr 22:05:44.771 * Connecting to MASTER...
[8465] 10 Apr 22:05:44.771 * MASTER <-> SLAVE sync started
[8465] 10 Apr 22:05:44.772 # Error condition on socket for SYNC: Connection refused
[8465] 10 Apr 22:05:45.774 * Connecting to MASTER...
[8465] 10 Apr 22:05:45.775 * MASTER <-> SLAVE sync started
[8465] 10 Apr 22:05:45.775 # Error condition on socket for SYNC: Connection refused
[8465] 10 Apr 22:05:46.777 * Connecting to MASTER...
[8465] 10 Apr 22:05:46.777 * MASTER <-> SLAVE sync started
[8465] 10 Apr 22:05:46.777 # Error condition on socket for SYNC: Connection refused
[8465] 10 Apr 22:05:47.779 * Connecting to MASTER...
[8465] 10 Apr 22:05:47.780 * MASTER <-> SLAVE sync started
[8465] 10 Apr 22:05:47.780 # Cluster state changed: fail
[8465] 10 Apr 22:05:47.780 # Error condition on socket for SYNC: Connection refused
[8465] 10 Apr 22:05:48.782 * Connecting to MASTER...
[8465] 10 Apr 22:05:48.783 * MASTER <-> SLAVE sync started
[8465] 10 Apr 22:05:48.783 # Error condition on socket for SYNC: Connection refused

...and it continues like this.
This is a test using localhost. I've tried to use one maser and one slave and one master and two slaves, but the output is the same.

This is my redis.conf file:

port 1234
cluster-enabled yes
cluster-config-file nodes1.conf
cluster-node-timeout 5
appendonly yes

The other redis.conf files are just like this, but the port and the config-cluster-file are different.

This is the cluster nodes output on the slave before the failure:

redis 127.0.0.1:1235> cluster nodes
ffc6464df68df47864fde0ea9eb4d9d9191dc1b1 127.0.0.1:1234 master - 1365624792 1365624792 connected 0-16383
c785f0d5f4752aab965f39c09d66cb239d51772f 127.0.0.1:1236 slave ffc6464df68df47864fde0ea9eb4d9d9191dc1b1 1365624793 1365624793 connected
b37d87bb080a9021bc265f99c9f114005d8ef0f2 :0 myself,slave ffc6464df68df47864fde0ea9eb4d9d9191dc1b1 0 1365624792 connected

This is the cluster nodes output on the slave after the failure:

redis 127.0.0.1:1235> cluster nodes
ffc6464df68df47864fde0ea9eb4d9d9191dc1b1 127.0.0.1:1234 master,fail? - 1365624843 1365624807 disconnected 0-16383
c785f0d5f4752aab965f39c09d66cb239d51772f 127.0.0.1:1236 slave ffc6464df68df47864fde0ea9eb4d9d9191dc1b1 1365624843 1365624843 connected
b37d87bb080a9021bc265f99c9f114005d8ef0f2 :0 myself,slave ffc6464df68df47864fde0ea9eb4d9d9191dc1b1 0 1365624840 connected

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions