0

I have 6 nodes in a redis cluster - 3 master and 3 slaves. All the nodes are running just fine except for one master and it's slave.

On this troublesome master, when i run the command (on either master or slave), i get the following error:

Error: Server closed the connection

There is nothing in the redis error logs following this error. The redis error logs shows that everything is running smooth and this articular error does is not included in the logs.

However, the predis error logs show this error:

Error while reading line from the server [tcp: //192.168.2.100: 6379]

What can be causing this problem?

The Georgia
  • 1,005
  • 7
  • 23
  • 59
  • Which version of Redis cluster are you running ? – Rabea Oct 06 '15 at 19:24
  • I am running version 3.0.2. And they are running on Ubuntu 14.04 – The Georgia Oct 07 '15 at 00:27
  • I saw before an issue on Github unfortunately couldn't find now about a bug when the master/slave is not assigned their slots causing such an error. Anyway Redis 3.0.4 seams to show some fixes on the Redis cluster in general, it would be preferable to upgrade https://github.com/antirez/redis/blob/3.0.4/00-RELEASENOTES – Rabea Oct 07 '15 at 01:37
  • When you say that the master and ska not being their slots, you mean they at missing some slots? If any slots of the 16k + slots were not being cared for, the whole cluster would not work. All slots must be covered at all times. Unless missunderstood you? – The Georgia Oct 07 '15 at 01:55
  • You are right, other nodes will stop responding till all slots are assigned. I will keep checking from my side. that was an old issue where the master is responding but the other nodes are not https://github.com/antirez/redis/issues/1869 – Rabea Oct 07 '15 at 02:11
  • Checked the logs of the two instances and they keep repeating this message: CONFIG REWRITE executed with success over and over – The Georgia Oct 07 '15 at 06:41
  • I see you got that fixed. – Rabea Oct 07 '15 at 13:06

0 Answers0