Some questions

[Mirage74]

Lately i have started to use Redis as a queue, And i have some questions regarding my cluster (M-S-S):

1. Is there an option, other than log file, to get details regarding failovers? For example: Latest failover timestamp and reason, history of failovers, etc.
2. I went over my log file, and i have noticed that my cluster had a few failovers in the past few days. I couldn't find enough information for the failover reason. I can see the line "+odown master mymaster x.x.x.x 6379 #quorum 3/2" on the Sentinel log file, but i don't have any additional information regarding the reason for that. I didn't do any change to my master server, which was always up and running including his Redis service. How can i investigate this issue? 
3.  Here is a list of my Cluster's AOF and RDB files: 
   -rw-r--r-- 1 redis     redis       18M Nov 30 13:02 redis-master.aof

   -rw-r--r-- 1 redis     redis      5.1M Nov 29 13:07 redis-master.rdb

   -rw-r--r-- 1 redis     redis       14M Nov 30 13:02 redis-slave1.aof

   -rw-r--r-- 1 redis     redis      3.9M Nov 30 13:03 redis-slave1.rdb

   -rw-r--r-- 1 redis     redis       14M Nov 30 13:02 redis-slave2.aof

   -rw-r--r-- 1 redis     redis      3.9M Nov 30 13:03 redis-slave2.rdb
   How can it be that the master files are much bigger than the slaves files? (Currently no data been written into my Redis). Is this indicate for some kind issue? How can i investigate this issue?

Thanks

[Mirage74]

Anyone?

[David Geller]

Have you tried to correlate redis logs to sentinel logs? Have you checked the system performance during the time of your failure?  Was there a network issue by chance?

I had an issue where blocking aof writes would cause a disconnection of slaves and a failover and I found out by comparing aof "taking too long" messages with the failover in the sentinel logs.

The only thing sentinel is going to tell you is what you see... odown, sdown and the election process.