CODEWITHSUNDEEP
servercassandra
Vikram Mishra
Vikram Mishra

Reputation: 1

Cassandra is going down frequently without any error or exception

Cassandra is going down frequently in server, Checked for RAM and disk space. It's sufficient to run Cassandra. There is no error or exception in Cassandra logs file. Cassandra version: 2.1.12

Upvotes: 0

Views: 551

Answers (1)

Erick Ramirez
Erick Ramirez

Reputation: 16303

If the messages in the logs stop abruptly, it's most likely the Linux oom-killer that is terminating the Cassandra process. You can see evidence of this in the operating system logs.

You didn't mention the Cassandra version but if:

  • you're running C* 2.2 or newer
  • node data size is greater than RAM

There's a chance that you're hitting a known issue where all SSTables are getting mapped to memory (with mmap). If this applies to your cluster, you will need to explicitly configure disk_access_mode to mmap_index_only so C* doesn't use up all the memory on the nodes.

I've explained the cause and solution in detail in this post -- https://community.datastax.com/questions/6947/. Cheers!

Upvotes: 1

Related Questions