Bringing together the Apache Cassandra experts from the community and DataStax.

Want to learn? Have a question? Want to share your expertise? You are in the right place!

Not sure where to begin? Getting Started

 

question

pi_165798 avatar image
pi_165798 asked Erick Ramirez answered

Is there anything I should be aware of when upgrading Cassandra from 3.11.3 to 3.11.6?

Is there anything I should be aware of when upgrading Cassandra from 3.11.3 to 3.11.6?

I have a hard time finding any official documentation on the upgrade process. Would it be okay to do rolling upgrades for each node like this?:

  
                 
    
                  
  1. nodetool drain
    2. 
                  
  2. service cassandra stop
    3. 
                  
  3. apt-get install cassandra=3.11.6
    4. 
                  
  4. service cassandra start
    5. 
                  
  5. nodetool upgradesstables (after verifying everything looks ok in logs)
    6.
cassandraupgrade
10 |1000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 1.0 MiB each and 10.0 MiB total.

1 Answer

Erick Ramirez avatar image
Erick Ramirez answered

@pi_165798 I agree about the documentation and we are going to do more in that space by contributing both documents and technical writers' time to the Apache project (discussions are in progress so stay tuned).

In the meantime, you can have a look at this Apache Cassandra upgrade procedure document. It is for upgrading from OSS C* to a special DataStax distribution of C* but the general advice in the document still works for regular OSS C* to OSS C* upgrade (we are rewriting it for OSS).

Some of the things you need to note from that document are:

  • backup the cluster data
  • backup the C* configuration files, specially cassandra.yaml and cassandra-env.sh
  • make sure that repair jobs are temporarily disabled during the upgrade
  • do not TRUNCATE any tables or make schema changes during the upgrade
  • after the binaries are installed, verify that your configuration files have not been overwritten

In addition to the above, I recommend that you always check NEWS.txt for any special upgrade information &/or instructions on (1) the version you're upgrading to, and (2) all the versions in between. You'll see that there's a new property repair_session_max_tree_depth was added. It allows you to cap the size of Merkle trees so repairs don't put too much pressure on the heap.

And finally, test, test, TEST! Practice makes perfect and gives you an idea of what to expect when it's time to implement in production.

Good luck and keep us posted on how you're going with the upgrade. Cheers!

Share
10 |1000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 1.0 MiB each and 10.0 MiB total.