What happens when the trial expires?

Follow

Aviram Alter

• February 12, 2017 11:13

I currently have setup Percona replication, two nodes. What happens when the trial expires, though?

Will my nodes continue to function the same and I will just lose the ability to manage them via ClusterControl (can still manage manually I assume?) ?

 

Thanks

0

• 

  Ashraf Sharif

  • February 12, 2017 12:12

  Hi there,

  ClusterControl will default to community edition once it's expired. Details at https://severalnines.com/pricing and click on "Expand full list of features". And yes, you still able to manage your Replication manually after the trial expires.

  Regards,
  Ashraf

  0

  Comment actions Permalink
• 

  Aviram Alter

  • February 12, 2017 15:22

  Thanks

  One more question:

  Will failover still occur automatically in the community edition, or will this also become manual?

  0

  Comment actions Permalink
• 

  Ashraf Sharif

  • February 13, 2017 02:15

  Hi,

  Database failover via load balancers like HAproxy/MaxScale/ProxySQL will happen automatically as it should be after the trial expires. Simply said, what has been deployed will retain after the trial expires. You will miss the scaling, management and monitoring granularity features in community edition.

  Regards,
  Ashraf

  0

  Comment actions Permalink
• 

  Ashraf Sharif

  • February 13, 2017 07:00

  Hi,

  I guess my previous reply was in the wrong context. If you mean Percona Replication failover of the master-slave topology (automatic cluster/node recovery by ClusterControl) is only available in the commercial edition.

  Regards,
  Ashraf

  0

  Comment actions Permalink
• 

  Aviram Alter

  • February 13, 2017 09:36

  Thank you for the reply Ashraf.

  If I could trouble you with another question on a different topic...

  If I put HAproxy in front of my replication topology (three nodes, master+ 2 slaves), will this work correctly? Will HAproxy make decisions also based on replication lag with the default settings, or should I change anything? the documentation did not make this clear for me.

   

  Thanks

  0

  Comment actions Permalink
• 

  Ashraf Sharif

  • February 13, 2017 13:37

  Hi Aviram,

  If you deploy HAproxy using ClusterControl, then it should detect the replication topology accordingly based on replication role and read_only flag. By default, the slave lag limit is 5s. You can change it by going to /usr/local/sbin/mysqlchk on every DB node under SLAVE_LAG_LIMIT variable.

  Regards,
  Ashraf

  0

  Comment actions Permalink
• 

  Aviram Alter

  • February 13, 2017 14:35

  Any guide for what settings I should use for a deployment of this kind? For example load balancing method, I'm sure round robin is not a good choice...

  0

  Comment actions Permalink
• 

  Ashraf Sharif

  • February 14, 2017 02:17

  Hi Aviram,

  You can go to our HAproxy tutorial page for details:

  https://severalnines.com/resources/tutorials/mysql-load-balancing-haproxy-tutorial

  Look at the bullet points 3.1.2. (mysqlchk for MySQL Replication) and 5. (Read/Write Splitting with HAProxy). Also, we have covered an example use case on MySQL Replication with HAproxy in this blog post:
  http://severalnines.com/blog/high-availability-read-write-splitting-php-mysqlnd-mysql-replication-and-haproxy

  Regards,
  Ashraf

  0

  Comment actions Permalink

Please sign in to leave a comment.