5.14.5. Upgrading or Updating your JVM

When upgrading your JVM version or installation, care should be taken as changing the JVM will momentarily remove and replace required libraries and components which may upset the operation of Tungsten Clustering while the upgrade or update takes place.

For this reason, JVM updates or changes must be treated as an OS upgrade or event, requiring a master switch and controlled stopping/shunning of services during the update process.

A sample sequence for this in a 3-node cluster is described below:

Step Description Command host1 host2 host3
1 Initial state   Master Slave Slave
2 Shun slave host2 datasource host2 shun Master Shunned Slave
3 Stop all services on host2. stopall Master Stopped Slave
4 Update the JVM   Master Stopped Slave
5 Start all services on host2 slave. startall Master Slave Slave
6 Recover slave back datasource host2 recover Master Slave Slave
7 Ensure the slave ( host2 ) has caught up ls Master Slave Slave
8 Shun slave host3 datasource host3 shun Master Slave Shunned
9 Stop all services on host3. stopall Master Slave Stopped
10 Update the JVM   Master Slave Stopped
11 Start all services on host3 slave. startall Master Slave Slave
12 Recover slave back datasource host3 recover Master Slave Slave
13 Ensure the slave ( host3 ) has caught up ls Master Slave Slave
14 Switch master to host2 switch to host2 Slave Master Slave
15 Shun host1 datasource host1 shun Shunned Master Slave
16 Stop all services on host1. stopall Stopped Master Slave
17 Update the JVM   Stopped Master Slave
18 Start all services on host1 slave. startall Slave Master Slave
19 Recover host1 back datasource host1 recover Slave Master Slave
20 Ensure the slave ( host1 ) has caught up ls Slave Master Slave
21 Switch master back to host1 switch to host1 Master Slave Slave