All servers that are configured for the HA feature must
be on the same version. Thus, all servers must be upgraded at the
same time.
Note: During the upgrade of
IBM DevOps Deploy (Deploy), the contents of
the appdata/patches directory (including all .jar files) will be appended with
.off, preventing the existing test fixes from being loaded once the
Deploy server
is started post-installation. All UI test fixes applied to the Tomcat directory will be
wiped by the upgrade and will not be recoverable.
If a text fix received for a reported
defect is not fixed in the upgraded version, you may need to open a Support Case to
request a new build.
-
Back up the database.
-
Ensure that no application and generic processes are running.
Open the Dashboard and confirm that no processes are running. If
processes are running, you can wait for them to complete or by clicking
Cancel in the same row as each process, cancel them.
You can use the Message of the Day feature () to notify users of the planned upgrade, and request that they do
not start new processes.
-
Download and extract the installation package.
Upgrades are done with the same package used to perform new installations.
-
Shut down the servers.
Warning: Ensure that all servers are stopped. If any servers are accessing the
database when the upgrade happens, the upgrade will cause errors.
-
Ensure that each server is stopped and unable to connect to the load balancer.
-
Check that there is no server.pid file in the
server_install/var
folder.
-
Check for running system processes.
On Linux™, you can use the
netstat or ps commands. For
example, if the server is running on port 8443, use the command
netstat -nap | grep 8443
. You can also run
ps -ef | grep server
and look for the server system
process in the results.
-
Stop any running server process.
If you try to upgrade with a running old server process, you get an
error that says A previously installed version of Deploy
is running. Please shutdown the running Deploy and start the
installation again. .
If you try to upgrade with a running old server process, you get an
error that says A previously installed version of Deploy
is running. Please shutdown the running Deploy and start the
installation again. .
-
Update the network storage to match the folders in the new version.
For information about how to structure the folders, see
Setting up high-availability clusters.
Depending on
how you set up network storage for the high-availability cluster in the previous
version, you might need to move files and folders to put them in the place that the new
version of the server expects to find them. When you run the installation program,
specify the location of the app_data folder in
the Where should the server store application data such as logs,
plugins, and keystores? prompt.
For more information, see
Setting up high-availability clusters.
- Run the server installation script.
-
From the command line, run the server installer program.
- On Windows™, run the
install-server.bat file.
- On Linux, run
./install-server.sh.
-
To install a FIPS-compliant server, add the
-fips-semeru
parameter
to the command, such as
install-server.sh -fips-semeru
.
Note:
- If you install a FIPS-compliant server, cloud provisioning is not available.
- Deploy
supports IBM Semeru versions 11.0.24, 17.0.12, 21.0.4 or higher.
Depending on your system settings, you might need to run this file as an administrator.
- When prompted for the location of the installation directory,
enter the path to one of the servers.
By specifying an existing installation, you are asked if you want to upgrade the current
installation. If you answer yes, the script leads you through the upgrade steps.
-
Restart the upgraded server.
-
After the first server is back online, upgrade the other servers. After you upgrade the first
node, you will not be prompted to upgrade the database when you upgrade the other nodes.