Upgrading Netezza Performance Server Replication Services

Learn how to upgrade Netezza Performance Server Replication Services.

Refer to the Upgrade paths section of the Netezza Performance Server Replication Services 3.0.X.X release notes for supported upgrade paths.

If you upgrade to the same version (for example, 3.0.1.0 to 3.0.1.0), all of the NRS containers are redeployed.

Note: Fresh installations or upgrades to NRS version 3.0.x.x are not currently supported on Cloud Pak for Data System version 1.0.9.0 and above. However, if NRS is already installed and running on version 1.0.8.x, it will continue to function after upgrading the Cloud Pak for Data System to version 1.0.9.x or later.

Note: If you plan to upgrade to NRS 3.0.5.1, ensure the upgrade is performed on both primary and replica nodes before enabling NRS.

Before you begin

Contact IBM Support to download the Netezza Performance Server Replication Services upgrade bundle.
On each system, unpack the installation bundle on the e1n1 node.
On each system, stop replication. For more information, see Starting and stopping replication on a given node.
Ensure that before you restart NRS on the participating NRS nodes , you upgrade all the systems with the same version of NRS software versions to avoid potential functionality mismatch issues.

Procedure

As root, upgrade Netezza Performance Server Replication Services.
```
./cpds_deploynrs -u
```
Note: If you are using a version of NRS 3.0.4.0 that includes support for lowercase, ensure you include the -l option along with -u during the upgrade process.
On each system, start replication.For more information, see Starting and stopping replication on a given node.
Log in to the Netezza Performance Server host container as the nz user and verify whether upgrade has been completed successfully.
```
./nzdr version
```

Known issues

If the container fails to come up, you might see the following output:

The following NRS container(s): nrs_replmgmt did not come up. NRS is currently offline. To start NRS, please bring up the container(s) listed using docker -H e2n1 start <container_name>.

To fix this issue, perform the following steps:

Bring up the container or containers that are listed in the output:
```
docker -H e2n1 start CONTAINER NAME
```
Verify that the containers are running:
```
docker -H NRS NODE ps
```

Copy the ssh keys from the NRS container to the NPS container:

docker -H NRS NODE exec -it nrs_replmgmt bash -c "ssh-keyscan -p 62000 -H localhost" >
/opt/ibm/appliance/storage/nrs/home/replmgmt/.ssh/known_hosts

docker -H NRS NODE exec -it nrs_replmgmt bash -c "ssh-keyscan -p 51022 -H NPS HOST" >>
/opt/ibm/appliance/storage/nrs/home/replmgmt/.ssh/known_hosts

Start the mirror maker:

docker -H NRS NODE exec -it nrs_mirrormaker mm-start

Run the post upgrade script:

upgrade/nrs_upgrade_helper.py --source-version SOURCE NRS VERSION --target-version TARGET NRS VERSION --post-upgrade-actions --oci=”docker” --host="DOCKER HOST"

Example:

upgrade/nrs_upgrade_helper.py --source-version 3.0.0.0 --target-version 3.0.1.0 --post-upgrade-actions --oci=”docker” --host="e2n1"