Upgrading to Cloud services 1.1.4 from 1.1.2.x

This topic describes the procedure for upgrading to Cloud services 1.1.4 from 1.1.2.x

Ensure the following before you start the upgrade the process:

Uninstalling the old rpm (rpm -e command) and then installing the new rpm (rpm -i) will not work. The upgrade must be done with the rpm -Uvh command only.
Transparent cloud tiering service is stopped on all the server nodes by using the mmcloudgateway service stop command.
IBM Storage Scale is at release 5.0.0. Upgrade might fail if your IBM Storage Scale release is different.
File system version is upgraded to the latest (18.00 and beyond) by using the mmchfs command, to leverage the default transparent recall policy. If file system version is not upgraded to a given level prior to Cloud services rpm upgrade, enabling transparent recalls for a container that is created with earlier Cloud services versions will fail. If file system version is upgraded to 18.00 and beyond after upgrading the Cloud services rpms, you must restart Cloud services by using the mmcloudgateway service command to be able to effectuate transparent recalls for containers.
No data migration or recall is in progress.

Perform the following steps:

Copy the gpfs.tct.server-1.1.4.x86_64.rpm file to each of the nodes that is specified as Transparent cloud tiering server nodes.

Run this command:

rpm -Uvh gpfs.tct.server-1.1.4.x86_64.rpm

To upgrade the client nodes, issue this command:
```
rpm -Uvh gpfs.tct.client-1.1.4*.rpm
```
To upgrade the Debian package, issue this command:
```
sudo dpkg -i gpfs.tct.client-1.1.4*.deb
```

Verify the version by using the mmcloudgateway service version command. The system displays output similar to this:


Cluster minReleaseLevel:   5.0.0.0

 Node  Daemon node name             TCT Type          TCT Version     Equivalent Product Version
------------------------------------------------------------------------------------------------
    1  c350f1u1b7.pok.stglabs.ibm.com  Server            1.1.4           5.0.0

Note: This upgrade takes longer than others because of the database schema change. The larger the database of migrated files, the longer this upgrade will take.

Note:

After the upgrade, the previous sensors are no more valid and Cloud services performance monitoring will fail on the GUI. Therefore, you must remove the old sensors and populate new ones. For more information, see Upgrading the Cloud services sensors.
After the upgrade, do not use any transparent recall policy that was applied in IBM Storage Scale 4.2.x and earlier releases. You might experience hangs during transparent recalls if the old policies are reused. Therefore, you must remove the older policy (which was applied before upgrade using mmchpolicy <file-system> -P Default) and enable the transparent recall for the container using the mmcloudgateway containerpairset update command, after the upgrade.