In this blog, we will walk through the process of performing an offline upgrade of IBM Storage Scale using
the Ansible-based installation toolkit. This method allows for a faster upgrade of the all the nodes in the cluster using the offine method, where in you bring down the services cleanly, assign the nodes to offline upgrade and run the upgade steps. The toolkit simplifies many of the steps involved, ensuring that the upgrade is efficient
and minimizing downtime.
Step 1: Download the IBM Storage Scale Package
The first step is to download the latest version of IBM Storage Scale from Fix Central:
1. Visit the IBM Storage Scale Fix Central page.
2. Select the latest self-extracting package that fits your system architecture.
eg: Storage_Scale_Advanced-5.2.2.0-x86_64-Linux-install
3. Place the downloaded package on the installer node.
Step 2: Extract the Self-Extracting Package
Once the package is extracted on the installer node:
This will create the new directory structure at /usr/lpp/mmfs/Storage_Scale_Version/ansible-toolkit
Eg: /usr/lpp/mmfs/5.2.2.0/ansible-toolkit
Go to the above directory, and the CLI command will help you to do all the required operations. For
example:
cd /usr/lpp/mmfs/5.2.2.0/ansible-toolkit
./spectrumscale --help
Tip: After extraction, review the displayed instructions to understand the next steps, which include:
- Cluster installation and protocol deployment
- Upgrading an existing cluster
- Adding nodes, NSDs, or file systems to the cluster
Step 3: Preparing for the Upgrade
Before starting the upgrade, ensure that the current cluster state is accurately captured by the installation
toolkit.
1. Run the below command on the installer node which will install the required Ansible version and
verify the correct environment configuration:
./spectrumscale setup -s InstallNodeIP
2. Populate the cluster definition file by issuing the following command from the installer directory of
the target version:
./spectrumscale config populate -N Admin-Node
This captures the current state of your cluster, including nodes, NSDs, file systems, and protocol
configurations.
3. Perform the following steps to designate all the nodes to go through offline upgrade process.
- Suspend all the protocols nodes (mmces node suspend -N <protocol_node1, protocol_node2......>
- Stop all the services (mmces service stop -N <nfs/smb/hdfs/s3>
- shutdown the gpfs daemons on all the nodes in the cluster (mmshutdown -a)
4. Perform the following steps to designate all the nodes to go through offline upgrade process.
for ex: ./spectrumscale upgrade config offline -N all
Step 4: Run the Precheck
Before starting the upgrade, it is essential to run a precheck to catch any potential issues and also check how ansible toolkit is going to handle the non-protocol and protocol upgrade, which is taken up parallelly.
./spectrumscale upgrade precheck
Step 5: upgrade the cluster.
all the nodes in the cluster which are designated to be upgraded by the offline method use the following command
./spectrumscale upgrade run.
Step 4 :Post upgrade actions
once the offline upgrade process completes, make sure you follow the steps to resume the protocol nodes, start the services back and make sure the cluster is healthy.
- /usr/lpp/mmfs/bin/mmces node resume -N <protocol_node1,protocol_node2....>
- /usr/lpp/mmfs/bin/mmstartup -a ( this in turn will start the services back online, else can explicitly give the command to start them back)
------------------------------
DEEPA A
------------------------------