Adding nodes or building blocks

Additional nodes or building blocks in to an existing environment increase storage capacity of a cluster during the deployment or upgrade. You can add building blocks or nodes by using the essrun command or GUI.

Before you add nodes or building blocks, ensure that the following prerequisites are met:

Power9 management server and one or more nodes are deployed in the cluster, which includes three or more quorum nodes.
A server with two canisters is added without code upgrade in the existing management server
The file system includes 4M block size, 8+2p RAID code, 80% default set-size.

Complete the following steps to add nodes to an existing cluster:

Edit the /etc/hosts file to add a building block or nodes and I/O node names to an existing management server including management and high-speed network.
Reinstall the deployed code image if it is not in management server, and run the container clean to bring the new /etc/hosts in the container.
```
essrun -N ems,essio1,essio2,essio3,essio4 config load -p
```

Check whether nodes need any updates.. Start of change

essrun -N ems,essio1,essio2,essio3,essio4 update --precheck

Create network bonds for the added nodes. Start of change

essrun -N essio3,essio4 network --suffix=-hs

Outside the container, issue the essnettest command against the new nodes.
Run this command from one of the nodes in the cluster and use this command to test the health of the high-speed network connections. For more information, see the essnettest command.
Add new nodes to the existing cluster by issuing the following command from the container:
```
essrun -N ems cluster --add-nodes essio3,essio4 --suffix=-hs
```
Note: Add building blocks with the same node type at a time. For example:
```
essrun -N ems1 cluster --suffix=-hs --add-nodes ess3k1a,ess3k1b
```
Do not add building blocks with node type as follows:
```
essrun -N ems1 cluster --suffix=-hs --add-nodes ess3k1a,ess3k1b,ess5k1a,ess5k1b
```
Create node class and recovery groups by issuing the following command from the container:
```
essrun -N essio3,essio4 filesystem --rg-only --suffix=-hs
```
Get the list of recovery groups by issuing the following command from the system host:
```
mmvdisk rg list
```
Define the VDisk set.
Note: For IBM Storage Scale System 3500, you must keep 1.5 TB or more space free if future capacity MES is planned (performance to hybrid). Thus, it is recommended to not use all available space when you create a file system for the performance model. The default allocation is 80% of available space when you use the essrun filesystem command (for x86 nodes).
1. Use the ssh command to connect to one of the new nodes.
2. Define the VDisk set.
```
mmvdisk vs define --vs vs_fs1_essio3_hs_essio4_hs --rg ess3500_essio3_hs_essio4_hs 
--code 8+2p --bs 4M --ss 80% --nsd-usage dataAndMetadata --sp system
```
  Note: Assume that the existing file system has the attributes that are mentioned in the command. For VDisk set name, use a unique name similar to the existing VDisk set.
  For more information, see mmvdisk online command reference.
Add the VDisk set to an existing file system.
```
mmvdisk fs add --file-system fs1 --vdisk-set vs_fs1_essio3_hs_essio4_hs
```
Note: Assume that the original file system name is fs1.
Update the component database on the management server by issuing the following command from system host:
```
mmaddcompspec default --replace
```
Restripe the file system.
```
mmrestripefs fs1 -b
```
Add new nodes to the GUI for monitoring.
```
mmchnode --perfmon -N essio5-hs,essio6-hs
```
To monitor new nodes from the GUI, complete following steps:

Adding nodes to the GUI for monitoring

From the container, issue the following command to configure and start the performance monitoring sensors.
```
cd /home/deploy/ess_7.0.1.x_0604-21_dme_ppc64le.dir
./essmkyml
essrun -N essio3,essio4 gui --add-to-hosts
```
Note: Ensure that the passwordless SSH is configured from new nodes to the management server.
Check whether GUI web page shows new IBM Storage Scale System nodes.
Add the new nodes as sensors, reconfigure call home, and rerun the GUI wizard to identify the new nodes.