Installation issues

Use the troubleshooting tips and tricks in IBM Fusion installation.

Note: To log in to the provisioner node, as a kni user, SSH to service node. It is the same host that the IBM Service Support Representative (SSR) provided you before the stage 2 of the installation: https://<host IP address>:3000/isfsetup. For the actual step, see Installation procedure.

HA MR with static IP

Problem statement: The rack discovery fails for HA MR with static IP.

Resolution

For HA MR static IP setup, do the following workaround:

After you finish the stage 1 installation of rack 1, log in to the provisioning node RHEL.
Run the following command to change the directory:
```
cd /home/kni/isfconfig
```

Run the following command to copy the JSON file:

cp computeNodesDNSResult.json nodesDHCPResult.json

Complete the stage 1 installation of rack 2.
Log in to the provisioning node RHEL.
Run the following command to change the directory:
```
cd /home/kni/isfconfig
```

Run the following command to copy the JSON file:

cp computeNodesDNSResult.json nodesDHCPResult.json

In stage1 of rack3 (last rack of HA MR), go till the stage where all nodes are discovered and before you click Next to start the discovery of the neighboring rack.
Log in to provisioning node RHEL
Run the following command to change the directory:
```
cd /home/kni/isfconfig
```

Run the following command to copy the JSON file:

cp computeNodesDNSResult.json nodesDHCPResult.json

Go back to the stage 1 installation of rack 3 and click Next to start the rack discovery.

Red Hat OpenShift Container Platform cluster degradation

Problem statement: You might observe unpredictable Red Hat® OpenShift® Container Platform cluster degradation issues that include but not limited to severe performance and operator degradation issues.

Resolution

To avoid these issues, go through the Red Hat recommendations. For example, the following list has some of the high-level considerations (not all inclusive):

Your cluster must have sufficient CPU and RAM resources that are allocated to etcd and compute nodes.
The etcd disks must be high-performing with low read or write latency and high IOPs. You can use FIO test to check IOPS for etcd disks.
There must not be high network latency among etcd members.

For more information and a complete list of considerations, see Red Hat Documentation.

DeadlineExceeded error

If you notice that the operator installation or upgrade fails with DeadlineExceeded error, see Operator installation or upgrade fails with DeadlineExceeded error.

Duplicate host entries in Stage2 Network pre-check page

Problem statement

When you retry stage 2, installation can fail due to multiple entries found for nodes in the list.

For example:

For node at RU5 the lookup of IP address from hostname compute-1-ru5.testdomain.com failed with error DNS lookup of hostname compute-1-ru5.testdomain.com returning 2 IPs. Expected exactly 1 Ip....

Resolution

Run the following command to verify duplicate entries:
```
sudo vim /etc/hosts
```
Remove all the duplicate entries from file and save it.
In the Network validation page, click Retry pre-check connection.