Sap Data Hub Pods back-off restarting failed

[former_member592211]

Hello there

I'm trying to install sap data hub ond my demo landscape.

I have a master server with: Centos 7, 64 GB Memory, 4 cores and 128 GB harddisk.

3 Nodes with: Centos 7, 64 GB Memory, 4 cores and at least 64 GB harddisk.

On all Servers I'm useing Kubernetes version: 1.10.1, Docker version: 17.03.0-ce, Helm version: 2.9.1

For the installation I'm useing the command-line tool (install.sh).

The installation fails every time because some pods get stock in back-off restarting.

If I delete the namespace and restart the installation, other pods can also get this error. It seems like to be random wich pod work and wich not.

[pascaldp]

The restarting is because the pods failed to start correctly and after a timeout period Kubernetes just try to restart them. Did you check the pod events and pod logs for more information on why they do not start? If you doing an on-premise installation, how are your Persistent Volumes provided? Are to pods able to do a Persistent Volume Claim? Do you rollback the installation with install.sh --purge or --delete?

[architectSAP]

Hello Miguel,

Did you not state that you had:

• a master server with: Centos 7, 64 GB Memory, 4 cores and 128 GB harddisk
• 3 Nodes with: Centos 7, 64 GB Memory, 4 cores and at least 64 GB harddisk

That is not sufficient regardless what your underlying servers got.

Please check the Minimum Sizing of Kubernetes Cluster requirements.

Best regards

Frank

[architectSAP]

Hello Miguel,

Your environment does not fulfill the Minimum Sizing of Kubernetes Cluster requirements.

For the hana-0 pod to launch successfully there needs to be at least 24 GB main memory available on one of the Kubernetes nodes.

Best regards

Frank

[former_member592211]

Hello Frank,

All my servers have 64 GB of Memory.

So I dont think that this is the problem.

Regards

Miguel