csl client installation failing

[Smrithi]

We are trying to install csl client by running the client-bootstrap.sh. But in the end it fails and says client acknowledgement failed. We even reset the docker weave component. Still the issue persists.

Kindly help

[WengWai_C_Intel]

Hi @Smrithi ,

For the CSL client setup, have you added the CSL server IP, virtual IP and CSL client IP addresses into no_proxy for both the CSL server system and CSL client system? If not, you can follow the setup guide to do so to ensure both CSL server and CSL client system no_proxy being configured accordingly, and also run the 2 commands below to restart the services.

sudo systemctl daemon-reload
sudo systemctl restart docker

Take note, when rerun the client-boostrap.sh to setup CSL client, you will need to redownload again as it only can be run once.

[Smrithi]

Hi

We have actually changed all the proxy related files according to our system configuration and have also run the sudo systemctl daemon-reload and sudo systemctl restart docker and followed all the pre-requisites mentioned in the guide. And we are also downloading the  client-boostrap.sh every time we proceed with a fresh installation. Not sure where the problem lies.

[WengWai_C_Intel]

Hi @Smrithi ,

In this case, i'll assume you have done all the pre-requisite setup such as cert update, pre-requisite installation and proxy configured. Can you check further below in order to identify if any issue with the CSL server and CSL client system in your environment?

1) From CSL client system, ping the CSL server system IP to see if any network issue.

2) From the CSL client system, use the API command (refer to csl_api_doc.pdf file) to list the CSL server. Pls provide the output screenshot.

3) From the CSL client system, provide the output log for the command below:

 journalctl -u kubelet.service

4) For the CSL client system, can you check the disk space usage for this system in term of how many % free space available? Use the command below and provide the screenshot for the output.

df

5) And finally, you can try to reboot your CSL client system, run 'sudo swapoff-a' command then redownload the client-boostrap.sh and rerun the client setup again. This is to verify if any network interfaces may not be refreshed by the weaven reset earlier that you have done.

Please share the outcome. Thanks!

[Smrithi]

Hi

I have attached the required SS of all the executed commands. It is still showing the same error while installation.

[WengWai_C_Intel]

Hi @Smrithi ,

The screenshot for the  journalctl -u kubelet.service  command output seem can't be seen as the messages were out from the terminal screen. Can you provide the full log message so we can check if any error or abnormality? And for the disk space utilization, can you also run the df command on CLS server system as well to check the available disk space on the system?

thanks!

[Smrithi]

Hi

I have attached the logs of the journalctl -u kubelet.service  command and also the SS of the df command on CLS server system. Please check.

[WengWai_C_Intel]

From the log, it seem that the weave having issue and not able to establish the connection.

Can you check if you have the 10.weave.conflist file in the /etc/cni/net.d directory and also in /etc/opt/csl/csl-node/config/net.d directory? If yes, can you share with me the content of the 10.weave.conflist file for that 2 directories so that we can run a comparison check?

In case you don't see that, you can check if the file is in a different location such as /etc/kubernetes/cni/net.d directory? If yes, you can try move the net.d directory to /etc/cni directory, then run the commands below before rerun CSL client setup again.

sudo systemctl restart kubelet

sudo systemctl daemon-reload

sudo systemctl restart docker.service

[Smrithi]

Hi

I could not find any of the mentioned directories in the etc directory and hence could not locate 10.weave.conflist file. I did file search in the etc directory too, but could not locate the file anywhere.

[WengWai_C_Intel]

In this case, try to create the /etc/cni/net.d directory and put in the 10.weave.conflist file that i provided in the attachment. Then run the commands below before rerun CSL client setup again.

sudo systemctl restart kubelet

sudo systemctl daemon-reload

sudo systemctl restart docker.service

Note: change the 10-weave.conflist.txt file name to 10-weave.conflist before put it into the net.d directory.

[Smrithi]

I created the /etc/cni/net.d directory and put in the 10.weave.conflist file as suggested. But while running  sudo systemctl restart kubelet command I am getting an error saying kubelet.service not found. I have also attached the SS of it. Meanwhile I proceeded with the installation and am still getting the same error and the installation fails.

[WengWai_C_Intel · Accepted solution]

So far, from the information gathered and looking at the log file, the setup failure is related to your network environment that causing the weave is not being initialized and not started properly. Hence, it caused the client node is not being detected by the CSL server within the cluster. Can you verify your network  configuration, is there any restriction/block on port 6443? This port need to be opened as it is  being used for Kubernetes API server communication.

[Smrithi]

The failure was caused by restriction on port 6443. It has installed now.

[WengWai_C_Intel]

ok, thanks for the update and issue has been resolved.