Edge Private Cloud 4.19.01 5-node installation- Cassandra failed to start

[priniren]

I am attempting to install a 5-node topology of Edge Private Cloud 4.19.01 on Linux RHEL 7.6. I have completed the prerequisites mentioned in https://docs.apigee.com/private-cloud/v4.19.01/installation-overview and downloaded the required RPMs. Key OS related details/modifications :

1) iptables is disabled

2) SELinux is disabled

3) Installation commands being run as root user.

4) SMTP disabled in Edge

5) Substituted HOSTIP with the known IP address. (Reason : The command ($hostname -i ) was returning a MAC address, followed by an IP address. )

I could successfully install Zookeeper and Cassandra on the first node and start them. However, on the second node, Zookeeper installed and started successfully, but Cassandra failed to start. The error log from Cassandra(2nd node) and the system root logs(1st node) is attached. The issue persists even if I uninstall Edge and completely clear all the RPMs.

Any help/ideas are appreciated!

[priniren]

@Dino-at-Google @Anil Sagar @ Google @Sudhee Sreedhara Can you assist?

[dchiesa1]

I don't think I will be of much assistance but maybe @Paul Mibus @Carlos Frias will be able to help or refer this question to someone who can help.

[paulmibus]

It's not clear from the logs what triggered the problem, but right now it's complaining that some files are missing. You mentioned it persists across reinstalls; did you also stop all processes and clear data? Try this on the second node:

/opt/apigee/apigee-service/bin/apigee-all stop
pkill -9 -u apigee
yum -y remove 'apigee-*' 'edge-*'
rm -rf /opt/apigee/*

Then reinstall from step 1. If you still see the error can you post your setup-root.log and /opt/apigee/var/log/apigee-cassandra/* logs?

[priniren]

Thanks for the response @paulmibus

Yes i attempted stopping all the processes. Later, I figured out that the ports were not getting released. I killed the PIDs of the processes which were holding up the ports and the issue got solved. I could manage to install components till Node 4 which brings up Postgres in master mode. I can live without the Postgres slave for now, till I test the installation.

However, I am facing a new issue regarding the uploading/creation/deployment of API proxies. When I uploaded the helloworld proxy to the newly onboarded org and env, the proxy did not get uploaded. Where should I look for the logs? I checked up the logs in /opt/apigee/var/<component name>/logs for both MS, Router and MP but I did not find any logs being written regarding the below issue.

Any clues for debugging will be appreciated.

[paulmibus]

You can try checking /opt/apigee/var/log/edge-management-server/logs/system.log to see if any errors are occurring during the upload. If you don't see errors there it's possible it's a client-side error instead. Check the debug console in your browser to see if it's having trouble accessing any remote resources.

[priniren]

Thanks @paulmibus. There was an error in the imported zip. It included a "secure" VirtualHost which was not yet configured in the Edge server. I removed it from the proxy bundle and the import was successful.