I'm running Vora 1.4 on a Centos 7.5 single node. The installation was fine, and I can access the Vora Manager.
However the Vora Worker always stops. Here is what I see in /var/log/vora/vora-manager/vora-manager.log:
May 7 22:47:29 centos vora.vora-manager-worker: [c.92eecff1] : Stopping component due to task status. #012Service task scheduler-client status is 1 May 7 22:47:29 centos vora.vora-manager-worker: [c.92eecff1] : Stopping due to waitfortasks error. May 7 22:47:29 centos vora.vora-manager-worker: [c.92eecff1] : Stopping component May 7 22:47:29 centos vora.vora-manager-worker: [c.92eecff1] : Stopping service registration schemes May 7 22:47:29 centos vora.vora-manager-worker: [c.92eecff1] : Stopping tasks May 7 22:47:29 centos vora.vora-manager-worker: [c.92eecff1][task-scheduler-client] : Killing task: scheduler-client pid: 12298 May 7 22:47:29 centos vora.vora-manager-worker: [c.92eecff1][task-scheduler-client] : Killed May 7 22:47:29 centos vora.vora-manager-worker: [c.92eecff1] : Deleting running path May 7 22:47:29 centos vora.vora-manager-worker: [c.92eecff1] : Component stopped
Some advise would be great. Cheers Gerd
The log entries seem to be unrelated to the failing VoraMgr worker - they actually show that it was working when those entries were created. The Vora Manager Worker should start a 'nomad' process. Could you check if this process is still running? If not, could you try to manually start it:
If this fails, please check if the error is covered in the Vora 1.4 Troubleshooting Guide.
Also could you please share your Hadoop distribution + version and Vora 1.4 version.
Vora 1.4 is rather old by now. I would strongly recommend upgrading to a current version such as Vora 2.1 or DataHub 1.3. Please be aware that in those current versions the deployment model has changed - the installation is now in a Kubernetes cluster instead of installing Vora alongside the Hadoop nodes. There is no more Vora Manager.