Question

Nutanix CE - Cannot get to Prism Page

  • 23 June 2019
  • 3 replies
  • 6228 views

I have setup Nutanix CE in a lab environment but I cannot get to the Prism page when browsing to the CVM IP address. I'm getting the message:

Prism services have not started yet. Please try again later.

Although I have run cluster status and the Prism service is up and running:



It is a single node setup on VMWare. I can ping the CVM IP and SSH onto the CVM.

Any assistance would be gratefully received!

Thanks.

This topic has been closed for comments

3 replies

Hi I had the same issue and thought I would share my findings.

I’m running the CE on a laptop (lenovo t490s 32GB) using windows 10 as the OS and using VMware Workstation.

After a successful installation (I assigned 16GB to the cluster) I noticed the same behavior the OP mentioned, Prism services running but error when going to the CVM IP on my browser.

 

I checked the /home/nutanix/data/logs/prism.out log to see what additional information I could find about the service and found the following:

nutanix@NTNX-186d6f38-A-CVM:192.168.144.11:~/data/logs$ grep -i restart prism.out
E0227 10:30:28.228792 17190 prism_monitor.cc:1202] notification=TomcatFrequentRestart cvm_ip=192.168.144.11 message=prism monitor starting service_vm_id=2 (is_resolve_notification=false)

So the tomcat webserver seems to have been restarting, I then went over to check the catalina.out logs, and found this:

/usr/local/nutanix/apache-tomcat/bin/catalina.sh: line 467: 25354 Killed "/usr/lib/jvm/jre-1.8.0/bin/java" "-Dnop" -Djava.util.logging.manager=org.apache.juli.ClassLoaderLogManager -Djdk.tls.ephemeralDHKeySize=2048 -Djava.protocol.handler.pkgs=org.apache.catalina.webresources -Dorg.apache.catalina.security.SecurityListener.UMASK=0027 -Xmx512M -XX:MaxMetaspaceSize=256M -XX:ReservedCodeCacheSize=100m -XX:+UseConcMarkSweepGC -XX:+CMSParallelRemarkEnabled -XX:+UseCMSInitiatingOccupancyOnly -XX:CMSInitiatingOccupancyFraction=65 -XX:ErrorFile=/home/nutanix/data/cores/prism_hs_err_%p.log -XX:+HeapDumpOnOutOfMemoryError -XX:OnOutOfMemoryError=kill_tomcat

So it seems that my CVM was somehow running out of memory and thus the prism service was not correctly working.

 

Coincidentally I was running out of battery during the troubleshooting and decided to plug my laptop into a power source. Prism then started working!

 

After some testing, it seems that the power settings of my OS were hindering the usage of the CPU and/or RAM while running on battery only.  So the solution for me was to plug in my laptop to a power source.

 

I hope this helps some one in the future :)

Badge
Did you ever get this resolved and if so what was your fix. I have run into the same issue. I have 1-one node cluster up an running and am attempting to get the second one node cluster up. After trying various option, swapping out hardware, changing ip subnets, after downloading new CE image, it worked. I shutdown and added additional memory on the host, now i am back to same issue, services all started up, but cant get to gui. i have blown away a couple of times with now luck. I have combed all the logs on the cluster that does not work and alot of the entries are the same with the cluster that i have up and running. I am not sure what is going on, after the install it indicates success, i can ping both host and cvm from any machine in my lab and the cvm and host can ping out with no issue. Unable to access webgui. i have tried all the options that have been posted. It just magically worked one time, but now not working again. I know it works on my hardware as i have it running up on one node and the other node has installed correctly once. I have a c6220 2 node chassis. Let me know

@ericgar haha! Danke dir! Wie crazy! :-D War bei mir die Lösung.