Vertica Management Console 6.0.2 shows all nodes in UNKNOWN/OTHER state when they are UP against 6.0
When using MC 6.0.2 (the only version available for download on 6.0 page) against a 6.0.0 cluster I see all nodes as UP for a brief moment after startup of verticad-console, then after a period of time (~3mins) all nodes go into UNKNOWN/OTHER state. Then the UI locks up. This cluster has public interfaces that I used when setting up the cluster, and the private interfaces are not routable. I see these errors in the mconsole.log file (10.90.141.X are the private, non-routed IPs): 04 Sep 2013 15:21:45,365 [org.springframework.scheduling.quartz.SchedulerFactoryBean#0_Worker-4] AgentConnectionServiceImpl INFO - getIpBeingTried>> !*!*!*!*! could not connect to 10.90.141.13 (from agentconn) 04 Sep 2013 15:21:45,367 [org.springframework.scheduling.quartz.SchedulerFactoryBean#0_Worker-4] AgentConnectionServiceImpl INFO - theCommand = https://10.90.141.13:5444/ httpMethod:GET 04 Sep 2013 15:21:45,378 [org.springframework.scheduling.quartz.SchedulerFactoryBean#0_Worker-4] DefaultHttpClient INFO - I/O exception (java.net.NoRouteToHostException) caught when connecting to the target host: No route to host 04 Sep 2013 15:21:45,378 [org.springframework.scheduling.quartz.SchedulerFactoryBean#0_Worker-4] DefaultHttpClient INFO - Retrying connect 04 Sep 2013 15:21:45,385 [org.springframework.scheduling.quartz.SchedulerFactoryBean#0_Worker-4] DefaultHttpClient INFO - I/O exception (java.net.NoRouteToHostException) caught when connecting to the target host: No route to host 04 Sep 2013 15:21:45,385 [org.springframework.scheduling.quartz.SchedulerFactoryBean#0_Worker-4] DefaultHttpClient INFO - Retrying connect 04 Sep 2013 15:21:45,392 [org.springframework.scheduling.quartz.SchedulerFactoryBean#0_Worker-4] DefaultHttpClient INFO - I/O exception (java.net.NoRouteToHostException) caught when connecting to the target host: No route to host 04 Sep 2013 15:21:45,392 [org.springframework.scheduling.quartz.SchedulerFactoryBean#0_Worker-4] DefaultHttpClient INFO - Retrying connect 04 Sep 2013 15:21:45,398 [org.springframework.scheduling.quartz.SchedulerFactoryBean#0_Worker-4] AgentConnectionServiceImpl ERROR - Error connecting to 10.90.141.13. Exception:No route to host 04 Sep 2013 15:21:45,399 [org.springframework.scheduling.quartz.SchedulerFactoryBean#0_Worker-4] AgentConnectionServiceImpl ERROR - java.net.NoRouteToHostException: No route to host I've tried one of the solutions of adjusting cache_size of the agent on all nodes and restarting them, then adjusting mctl for -Dvertica.commit_size=200 but we still have the same issue. MC node: [root@vertadmin1 ~]# rpm -qa | grep vertica-console vertica-console-6.0.2-0.x86_64 Vertica data node: [root@vert1 ~]$ rpm -qa | grep vertica vertica-6.0.0-3.x86_64
0
Comments
Try checking on the cluster node, you are importing from, whether the agent is running. And if not start it.
netstat -a | grep 5444
/etc/init.d/vertica_agent stop
/etc/init.d/vertica_agent start