AnsweredAssumed Answered

community edition - remove CLDB and NFS services to fix botched installation

Question asked by alexwoolford on Feb 11, 2015
Latest reply on Feb 11, 2015 by nabeel
According to the documentation, the MapR Community Edition "restricts how many instances of certain services can run" (e.g. 1 x NFS; 1 x CLDB). I overlooked this when I setup the cluster and added NFS/CLDB to all the nodes. I suspect that's what's causing the error messages.

I tried stopping the `mapr-warden` and `mapr-zookeepr` services, removing mapr-cldb and mapr-nfs from the nodes I don't want CLDB/nfs to run on via `sudo apt-get remove mapr...`, and running `configure.sh -C node1 -Z node1 node2 node3`. The MCS dashboard still shows that two CLDB's and two NFS Gateways have failed. I'm wondering if this is a threshold that has been set in the MCS dasboard since I don't see `cldb` or `nfs` listed when I do an `ls /opt/mapr/roles` on the nodes where I removed those services.

Q) Are the alarms purely cosmetic? Is the threshold configurable?

Outcomes