Cluster Monitoring

The following is a set of useful cluster commands that are available on the cluster lab. They can be run from the master node in the cluster.

• cchk Check the machines on the cluster by pinging the machines. Quick way to check if any machines are down.

• cdate Check the current time and date on each node of the cluster. The master node is a NTP time server (and also a client to another trusted source of accurate time). All the other nodes in the cluster are NTP time service clients of the master node.

• cmips Check the total computing power of the cluster. The computing power is reported in MIPS (using the bogoMIPS value reported by the kernel).

• cfree Reports memory usage across the cluster.

• cdisks Reports /tmp scratch disk space usage across the cluster.

• ctemp Reports CPU temperatures across the cluster. May not be available on all clusters.