The cluster depends on a number of servers. 

`uhhpc` -- our administrative login machine. This runs 
the job control systems, the system that forwards ssh 
connections to headnode1 or headnode2, the web server, 
a MariaDB server, and the LDAP server which controls 
user accounts. If it goes down the system is probably unusable. 

`headnode1` -- normal user login machine 1, serves /home. 
If it goes down the cluster is not usable for normal users. 

`headnode2` -- normal user login machine 2, serves /soft 
and runs some licence servers. Software won`t be available if this goes down. 

`stri-server` -- legacy main NFS server. If it goes down 
some NFS volumes won`t be available. 

`lofar-server` aka lofar.herts.ac.uk. Login machine for 
LOFAR-UK users. Serves /data/lofar. If it goes down LOFAR users will complain. 

`cair-cluster` aka cair-cluster.herts.ac.uk. Login machine and 
server for CACP (formerly CAIR) users. If it goes down those users will complain.