Login nodes are the entry point for all users into MASSIVE. Login nodes are for submitting jobs to the cluster, compiling code (single core), transferring files (single process) and light pre and post processing of data.
MASSIVE has the following login node configuration:
m1-login1.massive.org.au - general login to m1 (Synchrotron side)
m1-login2.massive.org.au - default desktop login (Synchrotron side)
m2-login1.massive.org.au - general login to m2 (Monash side)
m2-login2.massive.org.au - default desktop login (Monash side)
MASSIVE has the following aliases configured:
m1.massive.org.au -> m1-login1.massive.org.au
m2.massive.org.au -> m2-login2.massive.org.au
In general users who login via ssh will use the aliases. Users of the MASSIVE desktop will leave the defaults in the MASSIVE launcher.
These nodes are shared by a number of users and it is important to share the processing power fairly. Also overloading the login node can cause it to break all connections causing other users with interactive sessions to lose their data.
- Run jobs or CPU intensive tasks on the login nodes.
- Run scripts which poll the queue at less than 15 second intervals
Load on login nodes is monitored and if tasks are found to be preventing others from accessing MASSIVE we may need to kill them without notice.
If you have intensive tasks please make use of Interactive Sessions on compute nodes.