XT commands

There are several commands starting with xt which show information about the system configuration, jobs on computing nodes etc. (for more information see manual pages):

xtps
xtps -ef

• Displays all  processes running on login nodes.
• With option -ef shows full status including users
• Please note, that man page xtps(1) is partly out of date and the yod utility is not anymore available on Louhi. It is replaced with the ALPS aprun urility.

xtnodestat (replaced the commands xtshowcabs and xtshowmesh in CLE 2.2)

• The xtnodestat command displays current node (process) allocation and  information on running jobs. The output is a textual two-dimensional grid  where each node is represented by its status or currently assigned job.  If no options are specified, xtnodestat displays node status and allocation grid, a legend, and a job listing. Only node information is shown with with the option -a or -m and only job information is shown with the option -j.
• There are 3 cages (chassis) in a cabinet of XT4 or XT5, each cage contains 8 slots for blades (blade computers), and each blade contains 4 (compute blade) or 2 (service blade) nodes. For example, C0-0 is the cabinet 0 in the row 0, c0 is cage 0, s00 is slot 0 (for blade 0) and n0 is node 0.
• When job status is shown then all aprun processes running on login nodes, also those which are running applications interactively in interactive compute nodes and for which aprun command is given directly from a command line of a login node. The command apstat also shows all these jobs, but the command qstat does not show jobs started with interactive aprun processes.  Job IDs given by xtnodestat are those given by aprun to jobs (apids) and also shown by apstat. The command qstat shows job ID numbers given the the PBS queuing system.

              Please note that the commands  xtshowcabs and xtshowmesh were removed in CLE 2.2 and replaced with xtnodestat.

xtprocadmin

• Mainly for administrative purposes, but shows without options the node ids (NIDs) and node names similarly as two commands above, status of all nodes, including compute nodes (up or down), is the node a service or a compute node, is it for batch or purely interactive jobs, how many processors slots each nodes has and how many of them are free.

ping_node

• Shows information about compute nodes, including processess running on nodes

xpbs

• Graphical User Interface for all PBS Pro commands. It uses also different forms of qstat commands (for example the qstat -B -f command) and shows job, queue and server status in different formats. It can also be used for writing or editing loaded job scripts and the script can be submitted from xpbs for excution.

Status of login nodes

Commands which show information about processes and users on login nodes (commands starting by letters xt shows information for all login nodes). See above and manual pages for more information.

top

ps

xtprocadmin

xtps (the option -f or -ef must be given for full status)

xtwho

xpbs

See also

Job launching command: aprun

General information about the batch system