wiki:GCCluster

Version 2 (modified by Pieter Neerincx, 12 years ago) (diff)

--

GCC cluster

The GCC has its own 480 core cluster. The main workhorses are 10 servers with 48 cores, 256 GB, 1 GBit management NIC and a 10 GBit NIC for a dedicated IO connection to a 2 PB shared GPFS for storage.

Servers

FunctionDNSIPComments
User interface nodecluster.gcc.rug.nl195.169.22.156Login node to submit and inspect jobs.
Relatively powerful machine.
Users can run code outside the scheduler for debugging purposes.
scheduler VMscheduler01195.169.22.214Runs Torque's pbs_server resource manager and the maui scheduler.
scheduler VMscheduler02195.169.22.190Runs Torque's pbs_server resource manager and the maui scheduler.
Execution nodetargetgcc01192.168.211.191Runs Torque's pbs_mom
Dedicated test node: only the test-short and test-long queues run on this node.
Crashing the test node shall not affect production!.
Execution nodetargetgcc02192.168.211.192Runs Torque's pbs_mom
Production node: only the default gcc and priority gaf queues run on this node.
Execution nodetargetgcc03192.168.211.193Runs Torque's pbs_mom
Production node: only the default gcc and priority gaf queues run on this node.
Execution nodetargetgcc04192.168.211.194Runs Torque's pbs_mom
Production node: only the default gcc and priority gaf queues run on this node.
Execution nodetargetgcc05192.168.211.195Runs Torque's pbs_mom
Production node: only the default gcc and priority gaf queues run on this node.
Execution nodetargetgcc06192.168.211.196Runs Torque's pbs_mom
Production node: only the default gcc and priority gaf queues run on this node.
Execution nodetargetgcc07192.168.211.197Runs Torque's pbs_mom
Production node: only the default gcc and priority gaf queues run on this node.
Execution nodetargetgcc08192.168.211.198Runs Torque's pbs_mom
Production node: only the default gcc and priority gaf queues run on this node.
Execution nodetargetgcc09192.168.211.199Runs Torque's pbs_mom
Production node: only the default gcc and priority gaf queues run on this node.
Execution nodetargetgcc10192.168.211.200Runs Torque's pbs_mom
Production node: only the default gcc and priority gaf queues run on this node.

PBS software / flavour

The current setup uses the resource manager Torque 2.5.12 combined with the scheduler Maui 3.3.1.

Maui

Runs only on the schedulers with config files in

/usr/local/maui/

Torque

Torque clients are available on all servers.
Torque's pbs_server deamon runs only on the schedulers.
Torque's pbs_mom daemon runs only on the execution nodes where the real work is done.
Torque config files are installed in

/var/spool/torque/

Dual scheduler setup

Installation details

Attachments (9)

Download all attachments as: .zip