Hardware

Hardware Nodes and available Storage .xls sheet 

Currently  there are 22 machines which are part of T3_US_UMD cluster.

There are two  10 GB switches, one for public network and other is private private network. 

Identity and Access Management  is done using OpenLdap and is served by  hepcms-henrietta.

The nameservers (DNS) are running on hepcms-hn2 (10.1.0.40)  and hepcms-henrietta (10.1.0.100).

The /home 7.3 TB)disk is mounted on hepcms-hn.

Two data disks are mounted on hepcms-rubin and hepcms-henrietta.  

/dev/mapper/rubin-rubin        87T   52T   31T  63%  /data

10.1.0.100:/data2                     73T   62T   12T  85% /data2

hepcms-hn.umd.edu (10.1.0.1) serves as a gateway. In addition hepcms-hn2.umd.edu (10.1.0.4) is also configured as gateway.

hepcms-rubin.umd.edu and hepcms-henrietta.umd.edu is interactive machine for RHEL9 work.

Most of the machines are running RHEL-Alma9.  

hepcms-namenode.privnet (RHEL8) is the name node for hadoop. 

There are 11 hadoop nodes six on RHEL9 and five RHEL8.  Total hadoop disk is  321 TB.

 /home is being backed-up to /data3 which is mounted on r720-datanfs.privnet (RHEL9).

  hepcms-se2. umd.edu (RHEL8)  is the xrootd server. 

  hepcms-ce2.umd.edu (RHEL8) is the condor-host node. It is also running condor-ce which recieves jobs from the cms central systems. 

 The Frontier Squid server is running on hepcms-squid.umd.edu (RHEL8). It is http://hepcms-squid.privnet:3128 for /cvmfs disks.

  hepcms-franklin.umd.edu (RHEL9) is a R750XA machine which have four Nvidia GPU L40 installed. It also have 1.92 GB SSD disk for use with GPUs.