Installing  BioHPC on cluster

• Install files on head node (or server)
  
  • A network drive accessible from nodes and servers is required for file sharing. The drive designation should be H. Set up the H: drive on the file server (which may be the same as the web server or head node), export it for IDU. File will be stored in H\CBSU directory.
    
    
  • Copy files/directories from distribution directory network_drive to H:\CBSU
    
    
• Install and configure nodes for BioHPC
  
  • Copy the install_node directory from the distribution to the head node. Export this directory to the nodes (only required for installation).
    
  • Edit install_node scripts install.bat and set_perm.bat. There are several parameters that need to be set like server names, path to install_node directory.
    
  • Edit the Perl scripts in the mountdrive_node subdirectory. These scripts define the service allowing the H: drive to be accessed on the nodes, so server name and path must be updated.
    
  • Edit the Perl scripts in mountdrive subdirectory. This is same as the previous one, but for the head node.
    
  • Create local drives on the nodes for local file storage and scratch space. Recommended name is T:, but it can be mapped to T: as well. This drive must be exported with fll access rights to IDU in order to use BLAST database propagation tools. Exported share name should be t-drive.
    
  • Run script \\path-to-install-node\install_node\install\install.bat on the nodes. The script can be run via CCS feature "Run on node".
    
  • Make sure Perl file extension (.pl) is registered properly, i.e. double click on file script.pl will execute this Perl script with Perl engine. If by any reason .pl is note associated with Perl you need to run accoc.bat script from the same directory to create the association.