Using CISL's HPSS for Backups

HAO RAID servers (such as the arrays, mlsodata, hoahu, etc) are setup with RAID6, which means they can lose 2 disks simultaneously and still be ok. However, there is a chance that 3 disks could fail at once, and data will not be recoverable. Although I mention the arrays, the following email is true for any of your critical files.

** PLEASE KEEP A REDUNDANT COPY OF YOUR DATA ELSEWHERE **

CISL offers the High Performance Storage System (HPSS) as a replacement to the Mass Storage System. The following web site contains links for using HSI, storage policies, project space, and getting help:

http://www2.cisl.ucar.edu/docs/hpss

HAO has shared CISL project space of 100TB (path= /glade/p/hao). Folks are encouraged to create directories and store files there.

CISL gives users a 10GB home directory. They also provide 10TB of scratch space (not backed up, scrubbed every 60 days). See more information about GLADE file spaces There is also 512GB of user work space.

The command "gladequota" is useful on yellowstone and other CISL machines. Here is sample output:

[username@yslogin2 2010]$ gladequota
Current GLADE space usage: user

SpaceUsedQuota%Full
/glade/p/hao29.24 TB100.00 TB29.24 %
/glade/p/work/username354.40 GB512.00 GB69.22 %
/glade/scratch/username4.05 TB10.00 TB40.50 %
/glade/u/home/username7.38 GB10.00 GB73.80 %


Ben Foster put together an hsi "cheat-sheet," which includes examples of get and put files on the hpss via hsi:
http://www.hao.ucar.edu/modeling/tgcm/hsi_commands.html

Hsi is the interface to the hpss. One can login to an hsi session, use basic cd and ls commands to get around on the hpss, transfer files on the command line, etc.

For more info, see: http://www2.cisl.ucar.edu/resources/glade
and http://www.mgleicher.us