Документ взят из кэша поисковой машины. Адрес
оригинального документа
: http://www.stsci.edu/ops/computing/workstation-list.html
Дата изменения: Tue Mar 22 16:25:53 2011 Дата индексирования: Sat Mar 1 07:42:57 2014 Кодировка: Поисковые слова: п п п п п п п п п п п п п п |
The HST planning and scheduling computing environment primarily consists of the Unix "SOGS" cluster, which includes about 30 Sun machines and about 10 MacOSX machines.
Although a workstation is on your desk, we strongly encourage everyone to share workstations (CPU and disk space) to the degree consistent with getting your job done. When you consider that most workstations are idle most of the time, you see we have a tremendous amount of computing power to share. Look here for information on tapping into the CPUs on other workstations.
Likewise, we have an enormous amount of disk space - 659 Gb on the server and serveral times that amount on the workstation local disks. There's a lot of empty space, and the fact that all disks are cross-mounted makes it easy to use them. Look here for hints on how to do this.
Wizard is our high performance file server. It has 659 Gb of disk space and uses RAID 5.
Cerberus is the web server inside the SOGS firewall, and it handles mail for accounts that use procmail. Cerberus also propagates software to other servers outside the firewall (currently distributing Contact Scientist software to Sun server Tib).
Balin runs the Legato backup software to manage nightly file backups for all SOGS machines.
The operational Assist and PMDB databases live on the server HAL9000 which is a Fire 480R. (Since Assist and PMDB are on the same server, cross-database queries are possible and this simplifies operations.)
To help assure the performance of Hal9000, a copy of the data is replicated to another server, Zeppo. Read-only queries should be run on Zeppo whenever possible (especially during the debugging phase) so that Hal9000 has peak performance. Zeppo is also used to support development and testing.
clotho (Ultra 45) is the Phase I and Phase II submission server. It is located outside of the SOGS firewall.
Below is a ranking of machine types currently in use by Hubble operations. They are ranked in order from most powerful to least. There are several factors that determine a workstation's speed, so this list is an estimate; but it should give a general idea.