Disk types: Difference between revisions
(Created page with "Disk types in Toronto = home = Each user has a home directory. /nfs/home/userid/. There is a default quota of 1 TB. Home is backed up weekly. You will not be able to run Je...") |
(→scratch: asdf) |
||
| (15 intermediate revisions by 2 users not shown) | |||
| Line 1: | Line 1: | ||
Disk types | Disk types | ||
= Desktop Computer = | |||
Note that your desktop computers has the raids mounted but your home is located on the local disk. | |||
local home is here: | |||
/home/username | |||
The cluster's home is here: | |||
/nfs/home/username | |||
= home = | = home = | ||
Each user has a home directory. /nfs/home/userid/. There is a default quota of | Each user has a home directory on the cluster. /nfs/home/userid/. There is a default quota of 750 GB for lab members, 250 GB for guests. /nfs/home is backed up weekly. You will not be able to run Jens-scale docking runs in home, and you will need to use work instead for large jobs. There is also a home directory local to each workstation /home/userid, which is not on the cluster and is not backed up. We really hope this does not cause confusion! We suggest you create ln -s /nfs/home/userid clusterhome on the workstation. | ||
/nfs/home/ | |||
= work = | = work = | ||
Each user has a work directory. /nfs/work/userid/. There is a default quota of | Each user has a work directory. /nfs/work/userid/. There is a default quota of 750 GB for lab members, 250 GB for guests, and can be increased on request. Work is never ever backed up. For critical files, we urge you to be proactive and copy files to either home or store, below. | ||
/nfs/work/users/ | |||
/nfs/work/projects/ | |||
= store = | = store = | ||
Each user has a store directory. /nfs/store/userid/. There is a default quota of | Each user has a store directory. /nfs/store/userid/. There is a default quota of 750 GB for lab members, 250 GB for guests, and can increased on request. Store is backed up monthly. The purpose of store is to hold data such as x-ray data collection, completed or static projects, and other files that do not change but need to be accessed. Departed lab members may be moved to store. | ||
/nfs/store/users/ | |||
/nfs/store/archive/ | |||
/nfs/store/ucsf/ | |||
/nfs/store/backup/ | |||
= db = | = db = | ||
Each user may request a db directory. /nfs/db/userid/. There is | Each user may request a db directory. /nfs/db/userid/. There is a default quota of 750 GB for lab members, 250 GB for guests, and can be increased on request. DB is backed up at best quarterly. The purpose of db is to hold databases that are accessed on an ongoing basis, and change slowly if at all. | ||
/nfs/db/users | |||
/nfs/db/zinc/ | |||
= scratch = | = scratch = | ||
There are high speed local scratch areas available which can help your jobs run faster. Ask the [[sysadmin]] for a free consultation. | |||
The scratch directory is located on each node and uses that mechine's local disk. This scratch directory is located here: | |||
/scratch | |||
These files will only be readable by this machine and no other. This is for temporary use during calculations. | |||
[[Category:Internal]] | [[Category:Internal]] | ||
Latest revision as of 04:13, 3 June 2015
Disk types
Desktop Computer
Note that your desktop computers has the raids mounted but your home is located on the local disk.
local home is here:
/home/username
The cluster's home is here:
/nfs/home/username
home
Each user has a home directory on the cluster. /nfs/home/userid/. There is a default quota of 750 GB for lab members, 250 GB for guests. /nfs/home is backed up weekly. You will not be able to run Jens-scale docking runs in home, and you will need to use work instead for large jobs. There is also a home directory local to each workstation /home/userid, which is not on the cluster and is not backed up. We really hope this does not cause confusion! We suggest you create ln -s /nfs/home/userid clusterhome on the workstation.
/nfs/home/
work
Each user has a work directory. /nfs/work/userid/. There is a default quota of 750 GB for lab members, 250 GB for guests, and can be increased on request. Work is never ever backed up. For critical files, we urge you to be proactive and copy files to either home or store, below.
/nfs/work/users/ /nfs/work/projects/
store
Each user has a store directory. /nfs/store/userid/. There is a default quota of 750 GB for lab members, 250 GB for guests, and can increased on request. Store is backed up monthly. The purpose of store is to hold data such as x-ray data collection, completed or static projects, and other files that do not change but need to be accessed. Departed lab members may be moved to store.
/nfs/store/users/ /nfs/store/archive/ /nfs/store/ucsf/ /nfs/store/backup/
db
Each user may request a db directory. /nfs/db/userid/. There is a default quota of 750 GB for lab members, 250 GB for guests, and can be increased on request. DB is backed up at best quarterly. The purpose of db is to hold databases that are accessed on an ongoing basis, and change slowly if at all.
/nfs/db/users /nfs/db/zinc/
scratch
There are high speed local scratch areas available which can help your jobs run faster. Ask the sysadmin for a free consultation.
The scratch directory is located on each node and uses that mechine's local disk. This scratch directory is located here:
/scratch
These files will only be readable by this machine and no other. This is for temporary use during calculations.