Where are the scratch areas?

The following table describes all scratch areas

name	location	Size	Usage
/scratch/fs1a	Roque	10 GB	for general use
/scratch/fs4a	SLO	11 GB	for CFG use^*
/scratch/fs4b	SLO	67 GB	for Astronomy use^*
/scratch/inta	INT	100 GB	for INT Observer's use
/scratch/intb	INT	75 GB	for INT Observer's use
/scratch/jkta	JKT	8 GB	for JKT Observer's use
/scratch/whta	WHT	99 GB	for WHT Observer's use
/scratch/whtb	WHT	60 GB	for WHT Observer's use
/scratch/whtc	WHT	16 GB	for WHT Observer's use

Any recent additions to this list?: If you want to see if any more have appeared since this documentation was last updated type the following command
Any recent changes to size?: To see the size you must use two commands. First wake up the automounter using ls then use df
Notes

Any restrictions on use of scratch area?

If you are not the observer please do not use the areas reserved for "... Observer's Use"

At the time of writing the trivially small /scratch/whtb is served from a poor old sunos sparc. This sparc if feeble but has a role in the Real-Time Display. To conserve resources on this old sparc the area /scratch/whtb can not be accessed outside of the WHT.

Which scratch area should I use?

Usually you will use a scratch area which is nearest to you. If working at the Sea-Level Office you would use /scratch/fs4a. If working at the Roque you would use /scratch/fs1a. If you are the observer then use the scratch area associated with the telescope where you have your observing run.
The show_scratch command which was is in /usr/local/bin is not available - sorry

please keep the area tidy - create a top level directory derived from your name if on a general scratch area but by DATE if you are the observer.

Example 

cd /scratch/fs1a
mkdir guest12
cd guest12
...

Example 

cd /scratch/inta
mkdir 20000831
cd 20000831
...

What if the scratch area is full?

Scratch disk space management is dynamic. There is no need for the system manager or anyone else to be asked to delete files. Every 4 hours a crontab job examines free capacity on the scratch disks. If the free capacity on any disk drops below 20% free then the oldest files on the disk are deleted if they are less than 21 days old. By oldest files we mean the files with the oldest accession date. Files are unconditionally deleted after 60 days of not being accessed.

IMPORTANT

If you restore files from a tape to disk then they will appear to be "old". To avoid them being selected as targets you should use a dummy tar job to access all the restored files. Use the null device (/dev/null) for output - or simply a dummy tape. Using a scratch area with lots of free space will help to avoid this problem.
If you want scratch disk space then just go ahead and use it. If the disk fills up to 80% and you want more then fill it beyond 80% and either use another disk or wait for the disk monitor to react in at most 4 hours. The reaction time of the disk monitor is generally going to be faster than calling the system manager. Average response is 2 hour. In the meantime you could use another disk.
* This system does not operate on these disks.

How secure are my files?

Under the rules of the dynamic disk management you are guaranteed 21 days. The scratch areas are not backed up to tape. In the event of a disk failure or the accidental erasure of your files by another user you could lose your files. You could also lose your files if they are more than 21 days old and you or another user reduces free disk capacity below 20%. If you want to secure your files it is your responsibility to make a backup tape. Please obtain a dat tape and use "tar" or "dump" of your directory.

If by security you mean "privacy" the answer is dependent on the protection you give to your files and in particular the protection code on the directory you create in /scratch/lpssX. Use chmod to restrict access from other users.

This page is available on-line as
http://www.ing.iac.es/~cfg/pub_notes/scrdskman.htm