Managing Disk Space

Disk space, or the lack thereof, is a common problem for any computer system, and with the archive, it's especially true. The very nature of constantly adding data and removing very little will force you to manage the disk space you have. If you start to run out of disk space, you can look in several places to find files you can delete.

From a Unix standpoint, you may want to use the du -sh */ command. This is a disk usage command that will summarize -- display only a total for each argument -- and print the results in human-readable format. You can perform this at any directory level, but if you do it at the root level, it may take a while to display the results. This tip is from Charlie Steed at the Pensacola News Journal.

The df command is also useful. It will report filesystem disk space usage. On the HP-UX system there is a bdf command, and I've made an alias in our Linux system for bdf. The actual command is df -h, which displays the disk usage in human-readable format.

You also may want to run this script on HP-UX systems (it may also work on Linux, but I haven't tested it):

find / -name core -exec rm {} \;

This command searches for all core files and removes them. Core files are dumped when a process crashes, and it's contents can be used in troubleshooting. However, these files typically are huge, and can take up valuable disk space.

Other disk-space hogs are log files. They will grow to enormous size very quickly. Most of the Digital Collections processes trim themselves at about 2 MB, however, the digicol.messages log file will not. To check it, change to the software directory (type cds at a command prompt) and then type ls -lh(that's el, not one). On Red Hat Linux 6.2, this will list the contents of the directory and show you in human-readable format the amount of disk space (fifth column) the file is occupying. If you omit the h switch, you will get the same listing, but the size will be in bytes, and it's much harder to read (see below).

#cds
#ls -l
total 234688
-rw------- 1 digicol dcarchiv 72420 Jan 5 2001 DCJACK~1.GZ
drwxrwxrwx 2 digicol dcarchiv 4096 Jan 10 2001 dcatexin
drwxrwxrwx 2 digicol dcarchiv 4096 Nov 12 14:03 dcbuild
drwxrwxrwx 2 digicol dcarchiv 4096 Jan 26 2001 dccity
drwxrwxrwx 3 digicol dcarchiv 4096 Sep 18 2000 dccopy
drwxrwxrwx 2 digicol dcarchiv 4096 Dec 11 1997 dcdisplay
drwxrwxrwx 2 digicol dcarchiv 4096 Jul 5 08:02 dcfail
drwxrwxrwx 2 digicol dcarchiv 4096 Sep 18 2000 dcfind
drwxrwxrwx 2 digicol dcarchiv 4096 Sep 7 2000 dcftp
drwxrwxrwx 2 digicol dcarchiv 4096 Oct 3 2000 dcgetthumb
drwxrwxrwx 2 digicol dcarchiv 4096 Dec 11 1997 dchash
drwxrwxrwx 2 digicol dcarchiv 4096 Oct 15 21:36 dcinsert
drwxrwxrwx 2 digicol dcarchiv 4096 Dec 11 1997 dcisofs
drwxrwxrwx 2 digicol dcarchiv 4096 Dec 11 1997 dcisomak
drwxrwxrwx 2 digicol dcarchiv 4096 Jul 31 10:07 dcjackobit
drwxrwxrwx 2 digicol dcarchiv 4096 Dec 11 1997 dcjpeg
drwxrwxrwx 2 digicol dcarchiv 4096 Apr 1 2001 dcjukebox
drwxrwxrwx 2 digicol dcarchiv 4096 Oct 19 21:41 dckamp
drwxrwxrwx 2 digicol dcarchiv 4096 Sep 19 1994 dckocom
drwxrwxrwx 2 digicol dcarchiv 4096 Dec 6 1998 dclienau
drwxrwxrwx 2 digicol dcarchiv 4096 Nov 6 16:12 dclocalpic
drwxrwxrwx 3 digicol dcarchiv 4096 Jul 5 08:02 dcmixer
drwxrwxrwx 2 digicol dcarchiv 4096 Apr 1 2001 dcpacker
drwxrwxrwx 2 digicol dcarchiv 4096 Nov 22 20:41 dcphoto
drwxrwxrwx 2 digicol dcarchiv 4096 Oct 22 10:32 dcpiciptc
drwxrwxrwx 2 digicol dcarchiv 4096 Dec 11 1997 dcping
drwxrwxrwx 2 digicol dcarchiv 4096 Dec 11 1997 dcreadp
drwxrwxrwx 2 digicol dcarchiv 4096 Dec 11 1997 dcreadt
drwxrwxrwx 2 digicol dcarchiv 4096 Dec 11 1997 dcrender
drwxrwxrwx 2 digicol dcarchiv 4096 Sep 26 2000 dcshell
drwsrwsrwx 2 digicol dcarchiv 4096 Sep 26 2000 dcstartup
drwxrwxrwx 2 digicol dcarchiv 4096 Jul 5 08:02 dctcpserver
drwxrwxrwx 2 digicol dcarchiv 4096 Jul 5 08:02 dcthesaur
drwxrwxrwx 2 digicol dcarchiv 4096 Sep 24 1994 dctifflib
drwxrwxrwx 2 digicol dcarchiv 4096 Oct 22 10:53 dctrasher
-rw-rw-rw- 1 digicol dcarchiv 239861509 Nov 26 15:01 digicol.messages

You should also look in the /dot/support/appleshare/ directory. Apple volumes are mapped to this directory and files may be inside this directory that shouldn't be. Only you will know if they are valid or not. There is a bug in the digital collections Mac client that allow users to enter more than 31 characters as a filename when they export photos. Because the Mac can't read a file with more than 31 characters, you will never see these files by looking at the Mac volume. The only way to kill these files, is with the rm command in unix.

There also may be spool files in the printer directories. So go to the /dot/spool directory and poke around there for large files. You will always have some files in these directories (such as "errors", "lock" and "status"). They are OK. If you have very large files in those directories, more than likely they are wayward postscript files that never printed, or they may be files waiting to print. "More" the status file to see what's going on with your printer.

The last place to look is in the /var/logs directory. This is where the system log files are kept. Be careful in here.