Administering filesystems

Maintaining free space in filesystems

One important responsibility of the system administrator is ensuring that there is adequate disk space for the users. To do this, you must monitor the free space in each filesystem and take corrective action whenever free space gets too low. The amount of free space depends on the size of the disk containing the filesystem and the number of files on the disk. To determine how much free space is available in a filesystem, see ``Displaying filesystem and directory usage statistics''.

A UNIX system operates best when at least 15% of the space in each filesystem is free. If a filesystem has less than 15% free space, system operation typically becomes sluggish.

When a filesystem has little or no space left to work, the system displays the following:

   NOTICE: Out of space on EAFS dev hd (major/minor)
When the filesystem runs out of space, the system stops any attempts to write to the filesystem. (The error message list the filesystem by its major and minor device numbers, for example hd(1/42) for the root filesystem.) The only way to restore system operation is to delete or reduce files from the named filesystem.

If the free space falls below 15%:

  1. Remove files from the /tmp directory.

    By default, cron clears out the /tmp and /usr/tmp directories once each day. You can define which directories and how often they are cleared in the /etc/default/cleantmp file. See the cleantmp(ADM) manual page.

  2. Remind users to remove unused files.

    You can include a reminder in the /etc/motd (message of the day) file, send e-mail, or send a message to the terminals of all users currently logged in. See the mail(C) and wall(ADM) (``write to all'') manual pages.

  3. Locate large files and ask owners to remove them.

    See Finding files of a certain size.

  4. Locate system files to remove.

    See ``Finding temporary files''.

  5. Clear system log files.

    See ``Checking and clearing system log files''.

  6. Back up and remove unused data.

    For example, delete data in the /usr/adm, such as the sar data in the /usr/adm/sa directory or, if accounting is enabled, the data in the /usr/adm/acct directory.

    Back up these files using the Backup Manager before removing them. See ``Running unscheduled filesystem backups''.

  7. Compress infrequently used files.

    See the compress(C) manual page.

  8. Archive files into one larger file to free up inodes. See the tar(C) manual page.

  9. Add more disk space.

    See ``Adding disk space and restructuring filesystems''.

  10. Add another disk and increase the size of the filesystem using the Virtual Disk Manager.

    See ``Creating filesystems on virtual disks''.

  11. If you are unable to add more disk space and your filesystem is not DTFS, you can regain some space by recreating your filesystems and changing the filesystem type to DTFS. DTFS compression filesystems allow you to compress the data to increase storage capacity. See ``Adding disk space and restructuring filesystems'' for instructions on revising your disk layout. When recreating the filesystem, remember to change the filesystem type to DTFS.

Next topic: Displaying filesystem and directory usage statistics
Previous topic: about inodes

© 2003 Caldera International, Inc. All rights reserved.
SCO OpenServer Release 5.0.7 -- 11 February 2003