Administering filesystems

Checking and repairing filesystems

When you boot the system after an abnormal shutdown, the fsck(ADM) utility runs automatically on the root filesystem. You can also set up your other filesystems to be checked automatically before they are mounted -- See ``Modifying filesystem mount configuration''.

NOTE: You cannot check the root filesystem with the Filesystem Manager. To check the root filesystem, you put the system in single-user mode and run fsck(ADM).

To check filesystems manually, in the Filesystem Manager:

Select the filesystem to check from the list.
Select Check and Repair from the Mount menu, then select Selected Filesystem. If the filesystem is not in the list, select Other Local Filesystem.
NOTE: You can only check local filesystems. Check NFS filesystems on the system on which they reside. You can do this remotely using the Open Host selection of the Host menu. This requires user equivalence on the machine you plan to administer, as described in ``Adding user equivalence''. As on the local machine, non-root accounts require the sysadmin authorization to run the Filesystem Manager. See ``Assigning subsystem authorizations'' for more information.
Specify the type of checking and what to do if the filesystem is corrupted, then click on OK. See ``Check and repair options''.

The Filesystem Manager checks the filesystem and displays the output in a Status window; see ``Filesystem check phases (HTFS, EAFS, AFS, S51K)''. When the filesystem has been checked, click on Close.

NOTE: In cases where a file appears to be lost, check for the file in the lost+found directory at the top of the filesystem. Files that have become disconnected from the structure are stored there. If the file is not in lost+found, restore the file from your backups. See ``Restoring files from a scheduled filesystem backup''.

DTFS filesystems do not require lost+found directories.