Managing File Systems

The process of incorporating a file system into the existing directory structure is known as mounting the file system. The file system can be on a disk or disks connected directly to your system, that is, a local file system, or it can be on a disk on a remote system (see “Importing a File System (HP-UX to HP-UX)”) and it can be on either a logical volume or a non-LVM disk.

Mounting a file system associates it with a directory in the existing file system tree. Prior to mounting, the files, although present on the disk, are not accessible to users; once mounted, the file system becomes accessible.

The directory in the existing file system where the file is attached is known as the mount point or mount directory for the new file system, and the files in the added file system become part of the existing file system hierarchy.

The mount point should be an empty subdirectory on the existing file system. If you mount a file system to a directory that already has files in it, those files will be hidden and inaccessible until you unmount the file system. If you try to mount the file system to a directory whose files are in use, the mount will fail.

You can either use SAM or HP-UX commands to mount file systems.

If you are using SAM, proceed from SAM’s “Disks and File Systems” area menu. You can perform the necessary tasks as part of creating your file system, as already described. For help in mounting files using SAM, see SAM’s online help; instructions for using HP-UX commands follow.

The mount command attaches a file system, on either a non-LVM disk or a logical volume, to an existing directory.

You can also use the mountall command or mount -a to mount all file systems listed in the file /etc/fstab. (See mount(1M), mountall(1M) and fstab(4) for details.)

Mounting Local File Systems

To mount a local file system:

1. Choose an empty directory to serve as the mount point for the file system. Use the mkdir command to create the directory if it does not already exist. For example, enter:

   mkdir /joe

2. Mount the file system using the mount command. Use the block device file name of the file system followed by the name of the mount point, as arguments to the mount command.

   For example, enter

   mount /dev/vg01/lvol1 /joe

Refer to mount(1M) for details and examples.

	NOTE: If you are not using logical volumes, run ioscan -fn -H hw_path to determine the block device file name to use. You can use lssf to display the location associated with the device file and compare it with the actual hardware address of the disk. You can also use ioscan to display the devices connected to your system and their hardware path. If the block device file does not exist, create it using insf or mksf. See Configuring HP-UX for Peripherals, lssf(1M), ioscan(1M), insf(1M), and mksf(1M) for more information on these commands.

To mount a file system automatically at bootup, list it in the /etc/fstab file. See the entry for fstab(4) for details on creating /etc/fstab entries.

Here are some typical problems that are sometimes encountered when mounting a file system and the actions to take to correct the problem. See also “Troubleshooting NFS”.

You can use either SAM or the umount command to unmount file systems located on an NFS remote system.

If the server unmounts, the file system disappears from the client; if the client unmounts, this does not affect access to the file system on the server.

For information on unmounting NFS file systems using SAM, see SAM’s online help.

For information on configuring and troubleshooting NFS mounts, see “Sharing Files and Applications via NFS and ftp ”.

When you execute the shutdown command, the system attempts to unmount all of your mounted files systems except for the root file system which cannot be unmounted. For more information on shutdown, refer to “Shutting Down Systems”.

If umount fails to unmount a file system, check the following:

	CAUTION: Always unmount file systems contained on a mass storage device before removing the device from the system. Removing a device containing mounted file systems (for example, disconnecting or turning off the power to a disk, or removing a disk pack from a mass storage device) will likely corrupt the file systems.

Suppose the current size of a logical volume is 1024 MB (1 gigabyte). Assuming the users of the file system within this logical volume have consumed 95% of its current space and a new project is being added to their work load, the file system will need to be enlarged. To increase the size of the file system, follow these steps:

The following are symptomatic of a corrupt file system:

If you or other users cannot readily identify causes for the difficulties, check the file system for inconsistencies using fsck.

fsck, the file system checker, is the primary HP-UX tool for finding and correcting file system inconsistencies. fsck examines the HFS or JFS file system listed in /etc/fstab.

If the system fails, reboot the system and run fsck(1M). Additionally, if you suspect that a file system is corrupt, or to do periodic preventive maintenance, you should also check the file system.

Refer to fsck(1M), fsck_hfs(1M), and fsck_vxfs(1M) for more information.

Checking an HFS File System

To check an HFS file system, use the following procedure:

1. Before running fsck, make sure that a lost+found directory is present and empty at the root for each file system you plan to examine. fsck places any problem files or directories it finds in lost+found.

   If lost+found is absent, rebuild it using mklost+found(1M).

2. For mountable file systems, prepare to unmount the file system by terminating all processes running on it, closing any open files.

   For the root file system, execute shutdown (without -h or -r) to enter the single-user state. The root file system cannot be unmounted.

3. Unmount the (mountable) file system using SAM or the umount command.

4. Run fsck.

   	NOTE: The -n or -N options run fsck in nondestructive mode, and are the safest options available. You can run them on a mounted file system as a precautionary measure when you suspect you might be having problems. The following text documents the traditional -p option in more detail.

   The -p option of fsck allows you to fix many file system problems, running noninteractively. (See fsck(1M) for information on syntax and options.) If fsck either finds no errors or finds correctable errors, it corrects any such errors and prints information about the file system it checked. If fsck encounters a problem it cannot correct while running with the -p option, it will terminate with an error message.

5. Use the following table to determine what to do next based on three possible outcomes:

   Table 6-6 fsck Results

   If fsck reports...	Proceed to...	Then...
   No errors	Step 6	You are done
   Errors and corrects them all	Step 7	Step 10
   Any uncorrectable errors with an error message	Step 8	Step 9

    

6. Check for other causes of the problem.

   If fsck runs without finding errors, the problem is not a corrupted file system. In this case, consider other possible causes of problems with files:

   • A user deleted, overwrote, moved, or truncated the file(s) in question.

   • A program/application deleted, overwrote, moved, or truncated the file(s).

   • The file system associated with a particular directory when a file was created might not be mounted to that directory at this time (if any are).

   • A file (or group of files) was placed in a directory that now has a file system mounted to it. The files that were in the directory before you mounted the current file system still exist, but won’t be accessible until you unmount the file system that is covering them.

   • The protection or ownership bits on the file prevent you from accessing it.

   Because your file system is not corrupt, do not continue with the remaining steps in this procedure.

7. Restore any necessary files.

   Once fsck finds and corrects all errors it locates in the file system, you may assume that the file system is again structurally sound. If any necessary files were lost, restore them from a backup or from lost+found. Once fsck has repaired the damage, proceed to Step 10.

8. Prepare to run fsck interactively.

   If fsck terminates without correcting all the errors it found, you must run fsck interactively.

   Before doing so, move any critical files on this file system that have not yet been backed up (and are still intact) to another file system or try saving them to tape.

   When you run fsck interactively, it may need to perform actions that could cause the loss of data or the removal of a file/file name (such as when two files claim ownership of the same data blocks). Because of this, any backups of this file system at this point are likely to fail. This is another reason you should back up your system regularly!

   	IMPORTANT: Empty the lost+found directory before running fsck again.

9. Run fsck interactively by reexecuting fsck without the -p or -P option.

   As fsck encounters errors, it will request permission to perform certain tasks. If you do not give fsck permission to perform the correction, it will bypass the operation, leaving the file system unrepaired.

   After running interactively, in many cases fsck will request you do a reboot -n. Failing to do so might re-corrupt your file system. (Note, do not use reboot -n for normal rebooting activities.)

10. Examine files in the lost+found directory.

    Once fsck has repaired the file system, mount the file system and check its lost+found directory for any entries that might now be present. These are files, listed by inode number, that have lost their association with their original directories. Examine these files, determine their name, and return them to their proper location. To do this,

    • Use the file command to determine file type.

    • If they are ASCII text files, you can review them using cat or more to see what they contain.

    • If they are some other type, you will have to use a utility such as xd or od to examine their contents.

    • Run the commands what or strings to help you find the origin of your lost+found files.

Once you have returned the files in the lost+found directory to their proper locations, restore any files that are missing from your most recent backup.

	IMPORTANT: The following message CAN'T READ BLOCK ... may indicate a media problem that mediainit(1) can resolve. Otherwise, hardware failure has probably occurred; in this case, contact your local sales and support office.

If you have HP OnLineJFS, you can reduce the size of a file system using a single command (fsadm). (See fsadm_vxfs(1M) for syntax and also Disk and File Management Tasks on HP-UX for further information.)

If you do not have OnLineJFS, the steps are identical to those shown below for HFS and depend upon whether you are using logical volumes.

If an HFS file system is contained on a non-LVM disk, follow these steps to reduce its size:

If an HFS file system is contained within a logical volume, the logical volume resembles a container with the file system as its contents.

Once a particular file system has been created, you cannot simply issue one command to reduce its size, as you can to extend the file system (described in “Extending the Size of a File System Within a Logical Volume ”). First, you must reduce the size of its logical volume. However, reducing the size of a container too much, that is, to a size smaller than its file system contents, will destroy part of the file system’s contents. Once the container is reduced, you must subsequently recreate a new file system within the container using newfs or SAM, or else if you attempt to access the original file system, you may crash your system. The steps you need to follow are shown below:

Here are the main steps for setting up and turning on disk quotas:

When you unmount a file system, HP-UX automatically turns off disk quotas.

You can turn off disk quotas for a file system without unmounting that file system by using the /usr/sbin/quotaoff command. However, using this command is not recommended because once quotas are turned off, actual disk usage may become inconsistent with the usage information stored in the quotas file, thus requiring quotacheck when quotas are re-enabled. See quotaoff(1M) for more information.

After creating a file that causes a soft limit quota to be exceeded, a user on locally mounted file systems will see a message similar to this:

WARNING: disk quota (/home) exceeded

The user has a limited time to remove unnecessary files. The user will receive no further warnings until he attempts to exceed hard limits or allows the time to expire without reducing usage to normal levels. Once a user corrects his usage levels, the system removes any time constraints.

	NOTE: Users of remote file systems (such as NFS mounts) will not receive soft limit warnings. Thus, users having quotas on remote file systems can reach hard limits without prior warning, so they should frequently check their usage levels using the /usr/bin/quota command. For details on checking levels, see quota(1). Only a user with superuser privileges can use the user option of the quota command to view specific usage and quota information about other users.

When users reach a hard limit or fail to reduce their usage below soft limits within the allotted time, an error message appears on their terminal. If a user reaches a block limit, the following message appears:

DISK LIMIT REACHED - WRITE FAILED

If he reaches a file limit, he sees:

FILE LIMIT REACHED - CREATE FAILED

How to recover from reaching a hard limit depends on whether or not the user was using an editor when the message was received. The next sections describe both cases.

When Not Using an Editor

When not using an editor, follow these steps:

When Using an Editor

When using an editor, the user needs to remove files to a level below the quota limits and still preserve the recent changes made to the file being edited. If possible, a user can do this by opening a new window or by logging in from a remote node. This way, the user can get a shell prompt without aborting the editor. Alternatively, the user can follow these steps:

Or, when using a job-control shell:

You can configure mirroring by using either SAM or HP-UX commands. Whenever possible, use SAM.

Using SAM

SAM will perform the following mirroring set-up and configuration tasks:

	NOTE: The logical volume feature in SAM related to mirroring will not function unless the MirrorDisk/UX subsystem has been added to the system.

Using HP-UX Commands

Table 6-8 “HP-UX Commands Needed to Create and Configure Mirroring ” summarizes the commands you will need to do mirror set-up and configuration tasks when you do not use SAM. Consult Section 1M of the HP-UX Reference for the appropriate command line options to use.