RDF(1t,C) AIX Commands Reference RDF(1t,C)
-------------------------------------------------------------------------------
rdf
PURPOSE
Reports free space information on replicated file systems.
SYNTAX
rdf---------------------|
^ | -c | |
| | -f | |
| | -h | |
| | -n | |
| | -w | |
| | -L | |1
| | device | |
| | filename | |
| | gfs_number | |
| +------------+ |
+----------------+
-----------------
1 The default action is to provide information for each
file system which is currently mounted on any site in
the current TCF cluster.
Warning: See restrictions, Chapter 18, AIX Programming Tools and Interfaces.
DESCRIPTION
The rdf command writes to standard output information about total space and
available space on the specified file systems. If the file system is a
replicated file system, information is displayed about each mounted copy of the
file system and additional replicated file system information is also
displayed. The flag arguments restrict which file systems or which copies of
replicated file systems are considered. They also adjust what information is
displayed about each file system.
A file system may be specified by a device name (device), by the name of a
directory or regular file (filename) inside a mounted file system, or by the
global file system number (gfs_number) of a file system.
Each copy of a replicated file system maintains a separate allocation of free
blocks to be assigned to files. The rdf command can be used to monitor the
free space in each copy of a replicated file system. Each file system copy may
run out of free space at different times, for a number of reasons, so rdf
should be used to monitor when file systems run out of space and whether files
Processed November 8, 1990 RDF(1t,C) 1
RDF(1t,C) AIX Commands Reference RDF(1t,C)
have successfully propagated from the primary copy of the replicated file
system to the other copies.
Non-primary copies of replicated file systems may run out of space for the
following reasons:
1. The file system copy has a smaller total allocation of file system blocks.
This usually only applies to secondary file system copies.
2. A new version of a program was installed in the primary copy of the file
system, but a process on another site was running the copy stored there.
The disk space for the old version will not be freed until the process
terminates or the system reboots and fsck is run.
3. The use of shadow pages by the file system requires that new pages be
allocated to a propagated file before the old pages of the file can be
freed. AIX requires that the free space in a file system is as large as
the largest file in the file system.
4. A clri command was run on an inode in the primary copy of a replicated
file system. This deletion will not automatically propagate to the other
file system copies. If clri is not run on the other copies of the file
system, the disk space will remain in use until the inode is used for a new
file.
When a non-primary copy of a file system is full, the recommended procedure is
to delete some files in the file system or use the chfstore or store command to
discontinue the storing of some files in the full copy of the file system.
The fields displayed by rdf for each file system are:
GFS The global file system of the file system. This number is the
same for all copies of a replicated file system.
PK The global file system pack number of this copy of a replicated
file system. Each copy of a replicated file system has a
different pack number. If the file system is not replicated, the
pack number is always 1.
SITE The cluster site name where the file system is currently mounted.
FL File system flags. These flags identify the type of file system.
- Not replicated.
SP System Replicated, Primary Copy.
SB System Replicated, Backbone Copy.
S- System Replicated, Secondary Copy.
UP User Replicated, Primary Copy.
UB User Replicated, Backbone Copy.
U- User Replicated, Secondary Copy.
Processed November 8, 1990 RDF(1t,C) 2
RDF(1t,C) AIX Commands Reference RDF(1t,C)
TOTAL Total number of 1024-byte blocks in the file system.
USED Number of 1024-byte blocks in the file system which are currently
allocated to a file.
FREE Number of 1024-byte blocks in the file system which are currently
available.
%USED Percentage of the total number of blocks which are currently
allocated to files.
LWM Low water mark for this copy of a file system. The LWM will
always equal the HWM on the primary copy of a file system. On a
non-primary copy, the LWM value indicates which file changes
(creations, modifications, deletions) have been successfully
propagated from the primary copy to this non-primary copy of the
file system. Each file system change is assigned a number, and
this number defines the order in which changes are propagated. If
the LWM value of one copy equals the HWM value of the primary
copy, this copy has successfully propagated all changes made to
files in this file system that this site is supposed to store.
Non-replicated file systems do not have a LWM value.
HWM High water mark for this copy of a file system. This indicates
the highest numbered change that has occurred in this file system.
Except for very brief moments, all copies of a replicated file
system should have the same value for HWM. If the file system is
not replicated, the file system does not have a HWM value.
DEVICE The name of the device for the mounted file system. This device
usually resides in the /dev directory on the site where this file
system copy is mounted.
DIR The directory on which a file system is mounted. All copies of a
replicated file system must be mounted on the same directory.
This field is normally runcated to eight characters unless
the -w flag is specified.
FLAGS
-c Display information on only the current synchronization site (CSS) for
the specified file systems. The CSS for a replicated file system is
one of the mounted copies which keeps track of which files are open and
what file record locks are held. If the primary copy of a file system
is mounted, this site will be the CSS.
-f Fast option. Suppresses the DEVICE and DIR columns. The DEVICE and
DIR columns are determined by reading the /etc/mtab files on each site
in the cluster. The -f option can often speed up the command by only
providing the other output fields.
-h Suppress header line. This option displays information without a
header line and thus makes the output more easily analyzed by programs
such as awk.
Processed November 8, 1990 RDF(1t,C) 3
RDF(1t,C) AIX Commands Reference RDF(1t,C)
-n Display site number rather than site name.
-w Do not truncate mounted directory name.
-L Only display information about file systems or file system copies
mounted on the local cluster site.
EXAMPLES
To list information about the root file system:
$ rdf /
GFS PK SITE FL TOTAL UUSED FREE %USED LWM HWM DEVICE DIR
1 1 fafnir SP 115000 115000 4780 96% 654076 654076 diskroot /
1 3 atlas S- 25000 22820 2180 91% 654076 654076 hd2 /
1 4 swords S- 25000 22820 2180 91% 654076 654076 hd2 /
1 7 coins S- 25000 22820 2180 91% 654076 654076 hd2 /
1 13 hath SB 115000 110220 4780 96% 654076 654076 diskroot /
1 14 carmen SB 115000 110220 4780 96% 654076 654076 diskroot /
FILES
/etc/mtab Lists the currently mounted file systems on each cluster
site.
RELATED INFORMATION
See the following commands: "df," "fsck, dfsck," "where," "chfstore," and
"store."
See the mtab and fs descriptions in the AIX Operating System Technical
Reference.
Processed November 8, 1990 RDF(1t,C) 4