You must take the following issues into account when you create and administer XFS filesystems with XFS for Linux Release 1.1.
Currently XFS is limited to filesystems smaller that 2 terabytes. This is due to limitations in the Linux block device I/O layers.
The XFS team is working with Linux developers to improve the Linux I/O layers. The improvements will include the support necessary to exceed 2 Tbyte filesystems.
Due to known problems with later versions of gcc, gcc version 2.91.66 (aka egcs-1.1.2) is recommended when compiling XFS and the associated kernel. This is the default compiler on Red Hat Linux 6.2 systems, and is provided as "kgcc" for Red Hat Linux 7.0 and higher (see the next item for more information). Other distributions may be using a later version. The easiest way to check what version is in use is by running gcc -v.
gcc versions 2.95.2 and later may work when compiling the kernel but have not been extensively tested.
Red Hat Linux ships with 2 compilers, "gcc" and "kgcc." kgcc is essentially an older gcc (egcs-2.91.66) which should be used to compile the kernel. To use this compiler, comment out the following line in the top level kernel makefile:
CC = $(CROSS_COMPILE)gcc -V egcs-2.91.66And uncomment the following line:
CC = $(CROSS_COMPILE)kgcc
kgcc is available in the "compat-egcs" package on Red Hat Linux 7.1 and 7.2.
XFS userspace tools release 2.0.0 or later are required for this release of XFS Linux. Instructions for creating the XFS tools are provided Configuring and Installing the XFS Linux Kernel.
EOVERFLOW errors may result when using 2.4.18 as an NFS client and an IRIX server with version 1 directories.
Applying patch http://www.fys.uio.no/~trondmy/src/2.4.13/linux-2.4.13-seekdir.dif. to 2.4.14 and exporting the version 1 directory on IRIX with the "32bitclients" option alleviates this problem.
The official system calls for extended attributes are used in this release. These are getxattr(2), listxattr(2), removexattr(2) and setxattr(2). Extended attributes can be manipulated using the getfattr(1) and setfattr(1) commands.
Library calls implement all of the draft Posix 1003.1e standard functions, such as: acl_copy_ext(3), acl_delete_def_file(3), acl_dup(3), acl_free(3), acl_from_text(3), acl_get_fd(3), acl_get_file(3), acl_set_fd(3), acl_set_file(3), acl_size(3), acl_valid(3). Access control lists can be manipulated using the getfacl(1) and setfacl(1) commands. See the associated man pages for details.
User quotas and group quotas are now functional.
The Linux version of dump and restore can be used to migrate data from a non-XFS volume to an XFS volume. Linux dump, however, cannot be used to dump an XFS volume as it uses ext2 library functions.
xfsdump and xfsrestore can be used to migrate an IRIX XFS volume to a Linux XFS volume and vice versa.
xfsdump and xfsrestore should be used for backup and restore of XFS volumes.
xfsdump and restore may be used to move DMAPI-managed filesystems between IRIX and Linux. The HSM that uses that DMAPI filesystem must also support the moving of the filesystem between big- and little-endian machines.
The LVM volume manager is not currently supported for IA-64 architecture. We recommend using md for IA-64.
IRIX XFS supports two directory formats, referred to as version 1 and version 2. Linux XFS supports version 2 only. XFS filesystems that need to be migrated from IRIX to Linux systems must use the version 2 directory structure.
The mkfs command creates version 2 directory format filesystems by default in IRIX release 6.5.14 and later. You can determine what directory format version exists on an IRIX filesystem with the following command:
xfs_growfs -n mount-point | grep naming
For example:
# xfs_growfs -n / | grep naming naming =version 1 bsize=4096
For the current release of XFS, the filesystem block size is limited to the size of a memory page. On a x86 architecture that size is 4 Kbytes. On IA-64 architecture, the page size is set at compile time; the page size must match the filesystem block size of any XFS filesytem you wish to mount. You may need to recompile the kernel with 8k or 4k filesystem block sizes if you are migrating an IRIX filesystem.
Note that files systems created on an IRIX/MIPS platform must have been created with a 4 Kbyte block size in order to be mounted on a ia32 Linux system. File systems not created with a 4k block will fail to mount with an error indicating the mis-match.
The unwritten extent feature is not yet operational. Currently this is the default setting for mkfs.xfs on Linux.
Realtime volumes are not yet tested in XFS Linux and should not be used.
It is possible to remount an XFS filesystem to change its permissions from read/write to read-only. This support is needed for XFS to function as a root filesystem.
XFS 1.1 supports MD RAID0, RAID1 and RAID5. RAID5 with an internal log may perform slightly worse with XFS than with ext2. This will be addressed in a future release.
In general we do not foresee the need to run fsr on system partitions such as /, /boot and /usr as these will in general not suffer from fragmentation. There are also issues with defragmenting the files lilo uses to boot your system. Should these files be moved by fsr then you must rerun lilo before you reboot or you may have an unbootable system.