The Global File System is a Linux-based file system that enables the sharing of block files among various users.
GFS, also known as Global File System, is a highly efficient cluster file system. By allowing a cluster of computers to use a block device that is shared between them, it enables simultaneous reading and writing to the block device. GFS operates like a local filesystem, but with the added feature of a lock module that coordinates I/O, ensuring filesystem consistency. One of the standout qualities of GFS is its perfect consistency – changes made to the filesystem on one machine are instantly visible on all other machines in the cluster.
GFS is composed of a set of kernel patches and userspace programs. Depending on the lock module used, different dependencies will be required. For instance, GFS lock module lock_dlm relies on CMAN and DLM, lock_gulm relies on GULM, while lock_nolock requires no other dependencies. Additionally, some GFS tools depend on the iddev library.
Building and installing GFS is a straightforward process. For building from a cluster tree, navigate to the cluster root directory and issue the following commands: './configure --kernel_src=/path/to/kernel', 'make', and 'make install'. Kernel modules can also be built by applying the relevant kernel patches, which are included in the cman-kernel/patches, dlm-kernel/patches, and gfs-kernel/patches directories. After building GFS, users must install the device mapper user space with 'cd device-mapper', './configure', 'make', and 'make install', and the LVM2/clvm with './configure --with-clvmd --with-cluster=shared', 'make', and 'make install LVM2/scripts/clvmd_fix_conf.sh /usr/lib'.
Once GFS is built and ready to use, the relevant kernel modules must be loaded. This can be done by running 'depmod -a', followed by modprobing 'dm-mod', 'gfs', 'lock_dlm', 'dlm', 'cman', 'lock_harness', and device-mapper (if built as a module).
To start using GFS, users must follow the designated startup procedure. For each cluster node, the following commands must be run: 'ccsd' (starts the CCS daemon), 'cman_tool join' (joins the cluster), 'fence_tool join' (joins the fence domain), 'clvmd' (starts the CLVM daemon), 'vgchange -aly' (activates LVM volumes locally), and 'mount -t gfs /dev/vg/lvol /mnt' (mounts a GFS file system).
To safely shut down GFS, users must follow the designated shutdown procedure. For each cluster node, the following commands must be run: 'umount /mnt' (unmounts the GFS file system), 'vgchange -aln' (deactivates LVM volumes locally), 'killall clvmd' (stops the CLVM daemon), 'fence_tool leave' (leaves the fence domain/stops fenced), 'cman_tool leave' (leaves the cluster), and 'killall ccsd' (stops the CCS daemon).
Overall, GFS is a reliable and highly efficient cluster file system that offers seamless coordination between multiple machines. With its easy installation, straightforward startup and shutdown procedures, and consistent performance, it's an excellent choice for anyone looking for a robust and scalable file system
GFS is composed of a set of kernel patches and userspace programs. Depending on the lock module used, different dependencies will be required. For instance, GFS lock module lock_dlm relies on CMAN and DLM, lock_gulm relies on GULM, while lock_nolock requires no other dependencies. Additionally, some GFS tools depend on the iddev library.
Building and installing GFS is a straightforward process. For building from a cluster tree, navigate to the cluster root directory and issue the following commands: './configure --kernel_src=/path/to/kernel', 'make', and 'make install'. Kernel modules can also be built by applying the relevant kernel patches, which are included in the cman-kernel/patches, dlm-kernel/patches, and gfs-kernel/patches directories. After building GFS, users must install the device mapper user space with 'cd device-mapper', './configure', 'make', and 'make install', and the LVM2/clvm with './configure --with-clvmd --with-cluster=shared', 'make', and 'make install LVM2/scripts/clvmd_fix_conf.sh /usr/lib'.
Once GFS is built and ready to use, the relevant kernel modules must be loaded. This can be done by running 'depmod -a', followed by modprobing 'dm-mod', 'gfs', 'lock_dlm', 'dlm', 'cman', 'lock_harness', and device-mapper (if built as a module).
To start using GFS, users must follow the designated startup procedure. For each cluster node, the following commands must be run: 'ccsd' (starts the CCS daemon), 'cman_tool join' (joins the cluster), 'fence_tool join' (joins the fence domain), 'clvmd' (starts the CLVM daemon), 'vgchange -aly' (activates LVM volumes locally), and 'mount -t gfs /dev/vg/lvol /mnt' (mounts a GFS file system).
To safely shut down GFS, users must follow the designated shutdown procedure. For each cluster node, the following commands must be run: 'umount /mnt' (unmounts the GFS file system), 'vgchange -aln' (deactivates LVM volumes locally), 'killall clvmd' (stops the CLVM daemon), 'fence_tool leave' (leaves the fence domain/stops fenced), 'cman_tool leave' (leaves the cluster), and 'killall ccsd' (stops the CCS daemon).
Overall, GFS is a reliable and highly efficient cluster file system that offers seamless coordination between multiple machines. With its easy installation, straightforward startup and shutdown procedures, and consistent performance, it's an excellent choice for anyone looking for a robust and scalable file system
What's New
Version 6.1pre21: N/A