Setting up a cluster filesystem: Difference between revisions
(Add section about inconsistent device numbers) |
(→Limitations: Add section on lock coherence) |
||
(4 intermediate revisions by the same user not shown) | |||
Line 12: | Line 12: | ||
= Cluster file systems = |
= Cluster file systems = |
||
== Components == |
|||
Any cluster file system will have some or all of following components: |
Any cluster file system will have some or all of following components: |
||
Line 22: | Line 22: | ||
* User space tools for management |
* User space tools for management |
||
== Limitations == |
|||
Every clustered file system has its quirks and limitations. Some of the file system limitations will affect the configuration of file services (Samba or NFS). |
Every clustered file system has its quirks and limitations. Some of the file system limitations will affect the configuration of file services (Samba or NFS). |
||
Line 30: | Line 30: | ||
* Does file system have specific quorum requirements? |
* Does file system have specific quorum requirements? |
||
=== Checking uniformity of device and inode numbering === |
|||
⚫ | |||
File services (e.g. Samba or NFS) often generate file identifiers or handles from device and inode numbers. These services may not work correctly if these numbers are not uniform across nodes. |
|||
This can be tested using the stat(1) command as follows: |
|||
⚫ | |||
⚫ | |||
⚫ | |||
41:35820037 |
|||
⚫ | |||
⚫ | |||
41:35820037 |
|||
⚫ | |||
⚫ | |||
38:35820037 |
|||
Note that the device numbers are not consistent across nodes. File services sometimes provide a way of working around this (e.g. [[Configuring_clustered_Samba#Filesystem_specific_configuration|Samba]]). |
|||
Some cluster filesystems (especially some [https://en.wikipedia.org/wiki/Filesystem_in_Userspace FUSE]-based ones) do not provide consistent inode numbers across nodes. There is often no workaround for this. |
|||
=== Checking lock coherence === |
|||
Clustered Samba has a couple of dependencies on the cluster filesystem: |
|||
* If using [[Configuring the CTDB recovery lock|CTDB recovery lock]] then lock coherence of the cluster filesystem needs to be confirmed |
|||
* Samba, with POSIX locking enabled, requires I/O coherence |
|||
Both of these can be checked using the [[Ping pong|ping_pong]] tool. |
|||
⚫ | |||
Each clustered file system example will describe how to set up a clustered file system for 3 node cluster. The implementation can be scaled down to 2 nodes or scaled up to more nodes. |
Each clustered file system example will describe how to set up a clustered file system for 3 node cluster. The implementation can be scaled down to 2 nodes or scaled up to more nodes. |
||
Line 61: | Line 92: | ||
[https://oss.oracle.com/projects/ocfs2/ OCFS2] is a general-purpose shared-disk cluster file system for Linux capable of providing both high performance and high availability. |
[https://oss.oracle.com/projects/ocfs2/ OCFS2] is a general-purpose shared-disk cluster file system for Linux capable of providing both high performance and high availability. |
||
= Other cluster filesystems = |
|||
If you can't find documentation about your choice of cluster filesystem and clustered Samba then you might need to work around some limitations. |
|||
== Inconsistent device numbers == |
|||
''Note: This section probably wants to be in a future page about cluster filesystems and Samba configuration. It can be moved later...'' |
|||
Locking will not work if a cluster filesystem does not provide unique device numbers across nodes. |
|||
Consider the following example: |
|||
⚫ | |||
⚫ | |||
⚫ | |||
File: `/clusterfs/testfile' |
|||
Size: 1286700 Blocks: 2514 IO Block: 65536 regular file |
|||
Device: '''29h/41d''' Inode: 35820037 Links: 1 |
|||
Access: (0774/-rwxrwxr--) Uid: ( 3535/ foo) Gid: ( 513/Domain Users) |
|||
Access: 2016-11-03 19:51:46.000000000 +0000 |
|||
Modify: 2016-11-01 13:06:04.000000000 +0000 |
|||
Change: 2016-11-01 13:06:04.000000000 +0000 |
|||
⚫ | |||
⚫ | |||
File: `/clusterfs/testfile' |
|||
Size: 1286700 Blocks: 2514 IO Block: 65536 regular file |
|||
Device: '''29h/41d''' Inode: 35820037 Links: 1 |
|||
Access: (0774/-rwxrwxr--) Uid: ( 3535/ foo) Gid: ( 513/Domain Users) |
|||
Access: 2016-11-03 19:51:46.000000000 +0000 |
|||
Modify: 2016-11-01 13:06:04.000000000 +0000 |
|||
Change: 2016-11-01 13:06:04.000000000 +0000 |
|||
⚫ | |||
⚫ | |||
File: `/clusterfs/testfile' |
|||
Size: 1286700 Blocks: 2514 IO Block: 65536 regular file |
|||
Device: '''26h/38d''' Inode: 35820037 Links: 1 |
|||
Access: (0774/-rwxrwxr--) Uid: ( 3535/ foo) Gid: ( 513/Domain Users) |
|||
Access: 2016-11-03 19:51:46.000000000 +0000 |
|||
Modify: 2016-11-01 13:06:04.000000000 +0000 |
|||
Change: 2016-11-01 13:06:04.000000000 +0000 |
|||
Note that the device numbers are not consistent across nodes. Locks set for the file on the first 2 nodes will not affect the 3rd node. |
|||
To work around this, the following settings should be in the global section of the Samba configuration: |
|||
vfs objects = fileid |
|||
fileid:mapping = fsname |
Latest revision as of 22:43, 5 August 2020
Goal
Set up a clustered file system to be used with CTDB for providing clustered file services.
In addition,
- How to test if posix locking is supported on the file system?
- Limitations when using clustered file system
Setting up clustered file system has nothing to do with CTDB. This information is provided for completeness. Users should be aware of any limitations of particular clustered file system.
Cluster file systems
Components
Any cluster file system will have some or all of following components:
- Shared or distributed storage
- Kernel or user space file system driver
- User space file system daemon(s)
- User space distributed lock manager
- User space tools for management
Limitations
Every clustered file system has its quirks and limitations. Some of the file system limitations will affect the configuration of file services (Samba or NFS).
- Does file system provide a consistent view across all the nodes (for example - uniform device and inode numbering) ?
- Does file system provide posix locking semantics (cluster-aware locking)?
- Does file system have specific quorum requirements?
Checking uniformity of device and inode numbering
File services (e.g. Samba or NFS) often generate file identifiers or handles from device and inode numbers. These services may not work correctly if these numbers are not uniform across nodes.
This can be tested using the stat(1) command as follows:
# onnode all stat -c '%d:%i' /clusterfs/testfile >> NODE: 10.1.1.1 << 41:35820037 >> NODE: 10.1.1.2 << 41:35820037 >> NODE: 10.1.1.3 << 38:35820037
Note that the device numbers are not consistent across nodes. File services sometimes provide a way of working around this (e.g. Samba).
Some cluster filesystems (especially some FUSE-based ones) do not provide consistent inode numbers across nodes. There is often no workaround for this.
Checking lock coherence
Clustered Samba has a couple of dependencies on the cluster filesystem:
- If using CTDB recovery lock then lock coherence of the cluster filesystem needs to be confirmed
- Samba, with POSIX locking enabled, requires I/O coherence
Both of these can be checked using the ping_pong tool.
Implementation
Each clustered file system example will describe how to set up a clustered file system for 3 node cluster. The implementation can be scaled down to 2 nodes or scaled up to more nodes.
GPFS
GPFS is a proprietary cluster file system from IBM.
GFS2
GFS2 is a clustered file system supported by Red Hat.
Lustre
Lustre file system is an open-source, parallel file system that supports many requirements of leadership class HPC simulation environments.
GlusterFS
GlusterFS is a scalable network file system.
OCFS2
OCFS2 is a general-purpose shared-disk cluster file system for Linux capable of providing both high performance and high availability.