The ZFS Filesystem represents the backbone of most Qlustar storage solutions. Its data-protection features are the best Linux has to offer. Qlustar ZFS is based on the ZFS on Linux port. While the latter project is managed outside of the official Linux kernel tree our implementation is fully integrated with the Qlustar kernel.
No recompiling and checking for updates: ZFS is available out-of-the-box on any Qlustar cluster node and comes with full support and updates. ZFS configurations are easily defined and assigned to nodes using the QluMan GUI. With Qlustar, setting up reliable and performant storage nodes doesn’t require rocket scientists anymore.
ZFS has unique data protection features and Qlustar is ready to take full advantage of it:
- It supports self-healing by using meta-data and checksums to detect silent and noisy data corruption. ZFS checksums data every time it is read from disk and is able to correct errors on the fly.
- ZFS pools have intrinsic support for all RAID levels to be desired. This includes striping, mirroring, parity based RAIDz1/2/3 and combinations of these.
- ZFS employs Copy On Write. This means that blocks holding active data are never overwritten in place. Instead, a new block is always allocated to write modified data to it. Finally, any meta-data blocks referencing the original block are similarly read, reallocated, and written. As a result, your data is always in a consistent state.
Performance and Scalability
- ZFS uses efficient inline compression to save storage space and increase write speed.
- Files in a ZFS filesystem can be arbitrarily large.
- ZFS filesystems are 128-bit, allowing for 256 quadrillion zettabytes of storage. All meta-data is allocated dynamically, hence no limit is put on the number of files. Compare this with the fixed limits of other filesystems that require pre-allocation of inodes during their creation. Directories can have up to 256 trillion entries, and no limit exists on the number of filesystems in a pool.
Special Qlustar ZFS Features
QluMan’s powerful disk management module allows to define ZFS configurations in a simple to use syntax and assign them to a class of nodes. A node’s assigned ZFS configuration is then employed during its boot process to automatically create or activate (depending on a configuration setting) the corresponding ZFS pools and filesystems.
If a hardware setup is available that supports it, Qlustar ZFS can be run in high-availability (HA) mode. This is achieved by specially developed pacemaker resource types usable within a Qlustar corosync/pacemaker setup. Contact us, if you need assistance in designing, setting up or supporting proven ZFS (HA) configurations.
ZFS is a registered trademark of Oracle America, Inc.