Can you please correct any miss understanding I may have for ZFS and GlusterFS?
Two identical servers with twi drives partitioned with ZFS. We use ZFS because ZFS is a good substitute for LVM, it also incorporates a form if its own version of RAID.
But we want to take it a step further and user GlusterFS. We want to do this so we can offer the ZFS drives over the network and but also because we want replication of data in case of a major failure. This way if Server1 is completely destroyed, we can default to Server2 and still have access to the same data stored in the ZFS volumes.
And here are my questions, conclusions.
ZFS offers a huge degree of scalability. By adding glusterfs we ensure redundant drives available across the network.
GlusterFS is NOT a filesystem (like NFS, ext4 etc) are. It is technology that offers redundancy.
GlusterFS is not to be confused with NFS, even though they are similarities. NFS is a filesystem that is available across the network (similar as a Samba share). It is that and nothing else. On the other hand GlusterFS offers whatever filesystem you want to share (can be ext4, ext3,btrfs) across the network and it also allows for a huge degree of redundancy.
Server1 Server2 ------- ------- | ZFS | | ZFS | ===== ====== | sda | | sda | | sdb | | sdb | ------ -------- gluster1 <--PEERS--> gluster2 | | V Client
[link] [comments]