I would like to use glusterfs to:
- provide high availability,
- provide one big filesystem to the clients,
- easily add more storage.
So, something like RAID-10 (only that RAID-10 will not easily allow to
add more disks...).
High availability means AFR - I would like to keep a copy of each file
on two servers:
(10 TB) s1 -AFR- s2\
(10 TB) s3 -AFR- s4---- big filesystem --- clients
(10 TB) s5 -AFR- s6/
I would like to start with only two servers (s1 and s2) making 10 TB
storage.
Later, I will add s3 and s4, increasing the storage to 20 TB (visible as
one filesystem to the clients). And so on.
Can I achieve something like this with glusterfs?
I checked the docs briefly, but there, I only found AFR example where a
copy is stored on each and every server, which does not use the space
very efficiently in my usage scenario.
Is my above example possible with glusterfs? I'm new to glusterfs, so
it's possible I didn't see a useful example in the docs - I would be
glad if someone could provide me a pointer.
--
Tomasz Chmielewski
http://wpkg.org