On 8/14/13 11:37 AM, Brian Candler wrote:
On 14/08/2013 15:58, Christian Völker wrote:
Hi all,
I'm planning to use DRBD in a production environment. I prefer to use
CentOS as base system.
The reason to use drbd is the synchronisation, not the high
availability.
We'll have two locations connected through a 100Mbit line. On both
locations users will access the data at the same time. So I know I have
to use a cluster aware filesystem.
You mean, you are planning to use DRBD in dual-master mode??
Why not just have all the accesses go to one master node in one data
centre? Then you can use a regular filesystem, with something like NFS
on top, and you can still replicate to a slave.
This would be the best option, or a chunk fs might be better, still
having a single master entry point is the key.
Perhaps you expecting a very read-heavy workload, and you want all
reads to take place from the nearer copy? Using DRBD + dual master +
GFS is a risky way of achieving this, and as the GFS nodes will have
to communicate to grab locks anyway, it may not have much better
latency than the simple case.
I agree with this, I think you will have issues with locking. DLM,
proto C and OCFS2 in the same data center is questionable with multiple
writers.
Which filesystem is recommended? GFS? ZFS (experimental?)?
Apples and oranges. GFS and OCFS2 are cluster-aware filesystems which
you may be able to use in a multi-master scenario - if you are braver
than me. ZFS is a standard filesystem which you could only use in a
single-master setup. The Linux ZFS port is also a somewhat risky
option; if you want to use ZFS I'd suggest FreeBSD (or a FreeBSD-based
appliance like FreeNAS)
Agreed.
Regards,
Brian.
_______________________________________________
drbd-user mailing list
[email protected]
http://lists.linbit.com/mailman/listinfo/drbd-user
_______________________________________________
drbd-user mailing list
[email protected]
http://lists.linbit.com/mailman/listinfo/drbd-user
