Best practice for high availability ZFS pool

[Palle Girgensohn]

> 16 maj 2016 kl. 16:52 skrev Rainer Duffner <rainer at ultra-secure.de>:
> 
> 
>> Am 16.05.2016 um 12:08 schrieb Palle Girgensohn <girgen at FreeBSD.org>:
>> 
>> Hi,
>> 
>> We need to set up a ZFS pool with redundance. The main goal is high availability - uptime.
>> 
>> I can see a few of paths to follow.
>> 
>> 1. HAST + ZFS
>> 
>> 2. Some sort of shared storage, two machines sharing a JBOD box.
>> 
>> 3. ZFS replication (zfs snapshot + zfs send | ssh | zfs receive)
>> 
>> 4. using something else than ZFS, even a different OS if required.
> 
> 
> 
> There’s always GlusterFS.
> Recently ported to FreeBSD and available as net/gulsterfs (10.3 recommended, AFAIK).
> 
> At work, we use it on Ubuntu - but not with so much data.
> On Linux, I’d use it on top of XFS.
> 
> For our Cloud-Storage, we went with ScaleIO (which is Linux only).
> 
> You need more than two nodes with Gluster, though (for production use)
> I think my co-worker said four at least.

Yeah, it is interesting, but as you say, you really create a RAID5 setup at least.

> 
> If you have the money and don’t mind Linux, ScaleIO is probably the best you can buy at the moment.
> While licensed at the GByte-Level (yeah, EMC…) it can be used free of charge, unsupported.

Yeah that is definitely an option.

We already have an infrastructure based on ZFS, and I am not sure I do trust ZFS on Linux?

Palle

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 495 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://lists.freebsd.org/pipermail/freebsd-fs/attachments/20160517/b642d340/attachment.sig>