[Linux-cluster] Redundant (multipath)/Replicating NFS

[Gordan Bobic]

isplist at logicore.net wrote:
>> If you have 2 NFS server you need replicated and failed over, you can
>> use DRBD to replicate a virtual block device. You can have this in
> 
> I'm looking on the net about this and it looks interesting. I see that it is 
> available as RPM's but I don't see it as being part of RHEL?

It's not a part of RHEL, but if you add yum repositories for rpmforge 
and atrpms, the packages for RHEL5/CentOS5 are available.

> Is this something 
> that will run on one or more servers? I can't actually run anything on the NFS 
> servers since they are embedded OS's on hardware. I;m using NAS servers 
> offering up NFS shares from pools of RAID FC storage.

In that case you're out of luck. :-(

> I'd like to use something like this for the web data that the web servers need 
> access to. Your aggregator suggestion is what I've decided upon for the larger 
> ongoing needs of media storage. 
> 
> Static web stuff as above is relatively easy to deal with using the right 
> tools but I'm trying to figure out how to deal with large amounts of media 
> data. Obviously, I don't want users to lose their pictures, videos, files, 
> work, etc. 

If your NFS servers are proprietary hardware, I doubt you'll find 
anything to handle replication between the two of them, unless the 
appliance vendor has a module available for such a thing.

> What are your thoughts on even bothering to replicate when using RAID storage 
> devices to begin with? They are highly reliable as it is, using 12 drives, 
> hard to imagine a loss off data.

So you have 2 FC chasis with 12 drives each? I doubt mirroring them is 
vital provided they are running RAID5 or RAID6 on each. But if a chasis 
fails, you'll still have downtime until you can replace it, even if you 
can transplant all the drives into a new chasis. It depends on how much 
downtime you can live with, and on whether you need a system with 
absolutely no single points of failure.

> I have huge amounts of tape storage using robot libraries with multiple drives 
> but still, that doesn't seem like the way to deal with the media. Setting up 
> fully redundant storage seems to be a doubling of costs no matter how I cut it 
> so I am at a loss on how to deal with media. 

I don't think near-line storage is going to be workable for on-line 
media serving. I guess you have to decide how much redundancy and 
headroom you need. Based on this you might decide to run RAID 01 or RAID 
51 on the setup. But if you need mirroring for the NFS proxies, I 
suspect that's only going to be doable if the appliances support it.

Gordan