[DRBD-user] DRBD + NFS

Mustafa A. Hashmi mahashmi at gmail.com
Sat Oct 14 13:37:55 CEST 2006

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.

On 10/10/06, Dan Brown <danb at zu.com> wrote:
> So I've got a couple of webservers which mirror each other without the use
> of a fileserver backend.  Up until now they've been using csync2 to mirror
> the data between the servers and it has been working fairly well (and the
> built-in file backup has come in handy once or twice as well).
> Unfortunately csync2 has a good deal of limitations as well, such as the
> more files you have, the slower it is, and you can't sync more than one
> server to other servers (in this case only one) at the same time.  I'm
> running it in a master/master mode.  Overall it was designed to be more of a
> tool to update configs, or update relatively non-interactive systems.
> So recently I've been looking more and more at DRBD and NFS (before I spend
> a couple of grand on a well built fileserver).  So what I am looking at is
> something like this:
> server A (DRBD + active NFS)
>   + mounted by itself and server B
> server B (DRBD + disabled NFS)
>   + mounts server A NFS share
> So while server A is the active NFS server and server B is passive and only
> synching, both servers would mount via NFS server A.  When server A goes
> down for whatever reason (reboot, crash, etc) server B will take over and
> continue where server A left off.   One thing I keep seeing in various
> documents however is a warning not to have the server mount its own NFS
> shares.  There is never reason (even a vague reason) given why not to
> however.  I can see the obvious reasons (eg. infinitely nested filesystems
> via symlinks/mounts, crossmounts, etc), but my directory structures should
> not need anything like this at all.   Other than overall system complexity,
> I don't see any reasons to not be able to self mount NFS with a whole lot of
> trouble.

Odd -- the NFS-HA howto shows how to do exactly this. Also: we've
deployed this for mail and web without issues this far, albeit, in a
non-complex environment.

> The problems I see include Apache log files (I'd assume each server would
> have to write to separate logs),


> and what happens when the servers
> disconnect from each other (eg. NIC failure) but are both still active and
> both decide they are masters.  How are synchronization issues then resolved?

You generally run a serial cable between the nodes to prevent a split
brain and not rely entirely on network.

Mustafa A. Hashmi
mahashmi at gmail.com
mh at stderr.net

More information about the drbd-user mailing list