[ale] Fileserver/filesystem replication

Bob Toxen bob at verysecurelinux.com
Wed Dec 31 04:18:46 EST 2003


Ya want real-time replication.  I'll give ya real-time replication.

1. Build your ext3 or ReiserFS on top of Linux software mirroring.
2. Make the second (mirror) disk a network disk device.
3. Ensure that writes are not acknowledged until they percolate through
   to the remote real hard disk.

Done.

Not only do you get true real-time remote mirroring BUT with any decent
DB your transactions will be consistent in the event of failover.

Bob Toxen
bob at verysecurelinux.com               [Please use for email to me]
http://www.verysecurelinux.com        [Network&Linux/Unix security consulting]
http://www.realworldlinuxsecurity.com [My book:"Real World Linux Security 2/e"]
Quality Linux & UNIX security and SysAdmin & software consulting since 1990.

"Microsoft: Unsafe at any clock speed!"
   -- Bob Toxen 10/03/2002

> On Tue, Dec 30, 2003 at 10:06:20AM -0500, Matt Smith wrote:
> Does anyone have experience with any filesystem replication methods
> other than rsync alone?  I'm looking for something a little more real-time
> (preferably not scheduled) to do at a minimum a one-way replication to a
> "backup" server, and ideally, a two-way replication so that two servers
> could be used in parallel.

> I realize the former could be accomplished with a frequently scheduled
> rsync, but I'm hoping for something a little more robust than that.
> I can tolerate a one-way copy to the backup server as long as it's done
> very quickly after a file is modified, because if I ever have to fail
> over to the backup server for reads, it probably means the primary is down
> hard so I won't have to worry about copying updates to it any time soon.
> I'm handling the failover via a NFS connection through a F5 Load balancer
> that will point the connections to the primary server unless it's down...

> I've looked into coda, several clustering filesystems: pvfs, and a few
> others (some commercial, but I've got no budget for this, of course),
> but they all seem to either lack the nonscheduled replication I'm looking
> for, or are more focused on parallel access than replication for fault
> tolerance.

> Thanks in advance for any feedback on your experiences.

> --Matt



More information about the Ale mailing list