[CentOS] CentOS 4.4 lvm and drbd 0.8?

Sun Mar 4 18:29:34 UTC 2007
Ross S. W. Walker <rwalker at medallion.com>

> -----Original Message-----
> From: centos-bounces at centos.org 
> [mailto:centos-bounces at centos.org] On Behalf Of Les Mikesell
> Sent: Sunday, March 04, 2007 12:51 PM
> To: CentOS mailing list
> Subject: Re: [CentOS] CentOS 4.4 lvm and drbd 0.8?
> Ross S. W. Walker wrote:
> > 
> > If you were running a later kernel version of MD, it is conceivable
> > that you could create a mirror with a remote storage drive over
> > iscsi.
> > 
> > It would be up to you though to figure out how to fail-over to it
> > and to limit the bandwidth MD takes to that remote mirror and
> > releasize that it will always be fully synchronous and so
> > performance may not be the best over a WAN.
> > 
> > You can also use a pair of vise grip plyers to do the job of an
> > adjustable wrench, but it will probably strip the bolt in the
> > process.
> Unix has always been about combining tools that each do one 
> job well. If 
> we already have a tool (iscsi) that exports remote block 
> devices well, 
> following standards that would the actual storage to be on non-linux 
> devices, and another tool (md raid) that mirrors block 
> devices, why not 
> combine them instead of inventing yet another special purpose 
> tool?  I 
> realize that drbd and nbd were developed before iscsi, but now that 
> there is a standard cross-platform network block device, why 
> shouldn't 
> it be used?  MD might need some new options to make it work as 
> efficiently in this scenario, but that seems like a more 
> useful place to 
>   add features - that is, there might be other situations where MD 
> mirroring to an external iscsi partition would be useful, or even 
> combining many iscsi exports into one raid volume.

While yes combining tools is the Unix way, there are also some
tools that are better suited for a task than others.

For example in my second post I suggest if you have a bunch of
direct attached storage, say located on 20 servers in a local
area network. If you export the storage from those 20 servers
to a central server via iSCSI and use MD raid to create a large
RAID-{3,4,5,6} array of those iSCSI targets. Then you use LVM
to split those into differing volumes for re-export via iSCSI.
You can then use drbd on those LVM volumes with asynchronous
replication to a storage array off-site for DR purposes or to
a local storage device as a volume snapshot server (where both
copies need to be active at once). You can then re-export those
volumes via iSCSI to different OS platforms to use as storage.

This could work well, but if you plan on using MD RAID1 to
replicate data to off-site storage you will see very poor
performance. MD RAID1 may work well in replicating storage
synchronously between 2 local storage devices, but not to a
remote storage device. In that case drbd would work better in
that case. Also if you wanted both sides of a mirror to be
active at once drbd 8.X is the only way to go.


This e-mail, and any attachments thereto, is intended only for use by
the addressee(s) named herein and may contain legally privileged
and/or confidential information. If you are not the intended recipient
of this e-mail, you are hereby notified that any dissemination,
distribution or copying of this e-mail, and any attachments thereto,
is strictly prohibited. If you have received this e-mail in error,
please immediately notify the sender and permanently delete the
original and any copy or printout thereof.