[lug] LVM and disk failure

[Nate Duehr]

David L. Anselmi wrote:

[snipping around to make a point...]

> In the Linux world you don't want /tmp to use up all your / space and
> crash.  You don't want anything in /home executed (ever, really).  So
> you make different partitions for different purposes.  But then
> resizing, without LVM, on a single device really sucks.
> 
> So first decide how to arrange your drives for performance and
> reliability.  That will depend on number of drives, read vs. write load,
> etc.  And don't forget to consider your controllers which may give you
> extra data paths to use or single points of failure.  (I've seen lots of
> data lost off RAIDs.  I've never seen a drive fail in a RAID.  Granted
> my experience is a corner case.)

I like that!  I've never met anyone who was a corner case.  :-)

I've definitely seen disks fail in a RAID!  It's so nice to just call
the guy halfway across the country who works for some low-end
tech-on-call company and have him drive to the site, slap a new disk in,
and monitor the resync for an hour, than the alternative!

> No you wouldn't.  1000 disks is a SAN and you have loads of redundancy
> built in there.  Or you got ripped off.  Or you built it yourself
> instead of buying the SAN you should have.

Yeah, this is the reality of storage these days... once you get beyond a
certain size, it *is* cost effective to buy SAN devices or
network-attached storage devices and let them handle the disks.

I like to call LVM "poor-mans Veritas" sometimes, too.  If you've played
around with Veritas on Sun, you get frustrated with the manual building
blocks of LVM, software or hardware RAID, etc.  It manages ALL of this
stuff, and allows a ton of other things like snapshots, in a very
portable fashion -- at least on Sun boxes.

Supposedly later versions of Veritas if set up correctly also allow
deporting and importing of the same filesystem not only across different
servers in a farm, but also across OS's.  It's wicked neat stuff, but
you have to pay attention to what you're doing...

Never had a project where I needed Veritas on Linux, but I hear they do
it, these days...

Veritas + JBOD's or dumb cheap SAN's makes a very powerful system, but
not cheap.  Everything it can do can be done on Linux these days, but at
a much higher load on the admin, to set it all up.

> That's an interesting idea.  You should research whether that's easier
> than using quotas (the traditional approach) and what the trade offs are
> and then write a paper.  The LUGs here would be interested in it, and
> maybe even USENIX et. al.

Yeah, I like his idea too.  That's pretty nifty -- resize the individual
user's disk space as needed.  Smart.

Nate