[Gluster-infra] LVM allocation settings

[Jeff Darcy]

Often, when a regression test seems to be running much slower than
usual, I look in the logs and see this:

> Internal error: Reserved memory (14749696) not enough: used 19890176.
> Increase activation/reserved_memory?

It's pretty easy to fix this manually, by changing a value in lvm.conf,
and I recommend doing so, but what's the best way to automate it?