Don't override disk_cache settings for RBD disks

What software version are you running? Please provide the output of "gnt-
cluster --version", "gnt-cluster version", and "hspace --version".

gnt-cluster (ganeti v2.10.1) 2.10.1

Software version: 2.10.1
Internode protocol: 2100000
Configuration format: 2100000
OS api version: 20
Export interface: 0
VCS version: (ganeti) version v2.10.1

hspace (ganeti) version v2.10.1
compiled with ghc 7.6
running on linux x86_64

What distribution are you using?

Debian Jessie (testing)

What steps will reproduce the problem?
1. disk_cache settings other than none will be overridden by ganeti despite RBD 
being completely safe and well behaved when it comes to live migration and 
other cache sensitive issues.

2. See https://groups.google.com/forum/#!topic/ganeti/loF_Gp6hHew and 
especially the reference to 
https://www.mail-archive.com/ceph-users@lists.ceph.com/msg06890.html in that 
thread.

What is the expected output? What do you see instead?

ganeti should treat RBD like DRBD when it comes to cache safety. 

Please provide any additional information below.

Since the check that does the override lumps all externally mirrored disks into 
the same unsafe pot (and some of those probably are unsafe with caching) I 
don't know what the best way to solve this from a programming PoV is.
If making an exception for RBD is too involved maybe changing the override to a 
stern warning "you better know what you're doing" might be the faster way 
forward. 
Fixing this ASAP would vastly improve performance with RBD, especially with RBD 
Cache in the userspace model.

Original issue reported on code.google.com by chibi0...@gmail.com on 20 Mar 2014 at 8:22

olopez32 / ganeti

Don't override disk_cache settings for RBD disks #766