I believe the deadlock described in the article is nearly impossible when VM disk caching
is disabled (by default).
Anyway, there is a solution to avoid loopback NFS:
https://gerrit.ovirt.org/#/c/68822/10
On 30/05/2017, 17:55, "users-bounces(a)ovirt.org on behalf of Derek Atkins"
<users-bounces(a)ovirt.org on behalf of derek(a)ihtfp.com> wrote:
Hi,
Simone Tiraboschi <stirabos(a)redhat.com> writes:
5. When I install Hosted Engine from the Node web interface, I
select NFS4
for the storage and provide the correct host:/path. Other than that, and
setting up static network details and email details, I follow the default
prompts of the Hosted Engine setup.
AFAIK loopback mount an NFS share could lead to deadlocks:
https://lwn.net/Articles/595652/
This article is from 2014. Are you saying that this issue hasn't been
fixed in the intervening three years?
FWIW I'm running with Loopback NFS on my single-host Ovirt 4.0.6 system
and it's been running happily for a while now. I would have happily
used a local filesystem, but the hosted-engine code wouldn't let me.
Setting up loopback NFS was easier than trying to learn GlusterFS and
figure out how to loopback Gluster.
On the other hand, my system does have 256GB of RAM, with only about 50%
being used, so that also might work in my favor.
-derek
--
Derek Atkins 617-623-3745
derek(a)ihtfp.com
www.ihtfp.com
Computer and Internet Security Consultant
_______________________________________________
Users mailing list
Users(a)ovirt.org
http://lists.ovirt.org/mailman/listinfo/users