Ok well thanks for the help. Moving it by hand is not a huge deal. I just assumed that since it was recommended to use the command in the docs that it was worth mentioning. case closed I guess. unless we want to consider the NFS error worthy of a bug report? On Dec 31, 2014 3:34 PM, "Yedidyah Bar David" <didi@redhat.com> wrote:
From: "Steve Atkinson" <satkinson@telvue.com> To: "Yedidyah Bar David" <didi@redhat.com> Cc: users@ovirt.org Sent: Wednesday, December 31, 2014 8:22:14 PM Subject: Re: [ovirt-users] engine-iso-uploader unexpected behaviour (steve a) If I attempt to mount any of them via NFS from our management network
----- Original Message ----- they
work just fine. (moved around, read/write operations) Copied the ISO I needed directly to it and changed the perms/ownership by hand which seems to have worked as a short term solution.
Good. Not sure it's that short term - it was suggested several times here and people seem to use it.
I can see why the --iso-domain argument has issues as it is trying to use the our storage network, which isn't routable from the Engine as it only has the one network interface. Although that does seem like an oversight. Seems like this transfer should pass through the SPM and not try to directly mount the NFS share if the --iso-domain flag is used.
Indeed this is a somewhat complex subject, which is why it was somewhat neglected so far. There is some work to fix this recently, see e.g.:
https://bugzilla.redhat.com/show_bug.cgi?id=1122970
http://lists.ovirt.org/pipermail/devel/2014-December/thread.html#9481 http://lists.ovirt.org/pipermail/devel/2014-December/thread.html#9565 (yes, two separate discussions this month).
Best, -- Didi