[Gluster-devel] regression in master

[Niels de Vos]

On Tue, May 02, 2017 at 11:57:16AM +0530, Atin Mukherjee wrote:
> With latest HEAD, all volume set operation fails for a volume which is in
> STARTED state. I've figured out that commit 83abcba has caused it, what
> makes me baffled is this patch had passed all the regression. Are we
> installing nfs so files in slave machines which can only way to avoid these
> failures? With source install I hit it 10 out of 10 times.
> 
> https://review.gluster.org/#/c/17149/ addresses the issue. Appreciate your
> review.

The Gluster/NFS server .so is not built by default, but the test scripts
were extended to have the "./configure --enable-gnfs" option. If you do
source installs and want/need to test Gluster/NFS, you would need to
pass this option as well.

Our tests in the glusterfs repository assume that most configurable
features are available. I do not think we have a way to pass the
./configure or build options to the tests and run only a selection. Many
tests use Gluster/NFS, and skipping all NFS related tests when the
xlator is not present does not look straight forward to me.

HTH,
Niels
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: not available
URL: <http://lists.gluster.org/pipermail/gluster-devel/attachments/20170502/914d53ba/attachment-0001.sig>