On Wed, Mar 06, 2013 at 01:51:48AM -0300, Lucas Meneghel Rodrigues wrote:
> On Mon, Mar 4, 2013 at 8:29 AM, Satheesh Rajendran
> <[email protected]> wrote:
> > Hi Lucas,
> >
> > I had sent the reworked V2 patch aswell which includes the switch in runner,
> > Could you please validate both the options based upon the discussion here
> > and push to the appropriate to next.
> 
> I did take a look at this patchset for inclusion, and there are things
> that still need to be addressed:
> 
> 1) The output of "hostname" is good only if the host has a fully
> qualified domain name. If it's a laptop installed on a domestic
> network, this will fail and gluster will say that 'host blabla is not
> a friend'. Therefore, if the hostname can't be resolved, this must
> fall back to IP.

Agreed, I would check on this.

> 2) There is the use of the commands API throughout the module. I'd
> prefer this usage is replaced with utils.system, or utils.run,
> whatever is more convenient.

Sure, I will replace.

> 3) Indeed, you don't need to create a runner switch.
Thanks, I would revert in the next version.
> 4) With the introduction of variants in guest_hw.cfg, you necessarily
> have to update a lot of test.cfg files inside virt-tests to say things
> like 'no glusterfs_support', otherwise all control files and test runs
> will create more tests than they should.

Sure, Will add the details in tests.cfg. 
I can update the wiki and give the reference to the wiki in the test.cfg.

> 5) With all enabled, qemu-img will still try to create an image, which
> is incorrect behavior.
> 
> 01:37:26 ERROR| Could not create image, failed with error
> message:Command </home/lmr/Code/qemu/qemu-img create -f qcow2
> gluster://192.168.2.5:0/test-vol/f18-64.qcow2 10G> failed, rc=1,
> Command returned non-zero exit status
> * Command:
>     /home/lmr/Code/qemu/qemu-img create -f qcow2 gluster://192.168.2.5:0
>     /test-vol/f18-64.qcow2 10G
> Exit status: 1
> Duration: 0.00223588943481
> 
> stderr:
> qemu-img: Unknown protocol 'gluster://192.168.2.5:0/test-vol/f18-64.qcow2'
> 
> I'd say implementing glusterfs as a new image type, that does not use
> qemu img to create it seems better than the approach you're taking.
> 

The following infrastructure is required to run this test
Qemu- 1.3, 03Dec2012
GlusterFS-3.4
Libvirt-1.0.1, 15Dec2012 (This is not required for this patch, would be 
helpful for the libvirt support in future)

Agreed, I should have added this details in the first place in a wiki,
that would have cleared the confusions, I will update it the wiki.

> 6) Glusterfs usage has to be covered in the wiki
Sure.
> 7) Some example config files for the qemu test would also help.
This might not be required as it is going be just a enabler, then
all the tests can run on the image created on gluster volume.
> 
> So I'd say there are a lot of things to be fixed. Let's get to work on them.
> 
> Thanks!
> 

Thanks for the detailed review, Would send the updated patch soon.

Regards,
-Satheesh.

_______________________________________________
Virt-test-devel mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/virt-test-devel

Reply via email to