3:30 a.m.
----- Original Message -----
| From: "Ofri Masad" <omasad(a)redhat.com>
| To: "Giuseppe Vallarelli" <gvallare(a)redhat.com>
| Cc: "Livnat Peer" <lpeer(a)redhat.com>, engine-devel(a)ovirt.org
| Sent: Tuesday, June 25, 2013 8:19:30 AM
| Subject: Re: [Engine-devel] VNic QoS measure units and attributes.
|
| Hi Giuseppe,
|
| Thanks for the review. I am aware of the need to add conversion of units in
| the engine.
| As for the mandatory fields - we decided to make all 6 fields mandatory to
| the engine users (we know libvirt does not enforce this).
| Leaving some fields empty may cause some ambiguity since libvirt is filling
| the remaining fields by it self.
| for example: if libvirt implementation will change in future releases, the
| user may get different peak and burst values without even knowing about it.
I see, in this case your choice makes perfect sense, Thanks for sharing.
Giuseppe
|
| thank again
| Ofri
|
| ----- Original Message -----
| > From: "Giuseppe Vallarelli" <gvallare(a)redhat.com>
| > To: "Livnat Peer" <lpeer(a)redhat.com>
| > Cc: engine-devel(a)ovirt.org
| > Sent: Monday, June 24, 2013 9:22:59 PM
| > Subject: Re: [Engine-devel] VNic QoS measure units and attributes.
| >
| > ----- Original Message -----
| > | From: "Livnat Peer" <lpeer(a)redhat.com>
| > | To: "Giuseppe Vallarelli" <gvallare(a)redhat.com>
| > | Cc: engine-devel(a)ovirt.org
| > | Sent: Monday, June 24, 2013 6:31:43 PM
| > | Subject: Re: [Engine-devel] VNic QoS measure units and attributes.
| > |
| > | On 06/24/2013 03:07 PM, Giuseppe Vallarelli wrote:
| > | > Hi Ofri, I've been reading your design page for this feature [1] and
| > | > and
| > | > I've found out
| > | > that there's a possible unit of measure mismatch regarding: Average,
| > | > Burst
| > | > and Peak, clearly
| > | > makes sense to use Mbps on the gui but I wonder if you're aware that
| > | > the
| > | > vdsm api expects
| > | > respectivly: kB/sec, kB and kB/sec (for average, burst and peak), so a
| > | > conversion is needed.
| > | > As a side note the only compulsory attribute needed to define QoS for
| > | > incoming and outgoing
| > | > traffic is average, burst and peak are optional (see [2]), this is not
| > | > clear by looking
| > | > at the mockups and description.
| > | >
| > | > Cheers, Giuseppe
| > | >
| > |
| > | Hi Giuseppe,
| > | Do you know why VDSM API exposes KB/Sec? don't you think that Mbps is
| > | more intuitive?
| >
| > Hi Livnat, I made this choice because libvirt uses those unit of measure
| > and I considered vdsm to be almost at the same (low) level. This part of
| > the API has been already implemented/documented and discussed in a
| > previous thread [1]. I agree that Mbps is much more intuitive for the
| > end user. We can decide where we should make the conversion, if at the
| > engine level or in vdsm. I'm fine with both choices.
| >
| > Cheers Giuseppe.
| >
| >
| > [1] Add traffic shaping parameters for a network interface.
| >
| > |
| > |
| > | Livnat
| > |
| > | > [1] http://www.ovirt.org/Features/Network_QoS
| > | > [2] http://libvirt.org/formatdomain.html#elementQoS
| > | >
| > | >
| > | > _______________________________________________
| > | > Engine-devel mailing list
| > | > Engine-devel(a)ovirt.org
| > | > http://lists.ovirt.org/mailman/listinfo/engine-devel
| > | >
| > | >
| > _______________________________________________
| > Engine-devel mailing list
| > Engine-devel(a)ovirt.org
| > http://lists.ovirt.org/mailman/listinfo/engine-devel
| >
|