[ovirt-devel] VDSM sync meeting minutes July 8th, 2014

[Dan Kenigsberg]

We had relatively low attendence today and we did not discuss pending 3.5 issus
deeply enough. If you have urgent 3.5 issue that needs fix/review, please chime in!

live-merge:
- we have 2 complex patches by Adam
  http://gerrit.ovirt.org/#/q/status:open+project:vdsm+branch:master+topic:livemerge,n,z

jsonrpc
- we have 3 pending patches by Piotr
  http://gerrit.ovirt.org/#/q/owner:%22Piotr+Kliczewski%22+status:open+project:vdsm,n,z

  The most flaky of them adds lastClientIface element to getVdsCaps over
  jsonrpc.
  That element is used by Engine to deduce which network interface is used to
  carry management traffic, and build ovirtmgmt network on top of this.

  I suggest in
      Bug 1117303 - add a new getRoute() verb
  to introduce a new verb for this purpose, but until this happens in 3.6, we'd
  need a hack to expose it here.

pthreading's superfluous wakeups
- Nir's reported bug is still in effect, and his suggested patches awaits
  review.

qos patches
- Martin has plenty of patches. Most urgent is the log spam avoidence of
  http://gerrit.ovirt.org/#/c/29504/9/vdsm/virt/vm.py
  All should be reviewed by virt folks, due to the natural dependency.

copious rebases
- It is sometimes tempting to use a single topic branch for multiple patches
  that do not really depend on each other. The down size of this is that on
  every rebase, reviewers are swamped with emails; it is also less apparent
  that a certain patch can be merged without waiting for previous patches.

  So if you can break your serial branch to multiple branches with the same
  topic, you should probably do it.

Hear you in a fortnight,
Dan.