[sheepdog] [PATCH 0/4] Handle pending block/notify event during reconnect

[Hitoshi Mitake]

At Mon, 8 Jul 2013 16:46:34 +0800,
Liu Yuan wrote:
> 
> On Mon, Jul 08, 2013 at 10:07:43AM +0800, Kai Zhang wrote:
> > 
> > On Jul 6, 2013, at 10:56 AM, Liu Yuan <namei.unix at gmail.com> wrote:
> > 
> > > On Fri, Jul 05, 2013 at 06:59:04PM -0700, Kai Zhang wrote:
> > >> This series is to enhance the implementation of sd_reconnect_handler()
> > >> by handling pending block/notify event during reconnect.
> > >> 
> > >> The first three patches prepare for the fourth one.
> > >> 
> > >> Kai Zhang (4):
> > >>  zookeeper: remove unneccessary zk operation
> > >>  sheep: handle error of notify_vdi_add()
> > >>  sheep: handle block/unblock/notify error
> > >>  cluster driver: handle pending block/notify event during reconnect
> > > 
> > > no test script?
> > > 
> > 
> > It is hard to control when the session timeout happen by using test script.
> > 
> > However, I would like to use other technology (mock or fault injection) to test it later.
> > 
> 
> Yes, it is hard to test every code path, but your previous functional test
> can reproduce some hidden problems sometimes on my box, so I think we need it.

DR based fault injector would be helpful for this case. DR provides
hooking mechanisms for many points, e.g. function calling and
returning: http://dynamorio.org/docs/group__drwrap.html

Thanks,
Hitoshi