[SSSD] [PATCH] Automatically retry SBUS requests within the timeout period

Simo Sorce ssorce at redhat.com
Fri Mar 12 14:15:15 UTC 2010


On Thu, 11 Mar 2010 08:05:56 -0500
Stephen Gallagher <sgallagh at redhat.com> wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On 03/11/2010 07:57 AM, Stephen Gallagher wrote:
> > On 03/10/2010 04:44 PM, Stephen Gallagher wrote:
> >> On 03/10/2010 04:12 PM, Simo Sorce wrote:
> >>> On Wed, 10 Mar 2010 15:30:09 -0500
> >>> Stephen Gallagher <sgallagh at redhat.com> wrote:
> > 
> >>>> This patch adds a new helper function, sbus_conn_send() that will
> >>>> handle automatically resending a request if the connection is
> >>>> closed at the time of the initial send.
> > 
> >>> No patch :-)
> > 
> > 
> >> Whoops! Here we go.
> > 
> > 
> > I discovered a bug in my previous patch (I was testing reply ==
> > NULL, rather than pending == NULL). This new patch corrects that.
> > 
> > 
> 
> Sorry, I missed a hunk with my last patch. THIS one should be right...

Sorry I have to nack, to summarize the IRC discussion, the 50ms loop is
not really the best approach, a callback mechanism in the re-connection
code to which you can register and have pending requests queued (unless
they timeout meanwhile) and be pushed out once we reconnect is a better
approach and avoids having around a lot of calls that retry operations
every 50ms without any chance of success until re-connection is
performed.

Simo.

-- 
Simo Sorce * Red Hat, Inc * New York



More information about the sssd-devel mailing list