[Chandler-dev] proposal: patch alpha2 to extend timeout

[Katie Capps Parlante]

As noted in Bug 5996, dogfooders with big calendars are experiencing 
timeouts when syncing with cosmo-demo.

Initial investigation indicated that PROPFIND was slower than it ought 
to be. Jared upgraded cosmo-demo to 0.4, which included some work on 
Brian Moseley's part to speed up PROPFIND responses. This likely helped 
somewhat, but didn't solve the problem entirely.

While we dig into resolving the root problem, we could help dogfooders 
by extending the timeout in alpha2 (it is currently 30 seconds). This 
would allow dogfooders to continue using alpha2, even if it is slower 
than we'd like. We'd also continue working on the root problem.

 From Morgen's bug comments:

> Yeah, we probably need to have Chandler "wait a little longer" for Cosmo to
> respond.  On the trunk and the bgsync branch I have Chandler set to wait 2
> minutes before "timing out", in other words "giving up".  I would be happy to
> apply the increased timeout change to alpha2 if I have approval.
> 
> The patch is to add this method to WebDAV.py, class ChandlerServerHandle:
> 
> 
>     def addRequest(self, request):
>         # Make all requests going through this ServerHandle have a
>         # 2 minute timeout
>         request.timeout = 120 # seconds
>         return super(ChandlerServerHandle, self).addRequest(request)
> 

The proposal is that we go ahead and patch alpha2, spinning a new build 
to give to dogfooders.

Thoughts?

Cheers,
Katie