[Varnish] #849: Session timeout while receiving POST data from client causes multiple broken backend requests

[Varnish]

#849: Session timeout while receiving POST data from client causes multiple
broken backend requests
----------------------+-----------------------------------------------------
 Reporter:  lew       |       Owner:  tfheen                                                               
     Type:  defect    |      Status:  new                                                                  
 Priority:  normal    |   Milestone:  Varnish 3.0 dev                                                      
Component:  varnishd  |     Version:  2.1.4                                                                
 Severity:  normal    |    Keywords:  503, post, backend write error: 11 (Resource temporarily unavailable)
----------------------+-----------------------------------------------------

Comment(by phk):

 See also #1041.

 There are a couple of issues in this:

 1. We use sess_timeout also during client->varnish body transfer.  It
 might be better with separate first/between timeouts like we have on the
 backend side.

 2. We don't buffer the body we get from the client, so restart from VCL is
 never a good idea.

 3. We have the automagic "try another fd" pseudo-restart, and it might
 happen in the middle of a body from the client.   This would be less of an
 issue if 2. is fixed, but we still need a better way to handle that.

-- 
Ticket URL: <https://www.varnish-cache.org/trac/ticket/849#comment:5>
Varnish <https://varnish-cache.org/>
The Varnish HTTP Accelerator