Default Timeout for Push Operation

[Matt Hawley]

Hi Benoit,

Out of curiosity, I tried pushing without SSL, and things worked. I re-tried under SSL and the same behavior occurred. Unfortunately, this isn't a solution, as we need to have SSL enabled.

Regarding wireshark - I've never used it and got a little overwhelmed by what I need to do. Any guidance? Also, I've never ran with --debugger, so if you could point me to some info on that, it would be helpful.

Matt

-----Original Message-----
From: Benoit Boissinot [mailto:benoit.boissinot at ens-lyon.org] 
Sent: Tuesday, February 23, 2010 9:19 AM
To: Matt Hawley
Cc: Patrick Downey; mercurial at selenic.com
Subject: Re: Default Timeout for Push Operation

[please do not top-post, and try keep the cc-list when replying]

On Tue, Feb 23, 2010 at 04:56:23PM +0000, Matt Hawley wrote:
> Hi Pat,
> 
> We don't have any hooks in place, however it's quite possible maybe 
> the server (as you said) is performing some large work on the 
> backend... but why would that show no utilization of the CPU? I have 
> also modified the timeout for both CGI and the application pool, but 
> no matter what, the push operation will take ~280 seconds before 
> returning a failure. Oddly enough, none of the timeouts I've set 
> equate to this value, so that must be something on the client side.

A couple ideas of things to do client side:
- run with hg --debug to get an idea what the client is doing (hg
  --debugger can help too)
- get a network dump to see what is happening (is there some ongoing
  conversation between the client and the server).

Cheers,

Benoit

--
:wq