[Yum] yum timeout

[Troy Dawson]

Michael Stenner wrote:
> On Wed, Apr 16, 2003 at 11:06:53AM -0500, Troy Dawson wrote:
> 
>>Hi Seth,
>>The question came up about timeouts with yum.
>>
>>Here is the scenario, so that you know what I'm talking about.
>>
>>A user set's up some type of firewall.  They make alot of holes in it, and 
>>what have you, but for some reason they don't allow things to come in.  So 
>>basically yum connects to the server and just sits there because for some 
>>reason there isn't anything comming back.  Will yum eventually timeout?
>>
>>I'm not exactly sure this is the problem, I'm still investigating, but it's 
>>something I'm curious about anyway.
> 
> 
> Yes, sorta.  It will definitely time out.  I do not remember off the
> top of my head what kind of exception a timeout results in though,
> which will affect how it's handled.  If someone beats me to this,
> and it's not handled well (as in yum dumps a traceback) then let me
> know and I'll add that into the current "excepts".
> 
> The problem is that timeouts are not handled well in python 2.2.  You
> get whatever the default timeout is, and there's no way (unless you
> do a lot of work) to adjust it.  There's a module out there (called
> timeoutsocket.py, I think) that "fixes" this by _pushing_ socket
> methods into the standard socket module.  I've used it before and
> works nicely.  That could be used to sort-of impose timeouts on socket
> connections.  With a single import and function call, you can set the
> default timeout for all sockets.  You can also set timeouts
> per-socket, but that's less useful here since you don't normally have
> direct access to the sockets.
> 
> Bottom line: it's easy to make yum timeout nicely.  If you want it to
> have settable timeouts, that's a little harder, but still not too bad.
> Neither of these have been deliberately implemented or tested (that I
> know of).
> 
> 					-Michael

OK, thanks Michael,
That's what I wanted to know.
It turns out that our problem wasn't the firewall (though there was one 
involved) but a different version of python (2.1), so I'll know about the 
timeout's next time that comes up.
Troy
-- 
__________________________________________________
Troy Dawson  dawson at fnal.gov  (630)840-6468
Fermilab  ComputingDivision/OSS  CSI Group
__________________________________________________