[quagga-dev 3491] Re: [PATCH] non-blocking I/O from client daemonstozebra

Simon Talbot simont at nse.co.uk
Sun Jun 5 15:58:31 BST 2005


Thanks for that Jeroen,

You have missed a bit of the conversation -- we have *NOT* had any
problems with the MD5 patch either, we have been using it for over a
year without problems. The problems are only arising when we test a very
new (unstable) version of Quagga (0.99.1) with work queuing implemented
and not with vanilla Quagga. The observation that we are using MD5 was
really just saying, that could be a factor in why we see the problem and
other people don't.

Thanks anyway for your comments,

Simon 


Simon Talbot MEng, ACGI
(Chief Engineer)
Tel: 0845 6440972
Fax: 0845 6440971


-----Original Message-----
From: J. Simonetti [mailto:jeroens at office.netland.nl] 
Sent: 05 June 2005 15:38
To: Simon Talbot
Cc: Paul Jakma; Andrew J. Schorr; Paul Jakma; Quagga Dev
Subject: Re: [quagga-dev 3483] Re: [PATCH] non-blocking I/O from client
daemonstozebra

Simon Talbot wrote:
> Yup -- sure am


> On Fri, 3 Jun 2005, Paul Jakma wrote:
> 
> 
>>Highly odd. Even excessive memory usage by an application still 
>>shouldn't cause a panic.
> 
> 
> Ah, wait a minute - you're using the Linux TCP-MD5 kernel patch arent
> you?

So am I. I have had no kernel panics since the aprox. 1,5 year I'm 
running md5.

Jeroen Simonetti

PS, I'm not using vanilla kernel! (Fedora Core 3 kernel 2.6.10-1.741_FC3

with md5 patch, plus a route-cache mem-leak patch) I haven't had any 
problems with previous fedora kernels either..




More information about the Quagga-dev mailing list