Date: Sat, 21 Oct 2006 18:58:29 +0200 From: Milan Obuch <net@dino.sk> To: freebsd-net@freebsd.org Subject: Re: Somewhat weird net behavior Message-ID: <200610211858.30338.net@dino.sk> In-Reply-To: <20061021162630.I2879@fledge.watson.org> References: <200610211618.11472.net@dino.sk> <20061021162630.I2879@fledge.watson.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Saturday 21 October 2006 17:28, Robert Watson wrote: > On Sat, 21 Oct 2006, Milan Obuch wrote: > > I know this is really minimum information here, but I need just an idea > > what to look for. It is strange for me - is it some memory leak? How > > could it be cleared with simple ifconfig up? Did anyone seen something > > similar? > > It is pretty minimal information :-). > Agreed. I need some help in my investigation. I have no ide what to look for. > ifconfig down / up will cause the device driver to be reset, the link to be > re-negotiated, etc. The buffer full error from ping means that the > outbound interface queue has filled, and if the error doesn't go away, it > means it's likely wedged (no longer being processed). This could reflect a > bug in the device driver, a problem with the switch, etc. In the "wedged" > state, what does ifconfig report about link negotiation? When you reset > the interface, does the negotiated link type change? > I did only ifconfig up, not down/up sequence. That's a bit weird for me. Anyway, on next ocasion, I will do ifconfig before/after ifconfig up. Could there be some interesting information in some netstat output too? It happens once - twice a week, so I would rather be a bit prepared beforehand. Thanks for help. Regards, Milan -- This address is used only for mailing list response. Do not send any personal messages to it.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200610211858.30338.net>