[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAK6E8=fefVLUorMKCqJGSnq6Sc9LOKp7eUyOnh6h8rcOcJtQNQ@mail.gmail.com>
Date: Sat, 20 Jul 2013 23:04:56 -0700
From: Yuchung Cheng <ycheng@...gle.com>
To: Michael Sterrett <michael@...rretts.net>
Cc: netdev <netdev@...r.kernel.org>
Subject: Re: 3.10.0 network trace
On Sat, Jul 13, 2013 at 2:10 PM, Michael Sterrett <michael@...rretts.net> wrote:
> After changing those two sysctl settings the WARN hasn't reappeared.
> I'm willing to mess around with this box though if you want to try to
> narrow down the root cause with different settings and/or patches.
Could you try to disable F-RTO but enable loss probe by
sysctl net.ipv4.tcp_frto=0 net.ipv4.tcp_early_retrans=3
to help us identify if this is loss probe bug or f-rto. Thanks.
>
> On Fri, Jul 12, 2013 at 5:39 PM, Yuchung Cheng <ycheng@...gle.com> wrote:
>> On Fri, Jul 12, 2013 at 2:33 PM, Michael Sterrett <michael@...rretts.net> wrote:
>>> Thanks for the reply. The sysctl -a output is attached.
>>
>> Hi Michael: my guess is some of the recent changes like TLP or F-RTO
>> rewrite has bugs. It'd be best if you can provide tcp traces that
>> triggers this. If not and this WARN appears frequently enough, please
>> try i) sysctl tcp_early_retrans=2 ii) sysctl tcp_frto=0 and see if
>> that WARN goes away. Thanks.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists