(forgot to add the mailing list)
I am using kernel 3.2 under Debian but I am probably going to switch to
Ubuntu 14.04 shortly (so 3.13 it appears). I have completely redundant
hardware and secondary IP addresses on each link so this will be painless.
http://www.spinics.net/lists/netdev/msg204673.html
It looks like you've already dealt with this issue. I will manually clamp
MSS until upgrading to 3.13.
Thanks again,
Darren
On Tue, Apr 15, 2014 at 1:24 PM, Darren Ginter <[email protected]> wrote:
> I am using kernel 3.2 under Debian but I am probably going to switch to
> Ubuntu 14.04 shortly (so 3.13 it appears). I have completely redundant
> hardware and secondary IP addresses on each link so this will be painless.
>
> http://www.spinics.net/lists/netdev/msg204673.html
>
> It looks like you've already dealt with this issue. I will manually clamp
> MSS until upgrading to 3.13.
>
> Thanks again,
>
> Darren
>
>
> On Tue, Apr 15, 2014 at 12:45 PM, Timo Teras <[email protected]> wrote:
>
>> On Tue, 15 Apr 2014 11:57:48 -0400
>> Darren Ginter <[email protected]> wrote:
>>
>> > It appears that my diagnosis was incorrect as this worked out to be
>> > another MTU issue. With all of the IPsec and GRE overhead, I needed
>> > to reduce MTU down to 1414. Initially, I had done that on the GRE
>> > tunnels and the interface on my voice VLAN. But I had not done that
>> > on the VLAN with the PCs because everything seemed to work fine.
>> > This issue only presented itself on a handful of PCs and only a few
>> > times per day. Here's someone else who had the same issue (random
>> > Outlook EventID 26):
>> >
>> >
>> http://ask.wireshark.org/questions/652/connection-to-microsoft-exchange-has-been-lost-outlook-will-restore-the-connection-when-possible
>> >
>> > Yesterday, I tried to lower MTU on the PC LAN interface of the
>> > routers but that caused internet https sessions to break (the routers
>> > serve the default route for internet and intranet). A client-side
>> > reboot may have resolved this (PMTU?) but I was not able to test
>> > during the day just due to the sheer numbers. So I have subsequently
>> > increased the router's PC LAN MTU back to 1500 and reduced MTU on the
>> > email server. This is obviously not ideal but it appears to be
>> > working until I can get a proper lab configured for testing.
>>
>> I've have gone through great pain to fix PMTU issues. Which kernel
>> version you use on the nhrp nodes? (See my earlier mail on known kernel
>> regressions.)
>>
>> The GRE/IPsec overhead will depend on point-to-point tunnel basis based
>> on negotiated ciphers and depending if NAT is in use or not.
>>
>> Alternatively it could be some ISP pmtu issue causing blackholed
>> ESP packets? In that case doing clampmss on gre1 might help.
>>
>> - Timo
>>
>
>
------------------------------------------------------------------------------
Learn Graph Databases - Download FREE O'Reilly Book
"Graph Databases" is the definitive new guide to graph databases and their
applications. Written by three acclaimed leaders in the field,
this first edition is now available. Download your free book today!
http://p.sf.net/sfu/NeoTech
_______________________________________________
opennhrp-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opennhrp-devel
