On 03/04/2018 04:23 PM, Tom Eastep wrote:
> On 03/04/2018 04:01 PM, Udo Schacht-Wiegand wrote:
>> Hello,
>>
>> this is my first post to this list, and I hope I can reply to an already
>> existing thread.
>>
>> I think I'm facing the same problem as this user reported:
>>> From: Brian J. Murrell <brian@in...> - 2017-11-29 17:33:24
>>> I have a shorewall6/shorewall6-lite installation where the router has
>> multiple IPv6 connections to the Internet. 
>>> [...] I see major packet loss on the eth0.2 provider:
>>
>> My shorewall6 version is 5.0.4 (on Ubuntu 16.04 LTS). The providers file:
>> #############################################################################################################
>> #NAME   NUMBER  MARK    DUPLICATE       INTERFACE       GATEWAY         
>>                        OPTIONS                         COPY
>> mkn     1       -       -               eth0           
>> 2001:xxxx:xxx::1                        track,primary
>> htp     2       -       -               eth1           
>> fe80::464e:6dff:fe15:789a               track,fallback
>>
>> I x'ed out our businesses public IPv6. "htp" is not a typo of "http",
>> but the name of our secondary provider. As their addresses are dynamic,
>> I use the link local address as gateway. When our primary provider 'mkn'
>> should go down, fallback on 'htp' will be done. I use a script to detect
>> the link, and can initiate a failover. This setup works fine.
>> The gateway is a Cisco router/DSL modem of the provider #1 in our
>> premisses, where I have no acces on.
>>
>> However I'm facing massive packet losses (>50%, only on IPv6) on
>> provider #1 using this setup. If I leave out the providers file, there
>> are no losses, same as Brian reported above. To test the connection I
>> usually do a "ping6 google.com <http://google.com>" or so. After a
>> "shorewall6 restart" ping6s initially go through, but then stops (after
>> about 10 to 40 pings). When I do a "ping6 gateway-address" from another
>> terminal, the pings to google will continue to go through.
>>
>> So I built this as workaround: I send continuous pings (by a cronjob)
>> against the router, then there are next to no losses. To me this looks
>> very similar to the problem described
>> here: https://forums.gentoo.org/viewtopic-t-855990-start-0.html 
>>
>> I tried to open shorewall6 for all ipv6-icmp traffic to and from the
>> router, but it didn't do much of a difference, as ipv6-icmp was allowed
>> anyway.
>>
>> Would appreciate any help ...
>> Cheers
>> Udo
> 
> Be sure that your Kernel is fully patched. This sounds like a problem
> that I, along with a number of others, have experienced; it was
> corrected in a subsequent kernel update. The problem is that the kernel
> ignores NDP who-has requests, which will kills the link. The constant
> pinging keeps the upstream router from issuing those requests. I
> employed that same workaround until the problem was finally resolved.
> 

One correction -- it is NDP neighbor solicitation requests that are
being ignored; 'who-has' requests are part of (IPv4) ARP.

-Tom
-- 
Tom Eastep        \   Q: What do you get when you cross a mobster with
Shoreline,         \     an international standard?
Washington, USA     \ A: Someone who makes you an offer you can't
http://shorewall.org \   understand
                      \_______________________________________________

Attachment: signature.asc
Description: OpenPGP digital signature

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Shorewall-users mailing list
Shorewall-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/shorewall-users

Reply via email to