Default gateway on different net
Hi,
I have a server in location that provides ipv6 (not sure the ipv6 part is
important, but thats how I stumbled upon this). First a bit of information
about the server:
FreeBSD 8.1-RELEASE #0: Mon Jul 19 02:36:49 UTC 2010
1 NIC with ipv4 and v6 addresses.
I got an ipv6 subnet (/64) from the hosting provider but the default gateway is
not in the same subnet. The recommendation from the provider is simple: Create
a static route, that routes all traffic to the subnet with the gateway to the
interface. Then create default gateway.
What I have done is this:
ipv6_enable="YES"
ipv6_static_routes="defgw"
ipv6_route_defgw="2a01:::3180::1 -prefixlen 59 -iface re0"
ipv6_defaultrouter="2a01:::3180::1"
ipv6_ifconfig_re0="2a01:::3183::2 prefixlen 64"
( Notice the slight difference in the ip on the nick :3183: and the gateway
:3180: )
If I reboot the machine, the following happens:
*) The nic re0 got the ip: 2a01:::3183::2 - ifconfig:
inet6 2a01:::3183::2 prefixlen 64
*) The static route is in the routing table:
2a01:::3180::/59 6c:62:6d:0d:9b:e8 US re0
*) But there is no default route :(
I then stumbled upon line 1057 in /etc/network.subr:
ipv6_static_routes="default ${ipv6_static_routes}"
If I understand this correctly, the ipv6_defaultrouter is simply added in the
static_route configuration (which is added later).
But I also recall that I cannot create a default route to an ip not in my
subnet, if the static route is missing. And the above code place the default
route as the very first.
If I move default to the "last" position:
ipv6_static_routes="${ipv6_static_routes} default"
And reboots, the routing table shows my static route and the default route just
fine:
Internet6:
Destination GatewayFlags Netif Expire
default 2a01:::3180::1 UGS re0
2a01:::3180::/59 6c:62:6d:0d:9b:e8 USre0
But! I still can't get any ipv6 packets out of the server. Trying to ping6 the
gateway (or anything else outside) simply returns a: ping6: sendmsg: Operation
not permitted
I have tried tcpdumping ipv6 packets on re0, but absolutely nothing is showing
up when either ping6 or telnet to something outside on an ipv6 address
(firewall is disabled btw). If I tcpdump while ping6 from outside to
2a01:::3183::2, ipv6 packets is showing just fine.
If you are still reading this email, thanks :)
Now to the actual questions:
1) Is the hosting provider actually forcing me to do something "bad" og plain
wrong?
2) Should the default route really be the last route to be added in
/etc/network.subr:1057? (not sure if there is a reason to why its first - but I
feel guilty messing with that file directly :) )
3) Am I setting this up in a completely wrong way?
( At this point it feels like "everything is right, but I miss _that_
fundamental option/config somewhere )
Any hint, tips or trick is highly appreciated.
Best regards,
Lasse Brandt___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
Re: zfs very poor performance compared to ufs due to lack of cache?
on 15/09/2010 19:31 Andriy Gapon said the following: > on 15/09/2010 19:00 Andre Oppermann said the following: >> Is there a quick way of deciding within sendfile(2) whether a file resides >> on a filesystem that doesn't use the buffer cache? > > I don't know of any reliable way to do it. > If I understood correctly what Kostik suggested in a chat, then we could abstract core sendfile<->vm<->fs logic into a new vop, say vop_sendpages (don't mind the name much). std_sendpages would do what code in kern_sendfile does today and would be used by most of filesystems. But some filesystems could override it e.g. to avoid using VM page cache. Or, perhaps, we could even add EXT_ZFSARC type of mbuf that would know how to work with data buffers held in ARC and unhold them when done. That would make sendfile work on ZFS through a single cache. -- Andriy Gapon ___ freebsd-net@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
ROUTETABLES & IPv6?
Hi, am I correct in that multi-FIB support is currently only working for IPv4? At least "setfib route add -inet6 " adds the route into all FIBs, not only FIB . (For IPv4, it works correctly.) Or am I missing something? Thanks, Lars PS: Please CC me on replies.
Re: ROUTETABLES & IPv6?
Hello Mr.Eggert, Le Mar 21 septembre 2010 14:51, Lars Eggert a écrit : > Hi, > > > am I correct in that multi-FIB support is currently only working for > IPv4? At least "setfib route add -inet6 " adds the route > into all FIBs, not only FIB . (For IPv4, it works correctly.) > > Or am I missing something? > You're not, as far as I know setfib sends messages to all fibs without taking notice of the fib requested by the user. There is a PR in the base regarding this => PR 134931. http://www.freebsd.org/cgi/query-pr.cgi?pr=134931&cat= I don't think this patch is in the base, not even in HEAD. With regards to IPv6 and routing tables, Qing Li should be able to provide thorough feedback on this. > > Thanks, > Lars > > > PS: Please CC me on replies. Aman Jassal Wisdom comes from experience. Experience comes from a lack of wisdom. ___ freebsd-net@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
Re: ROUTETABLES & IPv6?
On 21.09.2010 14:51, Lars Eggert wrote: Hi, am I correct in that multi-FIB support is currently only working for IPv4? At least "setfib route add -inet6 " adds the route into all FIBs, not only FIB. (For IPv4, it works correctly.) It seems multi-FIB is incomplete for IPv6. Also radix multi-path is incomplete for IPv6 at the moment. Ingo Flaschberger has started working on radix multi-path by fixing bugs and porting further work from OpenBSD. IPv6 support is also on the map. He's working with Qing Li and me refine these changes to eventually get them into -current. Multi-FIB is a natural extension of this topic that may get tackled in a next step. -- Andre ___ freebsd-net@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
Re: ROUTETABLES & IPv6?
On 9/21/10 7:05 AM, Andre Oppermann wrote: On 21.09.2010 14:51, Lars Eggert wrote: Hi, am I correct in that multi-FIB support is currently only working for IPv4? At least "setfib route add -inet6 " adds the route into all FIBs, not only FIB. (For IPv4, it works correctly.) It seems multi-FIB is incomplete for IPv6. Also radix multi-path is incomplete for IPv6 at the moment. Ingo Flaschberger has started working on radix multi-path by fixing bugs and porting further work from OpenBSD. IPv6 support is also on the map. He's working with Qing Li and me refine these changes to eventually get them into -current. Multi-FIB is a natural extension of this topic that may get tackled in a next step. Multiple FIB support has not been added to the IPV6 code. it requires someone who actually understands IPV6 (which is not me) but it is not really rocket science. if there is someone who has access to, and understands IPV6, then I can guide them on what needs to be done. I know that CISCO STBU (ex Ironport) are looking for this to be done. Maybe they can sponsor someone to do it. (but I don't have time). ___ freebsd-net@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
Re: kern/146792: [flowtable] flowcleaner 100% cpu's core load
The following reply was made to PR kern/146792; it has been noted by GNATS. From: Yar Tikhiy To: bug-follo...@freebsd.org Cc: km...@freebsd.org Subject: Re: kern/146792: [flowtable] flowcleaner 100% cpu's core load Date: Wed, 22 Sep 2010 07:00:47 +1000 FWIW, this flowtable problem still can be triggered in 8.1-RELEASE. Yesterday it was seen on a machine with approx. 300 routes from OSPF -- no BGP full views involved. The machine was unable to reboot cleanly, was stuck forever waiting for vnlru to exit. Can it be a locking issue? Thanks, Yar ___ freebsd-net@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
mpd5 and mgetty functionality
Hi friends, I need to know if mpd5 can do something similar to: "/usr/local/sbin/megtty -R 30 /dev/cuau0" and "rings 2 in mgetty.config" exactly what I need it's something similar to -R 30 and ring 2 that gave me the same functionality. BUT with mpd5, i'm running 8.1, I got mpd5 with 2 profiles, one dial-out (ISP, works OK!) and one dial-in (From external to my home server) for dialin I run "mpd5 dialin" (see below) but when I dial into my server mpd5 answer the phone inmediatly, I don't want this, I want a similar functionality like mgetty with -R "SECONDS" and "rings 2" option, I need to dial and make let's say 2,3 or four rings and then hung up, mpd5 will never answer this call and when I hung up mpd5 will wait for a call and then when I dial again in a period of 20 to 30 seconds this will tell mpd5 that it's a conection attempt and answer the call otherwise dont's answer it. All this because in my home my wife use the phone too :) and I dont't want any friends call get this noise in his ears! below my dialin: section: ps: oh! sorry for my english ( not my native languaje :] ) LeaL :] #DIAL-IN CONF in mpd.conf dialin: create bundle static B1 set iface idle 900 set ipcp ranges 192.168.1.169/32 192.168.1.130/32 create link static L1 modem set link action bundle B1 set link no chap eap set link enable pap set link max-redial -1 set auth authname user_ppp set modem device /dev/cuau0 set modem var $DialPrefix "DT" set modem idle-script AnswerCall set link enable incoming -- Este mensaje le ha llegado mediante el servicio de correo electronico que ofrece Infomed para respaldar el cumplimiento de las misiones del Sistema Nacional de Salud. La persona que envia este correo asume el compromiso de usar el servicio a tales fines y cumplir con las regulaciones establecidas Infomed: http://www.sld.cu/ ___ freebsd-net@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
