|
High last hop lateency...
![[us]](/s/countries/us.gif) Shadow Hawkins on Wednesday, 13 September 2006 19:44:56
I am just curious. I just signed up and have yet to build enough credits for a subnet, but I have my tunnel setup and my assiged WAN IP pings. If I do an IPv4 traceroute from my location to my POP I am seeing a latency of around 14ms. But if I do a traceroute from the outside via IPv6 to my assigned IPv6 endpoint address, my latency is showing over 100ms.
Is this normal ir is something going on? I have done some searching and not found any info to help guide me.
My Tunnel: T9798
IPv6 traceroute:
FROM: Towardex Systems Group
Waltham, Massachusetts
United States [2001:4830:100:c0ff:ee::2]
TO: 2001:4830:1600:3b::2
1 mis.ge-0-2-0.r2.bos.twdx.net (2001:4830:100:20::1) 0.469 ms 0.189 ms 0.412 ms
2 0.ae2-4000m.c1.bos.twdx.net (2001:4830:1ff::193) 0.504 ms 0.491 ms 0.521 ms
3 0.ae5-2000m.c2.bos.twdx.net (2001:4830:1ff:224::226) 0.390 ms 0.581 ms 0.462 ms
4 0.10ge2-0.cr1.bos1.us.occaid.net (2001:4830:e1::1) 0.390 ms 0.583 ms 0.313 ms
5 192.ge0-0.cr1.ewr1.us.occaid.net (2001:4830:ff:15c4::1) 6.602 ms 6.529 ms 6.280 ms
6 ge-0-1-0.cr1.iad1.us.occaid.net (2001:4830:ff:f150::2) 12.627 ms 12.336 ms 12.474 ms
7 2001:504:0:2:0:3:3437:1 (2001:504:0:2:0:3:3437:1) 12.940 ms 12.805 ms 12.529 ms
8 sixxs-asbnva-gw.customer.occaid.net (2001:4830:e6:7::2) 13.362 ms 13.006 ms 12.987 ms
9 cl-60.qas-01.us.sixxs.net (2001:4830:1600:3b::2) 112.542 ms 112.070 ms 111.554 ms
Done!
IPv4 traceroute:
traceroute to 66.117.47.228 (66.117.47.228), 30 hops max, 38 byte packets
1 linksys.ropeguru.int (192.168.1.1) 1.351 ms 1.082 ms 0.777 ms
2 73.141.210.1 (73.141.210.1) 10.639 ms 12.425 ms 8.130 ms
3 68.86.175.161 (68.86.175.161) 7.683 ms 9.426 ms 10.273 ms
4 te-1-1-ur02.staplesmllrd.va.richmond.comcast.net (68.86.172.50) 7.791 ms 9.051 ms *
5 te-1-3-ar01.staplesmllrd.va.richmond.comcast.net (68.86.172.94) 9.467 ms 9.055 ms 18.823 ms
6 12.118.122.13 (12.118.122.13) 10.663 ms 11.955 ms 15.856 ms
7 tbr2-p011701.wswdc.ip.att.net (12.123.8.30) 14.383 ms 13.188 ms 14.273 ms
MPLS Label=31328 CoS=5 TTL=1 S=0
8 gar1-p3150.ascva.ip.att.net (12.122.80.197) 74.449 ms 13.070 ms 14.499 ms
9 12.117.208.26 (12.117.208.26) 15.972 ms 12.680 ms 12.086 ms
10 213.200.66.110 (213.200.66.110) 12.454 ms 11.574 ms 22.577 ms
11 209.222.130.22 (209.222.130.22) 15.115 ms 15.062 ms 13.333 ms
12 iad0-sixxs.hotnic.net (66.117.47.228) 14.417 ms 12.762 ms 11.258 ms
High last hop lateency...
But those are two different directions you are trying and there seems to be a hop at savvis which sometimes is reporting over 150ms of latency, even though it seems to be around the corner most of the time.
Better use Traceroute and trace from the PoP directly, you can see for yourself then how packets flow.
Note that ISP's sometimes ratelimit traffic that is unknown. As such proto-41 traffic might be put in the 'unknown' class and simply rate limitted.
High last hop lateency...
Shadow Hawkins on Wednesday, 13 September 2006 19:53:06
Ok, sorry about the noise. Idiot me did not try a traceroute to 2001:4830:1600:3b::1, which is SixXS' side of the tunnel. Even that address has high latency. Seems that the hop between Occaid and SixXS is the culprit.
High last hop lateency...
Do a v4 traceroute to the PoP and a v4 traceroute from the PoP to your endpoint and you should see the same things.
High last hop lateency...
Shadow Hawkins on Wednesday, 13 September 2006 20:12:49
Hi, thanks for the quick replys. Makes perfect sense and I am seeing low latency when testing as you suggested.
I have one last quick question.
If I traceroute from you PoP usqas01, which I am assigned to, from the tools on the SixXS web site, I cannot get any output for either 2001:4830:1600:3b::1 or 2001:4830:1600:3b::2. Additionally, 2001:4830:1600:3b::2 in my account is showing 100% packet loss and no ping stats. Yet traceroutes from your other PoP's do see my IP responding.
Is this normal?
High last hop lateency...
No that is not normal, and in combination with your traceroutes, I would suggest you to check your routing tables and make sure you are not sending your answers back over a different tunnel.
High last hop lateency...
Shadow Hawkins on Wednesday, 13 September 2006 21:14:21
I am not sure what is going on. I am running the aiccu on a Linksys router running dd-wrt v23 SP2. The conf file for aiccu has my info in it. The only other config I have going on in ref to the IPv6 tunnel is:
sleep 5
WANIP=$(ip -4 addr show dev vlan1 | grep 'inet ' | awk '{print $2}' | cut -d/ -f1)
if [ -n "$WANIP" ]
then
V6IP=$(printf '2002:%02x%02x:%02x%02x::1/64' $(echo $WANIP | tr . ' '))
ip tunnel add SixXS mode sit ttl 255 remote 66.117.47.228 local $WANIP
ip link set SixXS up
kill -HUP $(cat /var/run/radvd.pid)
fi
Which grabs the external IP, adds the SixXs tunnel, then brings up the link. I am still getting responses from every location I try except for the usqas01 selection on the SixXs web site.
BTW - Latency is down now also.
High last hop lateency...
2002:... that is 6to4, that has nothing to do with SixXS, and most likely you then also have some other odd routes (see 'ip -6 ro sho')
Clearly you are misconfiguring your setup.
Notez bien that AICCU takes care of completely setting up your tunnel.
Please read "Reporting Problems", there is a nice checklist there on what you can check to fix this issue.
High last hop lateency...
Shadow Hawkins on Thursday, 14 September 2006 21:27:38
Hi,
I hope this is my last question. I have successfully configured Aiccu and the link is up as I can traceroute to 2001:4830:1600:3b::1 and 2001:4830:1600:3b::2. If I shut Aiccu down then I cannot traceroute to either address.
Now my final question, which is something I asked yesterday. When I traceroute from any location listed in the traceroute utility here on SixXs, everything comes back fine. But, if I try and traceroute from the PoP I am assigned to, the results come back blank.
My question is, is this normal or is there a problem on usqas01.
If someone out there could possibly try a traceroute to their end point and the SixXS endpoint of their tunnel from the PoP they are assigned to, I would like to know if you get the same results I do. I am just wondering if maybe I do not get any results because it is only one hop in the traceroute.
High last hop lateency...
From the PoP:
$ traceroute6 2001:4830:1600:3b::1
traceroute6 to 2001:4830:1600:3b::1 (2001:4830:1600:3b::1) from 2001:4830:1600:3b::1, 64 hops max, 12 byte packets
1 gw-60.qas-01.us.sixxs.net 0.473 ms 0.136 ms 0.124 ms
$ traceroute6 2001:4830:1600:3b::2
traceroute6 to 2001:4830:1600:3b::2 (2001:4830:1600:3b::2) from 2001:4830:1600:3b::1, 64 hops max, 12 byte packets
1 cl-60.qas-01.us.sixxs.net 24.263 ms 11.039 ms 12.540 ms
2 gw-60.qas-01.us.sixxs.net 12.050 ms 10.853 ms 13.838 ms
3 gw-60.qas-01.us.sixxs.net 15.988 ms !A 11.219 ms !A 11.529 ms !A
Why are you routing back the ::2 address? :)
It still doesn't ping either btw.
From remote machines I see the same thing:
11 ge-0-1-0.cr1.iad1.us.occaid.net (2001:4830:ff:f150::2) 129.992 ms 128.407 ms 128.983 ms
12 2001:504:0:2:0:3:3437:1 (2001:504:0:2:0:3:3437:1) 130.994 ms 135.326 ms 130.985 ms
13 sixxs-asbnva-gw.customer.occaid.net (2001:4830:e6:7::2) 131.987 ms 129.27 ms 127.977 ms
14 cl-60.qas-01.us.sixxs.net (2001:4830:1600:3b::2) 138.994 ms 137.42 ms 136.981 ms
15 sixxs-asbnva-gw.customer.occaid.net (2001:4830:e6:7::2) 133.991 ms 132.278 ms 131.001 ms
16 sixxs-asbnva-gw.customer.occaid.net (2001:4830:e6:7::2) 131.957 ms !H 135.282 ms !H 133.985 ms !H
Reason why the PoP doesn't report a response though is because it is abusing an mtr for doing the traceroute and that ignores certain traces.
High last hop lateency...
Shadow Hawkins on Thursday, 14 September 2006 21:52:29
Thanks..
It all makes sense. I only set the routing of the ::2 address back because I wanted to check the latence and I also thought it was needed as I am still showing 100% PL in the graph on my account. I think I now know what is causing it and just need to figure out how to correct it.
I think it has to do with the AICCU client being on an XP machine behind a linksys router running dd-wrt. I have tried running the aiccu client for the dd-wrt firmware, but it does not seem to bring up the connection properly.
Now that I understand your setup better I will get it figured out.
High last hop latency... (Fixed)
Shadow Hawkins on Friday, 15 September 2006 14:46:28
For future search purposes I am adding this.
My setup is with a Linksys wrt-54g rev2 running dd-wrt v23 SP2. I have setup samb to connect to a shared drive and mounted jffs on the drive. I have installed aiccu from the dd-wrt packages (ipks install aiccu) and have configured the aiccu.conf file appropriately.
After completing all of this I still could not get aiccu to run properly. Even double checked the time on the router and made sure it matched, or was close enough, to the PoP. I finally got syslog up and running and ir was showing that the time was certainly the issue and was more than 14400 seconds out of sync. However, the issue in the end was that in the dd-wrt setup page, I had the "use local time" box checked. After unchecking this box, the aiccu program started successfully and my link came up.
So, even though I had read the info about making sure the time was in sync at both locations, it was a minro setting in the firmware itself that was causing problems.
Jeroen, thanks for you patience and time.
And the issue with the high latency was my initial setup I used off a wiki in where the remote endpoint was pointing to.
|
![[ch]](/s/countries/ch.gif)
on Wednesday, 13 September 2006 19:52:03
Posting is only allowed when you are