connectivity issue Internode to Linode

I am totally confused by a networking issue I have. I have 3 servers with Linode and am having problems connecting to 2 of them. I have contaced Linode and they insist everything is OK from their end. So the symptons. I cannot connect to nicbaxter.com.au using ssh, web, mail for the last 3 hours or so. I haven't knowingly made any changes to the server. I ran a mtr: [nic@albert ~]$ mtr --report nicbaxter.com.au HOST: albert.nicbaxter.home Loss% Snt Last Avg Best Wrst StDev 1.|-- gateway.nicbaxter.home 0.0% 10 1.3 1.3 1.1 1.9 0.2 2.|-- lns20.mel6.internode.on.n 0.0% 10 19.2 21.5 19.2 29.3 2.9 3.|-- te2-2.cor2.mel6.internode 0.0% 10 20.4 20.7 19.6 22.2 0.7 4.|-- gi8-1-110.bdr1.mel6.inter 0.0% 10 187.3 188.1 187.2 189.5 0.8 5.|-- te2-0-0.bdr1.cbr1.interno 0.0% 10 289.6 237.5 186.9 359.9 67.2 6.|-- te6-0-0.bdr1.syd4.interno 0.0% 10 187.7 188.0 187.0 189.8 0.8 7.|-- te0-0-0.bdr1.syd7.interno 0.0% 10 187.1 200.6 187.1 311.2 38.9 8.|-- po3-0-0.bdr1.sjc2.interno 0.0% 10 189.2 188.0 186.5 189.2 1.1 9.|-- te1-7.bbr01.eq01.sjc01.ne 0.0% 10 188.1 188.0 187.5 188.9 0.5 10.|-- ae7.bbr02.eq01.sjc02.netw 0.0% 10 188.1 188.3 187.6 189.2 0.6 11.|-- ae0.bbr02.cs01.lax01.netw 20.0% 10 193.5 194.4 193.5 195.7 0.8 12.|-- ae7.bbr01.cs01.lax01.netw 0.0% 10 194.6 194.4 193.8 196.2 0.7 13.|-- ae19.bbr01.eq01.dal03.net 10.0% 10 221.4 222.1 221.4 222.5 0.4 14.|-- po31.dsr01.dllstx3.networ 0.0% 10 224.7 224.3 223.6 225.4 0.5 15.|-- te2-4.dsr01.dllstx2.netwo 0.0% 10 225.1 225.3 224.7 226.3 0.5 16.|-- po1.car01.dllstx2.network 0.0% 10 224.3 231.4 224.1 291.5 21.1 17.|-- router1-dal.linode.com 0.0% 10 224.7 226.0 223.7 237.0 3.9 18.|-- ??? 100.0 10 0.0 0.0 0.0 0.0 0.0 [nic@albert ~]$ and tcpdump on the server 02:14:37.763601 IP (tos 0x0, ttl 64, id 44588, offset 0, flags [none], proto ICMP (1), length 64) nicbaxter.com.au > ppp118-209-136-214.lns20.mel6.internode.on.net: ICMP echo reply, id 61030, seq 61313, length 44 02:14:38.518998 IP (tos 0x0, ttl 1, id 24801, offset 0, flags [none], proto ICMP (1), length 64) ppp118-209-136-214.lns20.mel6.internode.on.net > nicbaxter.com.au: ICMP echo request, id 61030, seq 386, length 44 02:14:38.519043 IP (tos 0x0, ttl 64, id 44589, offset 0, flags [none], proto ICMP (1), length 64) nicbaxter.com.au > ppp118-209-136-214.lns20.mel6.internode.on.net: ICMP echo reply, id 61030, seq 386, length 44 02:14:38.560710 IP (tos 0x0, ttl 2, id 24802, offset 0, flags [none], proto ICMP (1), length 64) ppp118-209-136-214.lns20.mel6.internode.on.net > nicbaxter.com.au: ICMP echo request, id 61030, seq 642, length 44 I can connect using tor. Also Internode can connect without an issue. I have another server nicbaxter.com also with Linode and everything is fine. Any ideas? Cheers Nic

Hi Nic. Quick test: I can connect both your .com and com.au sites using IE from a WinXP Pc no problem from here - via iinet. All the best. John. ------------------------------------------------------------ John Simmons - Director Genesis Networks Phone: (03) 9754 6479 Email: jss@genesis.net.au Fax: (03) 9754 6514 Web: www.genesis.net.au Postal: 3 Gully Crescent Belgrave Victoria 3160 Australia ------------------------------------------------------------

Nic Baxter <nic@nicbaxter.com.au> wrote:
I am totally confused by a networking issue I have. I have 3 servers with Linode and am having problems connecting to 2 of them. I have contaced Linode and they insist everything is OK from their end.
Friends who run Linode instances reported having serious connectivity problems with them over the last week. Thus you are not alone in this respect. It seems the packets are reaching the destination, the echo replies are sent out but never arrive. How many hops are there between the last "good" router at the Linode end and the Linode host? If there is at least one, then that could be the problem. Have you tried a reverse traceroute from the server back to your machine? Perhaps the paths in the "forward" and "reverse" directions are different - I've seen that before. This doesn't help to solve the problem, but if you can work out where it's happening then you'll have data to present to Linode indicative of a network problem. Or maybe it's their upstream, in which case they are the ones who will presumably have to deal with it by contacting the affected "upstream" provider. I think they can use BGP to re-route the traffice if they must. Try to work out which host/network is having the packet loss.

On 03/07/12 15:22, Jason White wrote:
Nic Baxter<nic@nicbaxter.com.au> wrote:
I am totally confused by a networking issue I have. I have 3 servers with Linode and am having problems connecting to 2 of them. I have contaced Linode and they insist everything is OK from their end. Friends who run Linode instances reported having serious connectivity problems with them over the last week. Thus you are not alone in this respect.
It seems the packets are reaching the destination, the echo replies are sent out but never arrive. How many hops are there between the last "good" router at the Linode end and the Linode host? If there is at least one, then that could be the problem. Have you tried a reverse traceroute from the server back to your machine? Perhaps the paths in the "forward" and "reverse" directions are different - I've seen that before. Thanks everyone for the suggestions. It is puzzling. The forward and reverse traceroutes are the same so not much help there. Internode cannot see any problem with their connections. My Internode connection from Woodend uses Telstra as the wholesaler and three of my clients who had problems connecting use Telstra as their ISP. Others can connect. Who knows? Anyway I do enjoy a bit of gratuitous Telstra bashing so I will blame them. However that doesn't fix the problem so I decided to move the server to a Mammoth vps and so far so good. My remaining Linode instance is still working well, touch wood.
Cheers Nic

On Tue, 03 Jul 2012 13:23:53 +1000 Nic Baxter <nic@nicbaxter.com.au> wrote:
I cannot connect to nicbaxter.com.au using ssh, web, mail for the last 3 hours or so.
I haven't noticed any issues of late connecting to my Linode from Internode.
17.|-- router1-dal.linode.com 0.0% 10 224.7 226.0 223.7 237.0 3.9 18.|-- ??? 100.0 10 0.0 0.0 0.0
I guess that's Texas? Mine's in Freemont (or whatever, the last hop is router1-fmt.linode.com) so perhaps that's the issue. The fact that you get to a linode.com address before loosing packets would mean the issue is on the Linode side I'd have thought. Regardless, I can connect to your nicbaxter.com.au from my Internode connection here just fine. Regards, -- Todd Harbour ~o~

Todd Harbour <luv_list@datapax.com.au> wrote:
On Tue, 03 Jul 2012 13:23:53 +1000 Nic Baxter <nic@nicbaxter.com.au> wrote:
I cannot connect to nicbaxter.com.au using ssh, web, mail for the last 3 hours or so.
I haven't noticed any issues of late connecting to my Linode from Internode.
I can ping nicbaxter.com.au from my Internode connection without any problems. One suggestion: enable IPv6 at both ends, which may take a different path. Internode and Linode both support IPv6 at this point with no tunnels involved.
participants (4)
-
Jason White
-
John Simmons
-
Nic Baxter
-
Todd Harbour