brian@ucsd.EDU (Brian Kantor) (07/11/89)
The U.K. became unreachable again late this afternoon, and in the process of poking around to see what I could see, I got this strange result back from traceroute. It seems to me that either a routing loop or severe congestion might cause this effect. I'm not quite sure what I'm seeing here, and I'm really rather curious! Ideas? Brian Kantor UCSD Office of Academic Computing Academic Network Operations Group brian@ucsd.edu ucsd!brian BRIAN@UCSD Script started on Mon Jul 10 19:39:33 1989 PDT [ucsd] 1 : traceroute nsfnet-relay.ac.uk traceroute to nsfnet-relay.ac.uk (128.86.8.6), 30 hops max, 40 byte packets 1 ucsd-sdsc-gw (128.54.16.9) 19 ms 16 ms 17 ms 2 129.140.77.6 (129.140.77.6) 61 ms 55 ms 74 ms 3 129.140.79.13 (129.140.79.13) 151 ms 130 ms 134 ms 4 129.140.81.15 (129.140.81.15) 183 ms 181 ms 177 ms 5 129.140.72.17 (129.140.72.17) 218 ms 223 ms 216 ms 6 * * * 7 ford-gateway.jvnc.net (128.121.54.73) 250 ms 257 ms 236 ms 8 zaphod-gateway.jvnc.net (128.121.50.72) 237 ms 615 ms 233 ms 9 fenchurch-gateway.jvnc.net (128.121.54.78) 246 ms 256 ms 239 ms 10 ford-gateway.jvnc.net (128.121.54.73) 246 ms 261 ms 241 ms 11 zaphod-gateway.jvnc.net (128.121.50.72) 242 ms 1153 ms 337 ms 12 fenchurch-gateway.jvnc.net (128.121.54.78) 271 ms 254 ms 324 ms 13 ford-gateway.jvnc.net (128.121.54.73) 276 ms 268 ms 249 ms 14 zaphod-gateway.jvnc.net (128.121.50.72) 565 ms 316 ms 299 ms 15 fenchurch-gateway.jvnc.net (128.121.54.78) 538 ms 282 ms 254 ms 16 ford-gateway.jvnc.net (128.121.54.73) 250 ms 486 ms 998 ms 17 zaphod-gateway.jvnc.net (128.121.50.72) 310 ms 612 ms 290 ms 18 fenchurch-gateway.jvnc.net (128.121.54.78) 276 ms 262 ms 299 ms 19 ford-gateway.jvnc.net (128.121.54.73) 278 ms 1627 ms 788 ms 20 zaphod-gateway.jvnc.net (128.121.50.72) 1168 ms 1523 ms 1810 ms 21 fenchurch-gateway.jvnc.net (128.121.54.78) 919 ms 846 ms 278 ms 22 ford-gateway.jvnc.net (128.121.54.73) 1962 ms 268 ms 331 ms 23 zaphod-gateway.jvnc.net (128.121.50.72) 287 ms 292 ms 305 ms 24 fenchurch-gateway.jvnc.net (128.121.54.78) 660 ms 283 ms 281 ms 25 ford-gateway.jvnc.net (128.121.54.73) 310 ms 288 ms 825 ms 26 zaphod-gateway.jvnc.net (128.121.50.72) 593 ms 545 ms 777 ms 27 fenchurch-gateway.jvnc.net (128.121.54.78) 301 ms 287 ms 312 ms 28 ford-gateway.jvnc.net (128.121.54.73) 414 ms 368 ms 313 ms 29 zaphod-gateway.jvnc.net (128.121.50.72) 688 ms 283 ms 298 ms 30 fenchurch-gateway.jvnc.net (128.121.54.78) 280 ms 296 ms 328 ms [ucsd] 2 : exit [ucsd] 3 : script done on Mon Jul 10 19:40:59 1989 (BTW: our nameserver seems to be unable to find names for the 129.140 hosts in these paths; is no one in the DNS providing those PTRs?)
hwb@MERIT.EDU (Hans-Werner Braun) (07/16/89)
> (BTW: our nameserver seems to be unable to find names for the 129.140 > hosts in these paths; is no one in the DNS providing those PTRs?) Brian: We set up some servers the other day for the 129.140 in-addr domain and asked the SRI-NIC to add them to the root server, which should happen this coming week. I would like to point out, though, that this was done in response to friendly private suggestions from some sites to us and not to requests on public mailing lists. Generally, if you see problems like the one you have mentioned in your message (the routing problem) you should be able to rely on your local regional network (SDSC? CERFNET?) to get it forwarded and fixed, even if the problem is not local to the regional network you are connecting through. The person(s) responsible for your campus network should know the people responsible for the regional network which they connect through. -- Hans-Werner
heker@JVNCA.CSC.ORG (Sergio Heker) (07/17/89)
Brian, Thank you for the detailed information on your message. Certainly traceroute is a great debugging tool. I received your message today regarding the problems to reach the JANET network, and when looking at the date of your log is Mon Jul 10. Almost a week ago. After checking our Cisco/syslog messages I didn't find any reference of any entry for that date and timestamp that can correlate to the problem that you are pointing at. When looking at SNMP information that we collect in all our Cisco gateways (we have 30 operational and 14 more that are being installed) the findings are a high number of input errors coming from the ARPANET (Zaphod and Ford are directly on the ARPANET), plus errors from one of the ethernets that connects both routers. Of course to analyze a problem that was sampled for less than two minutes, six days ago is hard if not impossible due to the dynamic nature of the network. Two comments about this. The first one is that as soon as you notice a problem like this, please notify us immediately. You can do that by phone (609) 520-2000, or by e-mail "JVNCnet-noc@jvnca.csc.org". We operate the network 24 hours/7 days a week. The second thing, is the ability to resolve names for the JANET domains, we have been discussing the posibility of backing up those domains in the US. If we do that, the we will have resolved at least the name server part. As for the routing loop, we will try to make sense of the stats that we have to better understand why the routing loop happened in the first place, but the best way to deal with it is to work on the problem when the problem is happening. Again, thank you very much for your effort to isolate the problem by yourself, and our appologies for not having been of much help to you until know. Let us help you in the future if you have a problem again. -- Sergio ----------------------------------------------------------------------------- | John von Neumann National Supercomputer Center | | Sergio Heker tel: (609) 520-2000 | | Director for Networking fax: (609) 520-1089 | | Internet: "heker@jvnca.csc.org" Bitnet: "heker@jvnc" | -----------------------------------------------------------------------------