[Nfd-dev] Help with NFD FATAL errors

[Junxiao Shi]

Hi John

Having RibManager log entries doesn't mean the offending packet comes from
RibManager. Thus, tcpdump trace is necessary.
For full coverage, disable Unix socket listener and force local apps
(including RIB) to connect over TCP, and do tcpdump on the loopback
interface.

Yours, Junxiao

On Thu, Feb 21, 2019, 10:03 Dehart, John <jdd at wustl.edu> wrote:

> Junxiao,
>
> Here is some more context around one of them:
>
> Feb 19 23:22:49 vnetlab nfd[1100]: 1550618569.548951 INFO:
> [nfd.RibManager] Adding route /ndn/edu/ucla nexthop=270 origin=nlsr
> cost=27354
> Feb 19 23:22:49 vnetlab nfd[1100]: 1550618569.555526 INFO:
> [nfd.RibManager] Adding route /ndn/edu/ucla nexthop=260 origin=nlsr
> cost=29450
> Feb 19 23:22:49 vnetlab nfd[1100]: 1550618569.572069 INFO:
> [nfd.RibManager] Adding route /ndn/edu/ucla nexthop=274 origin=nlsr
> cost=29734
> Feb 19 23:22:49 vnetlab nfd[1100]: 1550618569.577982 INFO:
> [nfd.RibManager] Adding route /ndn/edu/ucla nexthop=261 origin=nlsr
> cost=29808
> Feb 19 23:22:49 vnetlab nfd[1100]: 1550618569.594084 FATAL: [nfd.Main]
> unrecognized TLV-TYPE 240
> Feb 19 23:22:50 vnetlab nfd[15182]: NFD version 0.6.4 starting
>
> If it is RibManager related where would the packets be coming from?
>
> John
>
> On Feb 21, 2019, at 8:34 AM, Junxiao Shi <shijunxiao at email.arizona.edu>
> wrote:
>
> Hi John
>
> It’s caused by malformed packets, plus missing error handling somewhere in
> the programs.
> Next step: run tcpdump and wait for crash, then inspect each packet
> arriving within 100ms of the crash.
>
> Yours, Junxiao
>
> On Thu, Feb 21, 2019 at 09:18 Dehart, John <jdd at wustl.edu> wrote:
>
>>
>> I have a node on the NDN Testbed that is getting NFD FATAL errors.
>> Here are a couple from last night:
>>
>> nfd.log.1:Feb 19 23:22:49 vnetlab nfd[1100]: 1550618569.594084 FATAL:
>> [nfd.Main] unrecognized TLV-TYPE 240
>>
>> ——
>>
>>
>> nfd.log.1:Feb 21 00:27:19 vnetlab nfd[16979]: 1550708838.808549 FATAL:
>> [nfd.Main] TLV-LENGTH of sub-element of type 21 exceeds TLV-VALUE boundary
>> of parent block
>>
>>
>> Any suggestions on what to look for?
>>
>> John
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/nfd-dev/attachments/20190221/8938423a/attachment.html>