[Nfd-dev] Emergency

Junxiao Shi shijunxiao at email.arizona.edu
Sun Aug 27 08:20:02 PDT 2017


Hi John

This is an Interest with a very long name. NFD would create many name tree
nodes, which takes long time and large amount of memory. If RIB sends a
request and doesn't get serviced in time, RIB crashes NFD.
I tested a similar scenario on my devbox two weeks ago, and it took several
minutes for NFD to process this packet before the system ran out of memory.
I'll enter a bug tonight.

Yours, Junxiao

On Aug 27, 2017 08:05, "Dehart, John" <jdd at wustl.edu> wrote:


All:

Well, having said all of that, we actually had a crash on most of the
Testbed nodes overnight.
I have located one log file with DEBUG turned on that has a FATAL error in
it.
I have it available here:

https://www.arl.wustl.edu/~jdd/REMAP_CRASH/nfd.log.10.gz

I am still going through it but one obivious thing is lines like this one:

1503809296.363506 DEBUG: [Forwarder] onIncomingInterest face=272
interest=/ndn/multicast/.../.../... <snip 15950 characters>
../.../.../.../.../.../.../.../.../…

Note that I removed 15950 characters to get it to fit into this email.
So, nfd is reporting a interest /ndn/multicast/…/…/   then with repeating
“…/“ for over 16000 characters.

The crashes all seemed to occur about 12:15AM Central time (11:15 Mountain,
10:15 Pacific, …).
Althought there are a few machines where it happened about 1 hour 15
minutes earlier.  So, perhaps there
were multiple times it happened, but I just have one in a log file.

Any thoughts?

John
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/nfd-dev/attachments/20170827/c4aabb25/attachment.html>


More information about the Nfd-dev mailing list