NFD

Is it on purpose that nfd-status doesn't refresh its output?

StatusDatasets from NFD Management are refreshed once every second.

sleep 2 shall do it.

The new log introduce 2 new errors:

1. test\_nfdc fails for scenario A & B.
2. test\_interest_aggregation fails.

I couldn't reproduce 1, but it seems to be the same problem of getting out-of-date output from nfd-status.
Regarding the second error, the produced test log is unreadable so I need some more time to figure out the exact problem.

An update:

I think that test_interest_aggregation fails because ndn-tlv-peek fails to send interests:

$ndn-tlv-peek ndn:/test-agg/ 

1425531594.411624 INFO: [UnixStreamFace]
[id=-1,local=unix:///run/nfd.sock,remote=fd://31] Creating face
1425531594.412065 INFO: [FaceTable] Added face id=277 remote=fd://31     local=unix:///run/nfd.sock
test-agADtmnfd_integration_testKEksk-142551004468ID-CERT?;F???x?,?߹?-?????????? P??C!l??O??U?=MR-0u6??????`8????Q?o7sL?
                                                                                                                    d.q?2?OU?Iү???????_?g&S?/,:    &l?D???u?p?&??A?????}2???k?p??  9j???z
?A??T???u??
       ??{???#?C??@?CD?r?go?IE????)???])???????7?u??
2E???`c\??}????)????߈|#-?K?'??d?hilata@b:/tmp/integration-tests$     1425531594.433362 INFO: [FaceTable] Removed face id=277 remote=fd://31     local=unix:///run/nfd.sock (Connection closed)

I need to continue with some more debugging, but before doing that I thought to ask whether anyone is familiar with this problem.

The mojibake in the output seems to be the Data packet, which means ndn-tlv-peek indeed sends the Interest.

#1384 has the spec for Interest aggregation scenario. Please find out which step is failing, so that I can give more specific suggestion.

I think that I found the problem. the test fails when running 2 instances of ndn-peek with the same name. The expected number of received interests is 1, but ndn-traffic-server reports on 2 incoming interests. I couldn't reproduce it manually, but it happens on every run of test_interest_aggregation.

A solution:
The script currently runs both instances of ndn-tlv-peek together. I was able to solve the error by running the first instance of ndn-tlv-peek, and then waiting for it to be completed before running the second instance. Therefore, it looks like nfd doesn't aggregate interests when arrive to close to each other. It's hard for me to point on the exact interval that causes the problem, but I wonder if that's related to the refresh period of the statusDatasets.

Thanks for the analysis in note-9. This problem is separated to #2621.

Please upload the patch for #2501 and #2502. I would treat Interest aggregation as an expected failure for now, so that the Change will be approved as long as there's no other failure.