-
Notifications
You must be signed in to change notification settings - Fork 1.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Negative values in loss report with UDP #914
Comments
I agree those values shouldn't be negative! At the moment I don't have any insight into how they got that way...it doesn't look like you are doing anything particularly unusual that might cause this problem. |
Hi, I'm also seeing similar behavior where negative lost packets are being reported. Like the original poster, it happens only sporadically. I've seen it present in two different ways. Setup:
Case 1:
Case 2:
|
Hi guys, have you overcome this problem?
(reformatted for clarity) |
Hi,
The -2 lost packets means that that the two packets lost in interval 30.00-31.00 were received in interval 31.00-32.00. The real total packets received in the interval is therefore "Total"-"Lost" (e.g. 173-(-2)=175 packets were received in interval 31.00-32.00) For the original case posted, data is missing regarding interval 20.000105-21.000194. As in interval 21.000194-22.000068, 146-(-111)=257 packets were received, it seems that there was a serious network buffering and reorder in interval 20.000105-21.000194, which caused a packet to arrive before 111 preceding packets, and the 111 packets arrived only during the following interval. It seems that the main issue is that sometimes the first packet is received with garbled packet count or is received truncated, so the packet counter is garbled. E.g. in:
probably there was a problem with the first packet assumed to be received, so its garbled packet count was 393539. Practically, 393539-393365=174 packets were received in the first interval. All following intervals shows negative number of lost packets, since each received packet is considered recovery of lost packet from the first interval, until more than 393539 packets will be received. Later I will send a pull request with suggested changes to guard against using garbled packet counter. |
Believed fixed by #1260 |
# Context
# Bug Report
Got a report with negative values in the lost_packets and losr_percent counters at server side logfile.
Expected Behavior:
{ "streams": [{ "socket": 6, "start": 19.000102, "end": 20.000105, "seconds": 1.0000029802322388, "bytes": 166520, "bits_per_second": 1332156.0298656528, "jitter_ms": 2.8512847190910904, "lost_packets": 0, "packets": 115, "lost_percent": 0, "omitted": false, "sender": false }], "sum": { "start": 19.000102, "end": 20.000105, "seconds": 1.0000029802322388, "bytes": 166520, "bits_per_second": 1332156.0298656528, "jitter_ms": 2.8512847190910904, "lost_packets": 0, "packets": 115, "lost_percent": 0, "omitted": false, "sender": false }
The above stream report just appeared as expected in the same logfile than the next stream report (the one with negative values)
Actual Behavior
{ "streams": [{ "socket": 6, "start": 21.000194, "end": 22.000068, "seconds": 0.99987399578094482, "bytes": 372136, "bits_per_second": 2977463.1729218694, "jitter_ms": 2110.7105786449633, "lost_packets": -111, "packets": 146, "lost_percent": -76.027397260273972, "omitted": false, "sender": false }], "sum": { "start": 21.000194, "end": 22.000068, "seconds": 0.99987399578094482, "bytes": 372136, "bits_per_second": 2977463.1729218694, "jitter_ms": 2110.7105786449633, "lost_packets": -111, "packets": 146, "lost_percent": -76.027397260273972, "omitted": false, "sender": false }
Steps to Reproduce
command server side: iperf3 -s -p 11002 -1 -J --logfile serverOut_11_to_2.json
command client side: iperf3 -c 10.0.0.2 -p 11002 -u -b 1675.289k -t 40 -J --logfile clientOut_11_to_2.json
I have reproduced several of such a test between different hosts and it just happens randomly so it is hard to reproduce.
I am employing Mininet 2.2.1 in the VM to deploy a network with 23 nodes and 36 links.
I am generating specfic amounts of traffic from hosts in my network regardind a traffic matrix (that is whay I set a value -b in the client side). I just realized about this issue when analyzing all the logfiles. What could I be doing wrong? I will appreaciate your help. Thanks!
The text was updated successfully, but these errors were encountered: