Disconnection with a Link Capacity test

AP PMP 450 m 15.1.3
PMP 450i 15.1.3

Rebooted AP and SM but the result it's the same.

When i try a link capacity test from the SM, the SM disconnects without providing the values.

SM doesn't see others AP. it had never happened to me.

Current Results Status

Error: Did not receive results from the remote device.

Downlink Status

Receive Power :
-68.9 dBm ( -71.0 dBm V / -73.0 dBm H )
Signal Strength Ratio :
2.0dB V - H
Signal to Noise Ratio :
22 V / 21 H dB

Subscriber Module Stats

Session Status :
REGISTERED VC 31 Rate 8X/6X MIMO-B
Session Uptime :
00:01:22

Ok, SM sees another AP, Receive Power -81. Here Link Capacity Test it's Ok.

The others SM with the initial AP (that of my first post) with the others SM (450 or 450 i) no problems with LCT.

Hi Tuvix_IT,

This is a strange problem. A few questions.

Does the LCT work when started at the AP ?

Is there anything different on this SM (e.g. QoS, CIR, MIR).

Is there a large traffic amount running to this SM?

Could there be high priority traffic that is stopping the report coming back in time.

Does the SM actually drop the link to AP or just not report the result ?

If the link drops does the SM also reboot ?

Are there problems with normal traffic you are trying to check with LCP ?

best regards,

Andrew.

From the AP the same Error: 
"Error: Did not receive results from the remote device."
and the SM, after the error, it's disconnected! (no reboot)
No, there isn't large traffic, it is almost unused. I tried these tests even at night but the result is the same.

Frame utilization it's low in the AP:

Frame Utilization
Downlink :
23 %
Uplink :
11 %
Bandwidth Requests :
99 %
 
SM it's preconfigured, no differences.

The result of LCT it's never showed. 

With CNMaestro a perpetual "In progress..:" and the SM disconnetced
 

Debug with CNMaestro:

Logs
In Progress...

started logging
2018-03-05T21:37:18Z 0.0.0.0 CANOPY PMP 450i: - - - state_table_event; table=SMSession; Event=SMSESMSG; MsgType=PAREQTO; NewState=REGISTERED; Flag=0;
2018-03-05T21:37:18Z 0.0.0.0 CANOPY PMP 450i: - - - state_table_event; table=SMSession; Event=SMSESMSG; MsgType=PASRSP; NewState=REGISTERED; Flag=20;
2018-03-05T21:37:19Z 0.0.0.0 CANOPY PMP 450i: - - - state_table_event; table=DiagState; Event=Loc Start; NewState=WaitLocTxCmplt; Flag=6;
2018-03-05T21:37:19Z 0.0.0.0 CANOPY PMP 450i: - - - event; link_test; user=none; Link capacity test started;
2018-03-05T21:37:19Z 0.0.0.0 CANOPY PMP 450i: - - - state_table_event; table=DiagState; Event=Loc TxCmplt; NewState=WaitRmtHndshk; Flag=6;
Link Test Mode :  extrapolated link test


This is the only test without problem but i see that the others test works once every 20 attempts, you can do a bit of tight tests and then the antenna disconnects.

Strange but not a serious problem.

EDIT: seem that 1714 bytes like packet lenght (default) are the problem.
if i set 1650 or less no problem :/

1 Like

EDIT: seem that 1714 bytes like packet lenght (default) are the problem.
if i set 1650 or less no problem :/


Thank you, changing the packet length was one of the other suggestions I was going to make :)

As you say 1714 is the default and will have been used in all our testing so I don't know why it would cause you a problem. The max MTU for ethernet is 1500 I think, so unlikely to cause you any problems if you are OK with 1650.

This (MTU, frame sizes, etc) isn't my area of expertise but I will raise this in our bug tracking system and see if anyone else has any ideas.

best regards

Andy.

I've raised a bug on this.

To help see if we can reproduce it can you provide the usual files after observing the fault

field_diags.cgi from the AP

and

engineering.cgi from the SM

thanks,

Andy.

Send them ;)

1 Like

got them.

Grazie.

Andy.

Hi Tuvix_IT,

I looked through the SM logs you provided.  The logs show that 42 SM initiated link test were attempted, with 36 succesful test and 6 failures.  The 6 failures resulted in a session drop.  Was the packet length set to 1714 for all 42 tests?

 

Thanks,

Anthony

Hi, no, in the last tests I tried to set a lower Packet Length and there was no MS disconnection.

Now i tried

Packet Lenght 1650: ok

Packet Lenght 1670: failed

Packet Lenght 1670: failed

Packet Lenght 1640: failed

Packet Lenght 1655: failed

But it's possibile that there is an interference? The Average Corrected Bit Errors in uplink many times it's 0 or 1, now it's 3 or 4

Stats for LUID: 29   Test Duration: 2   Pkt Length: 1655   Test Direction Bi-Directional

Link Test without Bridging

VC Downlink Uplink Aggregate Packet Transmit Packet ReceiveActual Actual

45 43.94 Mbps 26.29 Mbps 70.23 Mbps,  5186 pps 3898 (1949 pps) 6474(3237 pps)



Efficiency

Downlink UplinkEfficiency Fragmentscount Signal toNoise Ratio Efficiency Fragmentscount Signal toNoise RatioActual Expected Actual Expected

97% 177766 171652 28 dB V
26 dB H
88% 116329 103012 27 dB V
29 dB H


Link Quality 
Downlink

RF Path Modulation Fragments ModulationPercentage Average CorrectedBit Errors

V QPSK 29122 33% 1.167
V 16-QAM 29123 33% 1.197
V 64-QAM 29084 33% 1.373
H QPSK 27194 33% 4.016
H 16-QAM 27192 33% 4.109
H 64-QAM 26609 33% 4.417

Uplink

RF Path Modulation Fragments ModulationPercentage Average CorrectedBit Errors

V 64-QAM 48880 100% 3.746
V 256-QAM 4 <1% 9.000
H 64-QAM 56805 100% 1.877
H 256-QAM 1 <1% 9.000


Link Test ran on 00:20:05 03/15/2018 UTC 

Currently transmitting at:

VC 45 Rate 8X/6X MIMO-B


Others test with 1650 that are ok:

----------------------------------------------
uplink

RF Path Modulation Fragments ModulationPercentage Average CorrectedBit Errors

V 64-QAM 53049 100% 3.001
V 256-QAM 16 <1% 8.062
H 64-QAM 59593 100% 0.897
H 256-QAM 102 <1% 6.647

 

uplink

RF Path Modulation Fragments ModulationPercentage Average CorrectedBit Errors

V 64-QAM 53049 100% 3.001
V 256-QAM 16 <1% 8.062
H 64-QAM 59593 100% 0.897
H 256-QAM 102 <1% 6.647

Uplink

RF Path Modulation Fragments ModulationPercentage Average CorrectedBit Errors

V 64-QAM 53703 100% 2.948
V 256-QAM 16 <1% 8.187
H 64-QAM 60122 100% 0.796
H 256-QAM 96 <1% 6.812


Link Test ran on 00:23:41 03/15/2018 UTC 

Uplink

RF Path Modulation Fragments ModulationPercentage Average CorrectedBit Errors

V 64-QAM 54138 100% 2.846
V 256-QAM 21 <1% 6.142
H 64-QAM 59623 100% 0.756
H 256-QAM 92 <1% 6.130


Link Test ran on 00:23:55 03/15/2018 UTC 

1 Like

Thank you Tuvix_IT,

The 1650 bytes good is a definite clue, 64 bytes less than the 1714 max.

There is a relationship between the over the air fragments being dropped and the QinQ & VLAN extra bytes needed. We are managing to reproduce variations of the problem in the lab, so hopefully a fix will be found soon.

best regards,

Andy.