Often the system hangs probably due to problem with Bleutooth device

Topics: 5. Support
Sep 9, 2015 at 7:29 PM
After updating the system today to 3.1.1 I still experience that my RPi system hangs.
The last log of SBFspot shows some error messages see below:

SUSyID: 131 - SN: 2120214953
Pac max phase 1: 2000W
Pac max phase 2: 2000W
Pac max phase 3: 2000W
SUSyID: 131 - SN: 2120214953
Energy Production:
    EToday: 9.600kWh
    ETotal: 4159.200kWh                                
    Operation Time: 8397.50h                    
    Feed-In Time  : 7857.06h                    
getSpotDCVoltage returned an error: -1
SUSyID: 131 - SN: 2120214953
DC Spot Data:
    String 1 Pdc:   0.348kW - Udc:   0.00V - Idc:  0.000A
    String 2 Pdc:   0.000kW - Udc:   0.00V - Idc:  0.000A
getSpotACPower returned an error: -1
getSpotACVoltage returned an error: -1
recv() returned an error: 110
getSpotACTotalPower returned an error: -1
SUSyID: 131 - SN: 2SUSyID: 131 - SN: 2120214953
Pac max phase 1: 2000W
Pac max phase 2: 2000W
Pac max phase 3: 2000W
SUSyID: 131 - SN: 2120214953
Energy Production:
    EToday: 9.600kWh                              
    ETotal: 4159.200kWh                                
    Operation Time: 8397.50h                    
    Feed-In Time  : 7857.06h                    
getSpotDCVoltage returned an error: -1
SUSyID: 131 - SN: 2120214953
DC Spot Data:
    String 1 Pdc:   0.348kW - Udc:   0.00V - Idc:  0.000A
    String 2 Pdc:   0.000kW - Udc:   0.00V - Idc:  0.000A
getSpotACPower returned an error: -1
getSpotACVoltage returned an error: -1
recv() returned an error: 110
getSpotACTotalPower returned an error: -1
SUSyID: 131 - SN: 2120214953
AC Spot Data:
    Phase 1 Pac :   0.000kW - Uac:   0.00V - Iac:  0.000A
    Phase 2 Pac :   0.000kW - Uac:   0.00V - Iac:  0.000A
    Phase 3 Pac :   0.000kW - Uac:   0.00V - Iac:  0.000A
    Total Pac   :   0.000kW
send() returned an error: 107120214953
AC Spot Data:
    Phase 1 Pac :   0.000kW - Uac:   0.00V - Iac:  0.000A
    Phase 2 Pac :   0.000kW - Uac:   0.00V - Iac:  0.000A
    Phase 3 Pac :   0.000kW - Uac:   0.00V - Iac:  0.000A
    Total Pac   :   0.000kW
send() returned an error: 107

Does the above indicate an error with the bluetooth device during getting the information?
Coordinator
Sep 9, 2015 at 8:25 PM
Does the above indicate an error with the bluetooth device during getting the information?
It looks like you have a bad connection
Error 110 (Connection timed out)
Error 107 (Transport endpoint is not connected)

Did you try already with another BT dongle?
Sep 9, 2015 at 8:51 PM
No. This one shows as:
Bus 001 Device 005: ID 0cf3:3005 Atheros Communications, Inc. AR3011 Bluetooth
which seems to use the same chip as the ASUS dongle.
Coordinator
Sep 9, 2015 at 9:05 PM
What is the distance between RPi/Inverter?
Sep 9, 2015 at 9:38 PM
I have the RPi next to the inverter. Is too close?
Coordinator
Sep 10, 2015 at 6:36 AM
Don't think it can be too close, but you can try it with a few meters distance. If still not OK, take another BT adapter.
Sep 10, 2015 at 9:31 PM
Edited Sep 10, 2015 at 9:49 PM
dear f
Quote: "After updating the system today to 3.1.1 I still experience that my RPi system hangs. "

Can you give more info about your phrase "I still experience" and "my system hangs"

-?1.1- Is your status ~same as with previous version (or did it become worse)
-?1.2- is SBFspot "hanging" ??? what's to do to overcome the hang ??
-?1.3.- can you post a "full option" log-file -> by using the command switches -d5 -v5
do this for v3.1.1.
AND if possible do post also such a log file when errors occured with previous version

-?1.4- any other device that might influence the communication ??,
rational:
  • by scanning, after executing sbfspot, each log on warnings, i found several neighbour BTsenders
  • thanks to "defensive coding" in SBFspot - it survives these noizy radio&bt messages
-?2- your log-txt did not show the BT-signal strength (receive-level) - please control the value
!2.1-if in doubt and if you have a PC with SunnyExplorer at hand - you can use SunExplorer as second info source
  • connect your dongle to your PC - take care that the dongle is at ~same position AND direction as with RPi
    (see my point 3 below)
  • in the menu's of sunny explorer you can see the two receive-levels: -at the PC and -at the inverter
  • take the time to check that these values are stable - if they fluctuate then the dongle is no good
!2.2-not least: it is the weakest element that defines the comms-quality
! note that inverters have a class1 transmitter

-?3- About the distance - with your class 2 dongle a distance - less then 2meter should fit
we do not have experience that "very close" (<30cm) is BAD - but "antenna's" and electronic close to "high power units" !!
-3.1-it is not only distance - also position of "antenna's" will have an influence (valid for RPi as for inverter)
eg: "signal" attenuation depends on the position towards the inverter: below-top left-right front-back
there is an SMA document that indicates per type of inverter the bad / moderate / better positions
SMA-doc-ref: BT_Quick-UEN110711 wireless in practice
-3.2- also for this point the setup with "SunnyExplorer" can give extra information to eleminate causes

kr wim
Sep 11, 2015 at 8:23 AM
Can you give more info about your phrase "I still experience" and "my system hangs"

-?1.1- Is your status ~same as with previous version (or did it become worse)
The same
-?1.2- is SBFspot "hanging" ??? what's to do to overcome the hang ??
The whole RPi system hangs. I have no access any more via the network; used to be Wi-Fi, but now I use Ethernet.
-?1.3.- can you post a "full option" log-file -> by using the command switches -d5 -v5
do this for v3.1.1.
AND if possible do post also such a log file when errors occured with previous version
The only error messages that are relevant in my view are the attempts by SBFspot to access the SMA. I see 10 failures.
I check the output of SBFspot, written to a temporary file, for this, and if I have these failures it removes and adds the BT driver.
-?1.4- any other device that might influence the communication ??,
rational:
  • by scanning, after executing sbfspot, each log on warnings, i found several neighbour BTsenders
  • thanks to "defensive coding" in SBFspot - it survives these noizy radio&bt messages
hcitool scan only shows the SMA device

-?2- your log-txt did not show the BT-signal strength (receive-level) - please control the value
Varies between 73% to 75%
!2.1-if in doubt and if you have a PC with SunnyExplorer at hand - you can use SunExplorer as second info source
  • connect your dongle to your PC - take care that the dongle is at ~same position AND direction as with RPi
    (see my point 3 below)
  • in the menu's of sunny explorer you can see the two receive-levels: -at the PC and -at the inverter
  • take the time to check that these values are stable - if they fluctuate then the dongle is no good
I ordered another brand for the chipset, now I have an Atheros, but ordered a Broadcom.
!2.2-not least: it is the weakest element that defines the comms-quality
! note that inverters have a class1 transmitter

-?3- About the distance - with your class 2 dongle a distance - less then 2meter should fit
we do not have experience that "very close" (<30cm) is BAD - but "antenna's" and electronic close to "high power units" !!
-3.1-it is not only distance - also position of "antenna's" will have an influence (valid for RPi as for inverter)
eg: "signal" attenuation depends on the position towards the inverter: below-top left-right front-back
there is an SMA document that indicates per type of inverter the bad / moderate / better positions
SMA-doc-ref: BT_Quick-UEN110711 wireless in practice
I have the RPi to the right of the SB 2100-TL 10 to 15 cm, which seems a good position.
-3.2- also for this point the setup with "SunnyExplorer" can give extra information to eleminate causes

kr wim
I do another check regularly, which is simply a "hcitool scan", sometime this command "hangs" for quite a while.
I have a timeout of 1 minute after which I also reload the BT driver, which mostly fixes the problem.

But still the whole system may "hang", after which only power off and power on helps.
The Wi-Fi gave me also problems, so it was unclear where the problem was.
Now that I use Ethernet via a Powerline the most likely cause is the bluetooth dongle.

Thanks for your elaborate answer.

Freek
Sep 11, 2015 at 2:07 PM
Hello,

It appears that there is a more fundamental problem with the RPi. The RPi should not hang simply due to a BT issue. Have you performed an O.S. update recently?
Assuming raspbian is the O.S. :
sudo apt-get update
sudo apt-get upgrade

73
Ron / W4MMP
On 9/11/2015 04:23, fdekruijf wrote:

From: fdekruijf

Can you give more info about your phrase "I still experience" and "my system hangs"

-?1.1- Is your status ~same as with previous version (or did it become worse)
The same
-?1.2- is SBFspot "hanging" ??? what's to do to overcome the hang ??
The whole RPi system hangs. I have no access any more via the network; used to be Wi-Fi, but now I use Ethernet.
-?1.3.- can you post a "full option" log-file -> by using the command switches -d5 -v5
do this for v3.1.1.
AND if possible do post also such a log file when errors occured with previous version
The only error messages that are relevant in my view are the attempts by SBFspot to access the SMA. I see 10 failures.
I check the output of SBFspot, written to a temporary file, for this, and if I have these failures it removes and adds the BT driver.
-?1.4- any other device that might influence the communication ??,
rational:
  • by scanning, after executing sbfspot, each log on warnings, i found several neighbour BTsenders
  • thanks to "defensive coding" in SBFspot - it survives these noizy radio&bt messages
hcitool scan only shows the SMA device

-?2- your log-txt did not show the BT-signal strength (receive-level) - please control the value
Varies between 73% to 75%
!2.1-if in doubt and if you have a PC with SunnyExplorer at hand - you can use SunExplorer as second info source
  • connect your dongle to your PC - take care that the dongle is at ~same position AND direction as with RPi
    (see my point 3 below)
  • in the menu's of sunny explorer you can see the two receive-levels: -at the PC and -at the inverter
  • take the time to check that these values are stable - if they fluctuate then the dongle is no good
I ordered another brand for the chipset, now I have an Atheros, but ordered a Broadcom.
!2.2-not least: it is the weakest element that defines the comms-quality
! note that inverters have a class1 transmitter

-?3- About the distance - with your class 2 dongle a distance - less then 2meter should fit
we do not have experience that "very close" (<30cm) is BAD - but "antenna's" and electronic close to "high power units" !!
-3.1-it is not only distance - also position of "antenna's" will have an influence (valid for RPi as for inverter)
eg: "signal" attenuation depends on the position towards the inverter: below-top left-right front-back
there is an SMA document that indicates per type of inverter the bad / moderate / better positions
SMA-doc-ref: BT_Quick-UEN110711 wireless in practice
I have the RPi to the right of the SB 2100-TL 10 to 15 cm, which seems a good position.
-3.2- also for this point the setup with "SunnyExplorer" can give extra information to eleminate causes

kr wim
I do another check regularly, which is simply a "hcitool scan", sometime this command "hangs" for quite a while.
I have a timeout of 1 minute after which I also reload the BT driver, which mostly fixes the problem.

But still the whole system may "hang", after which only power off and power on helps.
The Wi-Fi gave me also problems, so it was unclear where the problem was.
Now that I use Ethernet via a Powerline the most likely cause is the bluetooth dongle.

Thanks for your elaborate answer.

Freek

Sep 11, 2015 at 7:24 PM
rlpatton wrote:
Hello, It appears that there is a more fundamental problem with the RPi. The RPi should not hang simply due to a BT issue.
Have you performed an O.S. update recently? Assuming raspbian is the O.S. : sudo apt-get update sudo apt-get upgrade
73 Ron / W4MMP
No, I am using openSUSE Tumbleweed with kernel 4.1.6.
On the bluetooth email list I read about a bug in the firmware for the device I am using which is solved but not yet completely available downstream.
I don't know if openSUSE suffers from this problem. It uses linux-firmware-20150715git.tar.xz
I am about to receive a Broadcom based dongle.
Sep 12, 2015 at 3:11 PM
dear f since i can not attach a doc for "discussions"

". . .access the SMA. I see 10 failures. "
-> in my system this corresponds with a problem with the BTdongle, the dongle might be OK for several days and then suddenly starts hanging
the only way to deblock (for a while) is UNPLUG & REPLUG such that the complete chain (BTdriver & RPi's USB-hub & BTdongle) is reset
reinstall driver is not enough
my other BTdongle - nearly never fails - and after failure it recovers (after ending the task) !!!cross fingers!!!

"hcitool scan only . . ."
from my experience not all devices are shown by "hcitool scan"
on top of the detections via the output of SBFspot I'm using a windows based tool from NIRSoft "BlueToothView"
it renders a list of all devices
the fact that hcitool does also "hangs" is also an indication for a problem with your dongle

"signal strength 73<->75" that should be OK

"SB 2100-TL to the right 10-15cm" if your in front of your inverter and looking to the inverter then -> then to the right is OK

by the way about your timeout of 1minute
-either reduce the number of retries - see cfg file to 5 or increase the timeout to 2minutes