GPRS stops working beyond a few meters on master

[shaddi]

When we try to do a download via GPRS using a Nexus S, things work fine as long as we're roughly in LoS of the BTS. After a point, we start getting this message in the logs when we try to load something:

Apr  2 03:13:40 endaga openbts: NOTICE 16889:17202 2015-04-02T03:13:40.3 TBF.cpp:1125:mtCancel:   03:13:40.3,110222: 03:13:40.3@@@failed tbf cause=108 TBF#256  mtMS= MS#3,TLLI=c0000001,80048002 mtDir=RLCDir::Down#012#011 channels: down=( 0:1 0:2 0:3) up=( 0:2,usf=0 0:3,usf=0)#012#011 mtState==TBFState::DataWaiting1 mtAttached=1 mtTFI=0 mtTlli=0xc0000001 size=0#012#011 mtMsgExpectedBits=0x2 mtMsgAckBits=0x0 mtExpectedAckBSN[1]=111204#012#011 mtUnAckMode=0 OnCCCH=1 mtAssignCounter=7 N3101=22#012#011 msNumDataUSFGrants=0 msAckNackUSFGrant=-1 msOldTlli=0x80048002 Pacch= idle=0#012#011 TimingError=(0.98 min=-1.68 max=2.94 avg=0.88 N=38662) RSSI=(-62 min=-63 max=-24 avg=-47.43 N=38662) CV=(41 min=0 max=63 avg=49.56 N=1120) ILev=(0 min=0 max=8 avg=2.67 N=3) RXQual=(0 min=0 max=4 avg=0.70 N=1075) SigVar=(3 min=0 max=63 avg=7.46 N=1078) ChCoding=(0 min=0 max=3 avg=0.10 N=38315)#012#011 VS=0 VA=0 TxQNum=0 stalled=0 mPrevAckSsn=0 mResendSsn=0#012#011 blocks: total=0 unique=0 grants=0  

This was the channel state immediately before the drop:

OpenBTS> gprs list  
 MS#3,TLLI=c0000001,80048002 rrmode=PacketIdle Bytes:158059up/754546down Utilization=0%  
         GMM Context: imsi=901550000000074 ptmsi=0x1 tlli=0xc0000001 state=GmmRegisteredNormal age=737 idle=1 IPs=192.168.99.1  
         TimingError=(0.98 min=-1.68 max=2.94 avg=0.88 N=38662) RSSI=(-62 min=-63 max=-24 avg=-47.43 N=38662) CV=(41 min=0 max=63 avg=49.56 N=1120) ILev=(0 min=0 max=8 avg=2.67 N=3) RXQual=(0 min=0 max=4 avg=0.70 N=1075) SigVar=(3 min=0 max=63 avg=7.46 N=1078) ChCoding=(0 min=0 max=3 avg=0.10 N=38314)  
         dataER:.3% (66912) recent:0% (0) tbfER:.19% (229)
         rrbpER:.17% (3584) recent:0% (0) ccchER:.0% (151) recent:1.0% (25) low:1.0% (2)  
 TBF#256  mtMS= MS#3,TLLI=c0000001,80048002 mtDir=RLCDir::Down  
         channels: down=( 0:1 0:2 0:3) up=( 0:2,usf=0 0:3,usf=0)  
         mtState==TBFState::DataWaiting1 mtAttached=1 mtTFI=0 mtTlli=0xc0000001 size=0  
 PDCH ARFCN=249 TN=1 FER=0%  
 PDCH ARFCN=249 TN=2 FER=11%
 PDCH ARFCN=249 TN=3 FER=12%  

This was the channel state a moment after:

OpenBTS> gprs list  
 MS#3,TLLI=c0000001,80048002 rrmode=PacketIdle Bytes:158059up/754546down Utilization=0%  
         GMM Context: imsi=901550000000074 ptmsi=0x1 tlli=0xc0000001 state=GmmRegisteredNormal age=739 idle=2 IPs=192.168.99.1  
         TimingError=(0.98 min=-1.68 max=2.94 avg=0.88 N=38662) RSSI=(-62 min=-63 max=-24 avg=-47.43 N=38662) CV=(41 min=0 max=63 avg=49.56 N=1120) ILev=(0 min=0 max=8 avg=2.67 N=3) RXQual=(0 min=0 max=4 avg=0.70 N=1075) SigVar=(3 min=0 max=63 avg=7.46 N=1078) ChCoding=(0 min=0 max=3 avg=0.10 N=38315)  
         dataER:.3% (66912) recent:0% (0) tbfER:.19% (229)  
         rrbpER:.17% (3584) recent:0% (0) ccchER:.0% (154) recent:1.0% (28) low:1.0% (2)  
 TBF#256  mtMS= MS#3,TLLI=c0000001,80048002 mtDir=RLCDir::Down  
         channels: down=( 0:1 0:2 0:3) up=( 0:2,usf=0 0:3,usf=0)  
         mtState==TBFState::Dead mtAttached=1 mtTFI=0 mtTlli=0xc0000001 size=0
 PDCH ARFCN=249 TN=1 FER=0%  
 PDCH ARFCN=249 TN=2 FER=11%  
 PDCH ARFCN=249 TN=3 FER=12%  

[shaddi]

We also have some mtCancels ending with cause=3101:

Apr  2 03:11:48 endaga openbts: NOTICE 16889:17202 2015-04-02T03:11:48.5 TBF.cpp:1125:mtCancel:   03:11:48.5,104632: 03:11:48.5@@@failed tbf cause=3101 TBF#210  mtMS= MS#3,TLLI=c0000001,80048002
 mtDir=RLCDir::Up#012#011 channels: down=( 0:1 0:2 0:3) up=( 0:2,usf=1 0:3,usf=1)#012#011 mtState==TBFState::DataTransmit mtAttached=1 mtTFI=18 mtTlli=0xc0000001#012#011 mtMsgExpectedBits=0x20 m
tMsgAckBits=0x2 mtExpectedAckBSN[5]=104630#012#011 mtUnAckMode=0 OnCCCH=0 mtAssignCounter=1 N3101=21#012#011 msNumDataUSFGrants=1723 msAckNackUSFGrant=1715 msOldTlli=0x80048002 Pacch= idle=0#012
#011 TimingError=(1.15 min=-1.68 max=2.94 avg=0.88 N=36226) RSSI=(-61 min=-63 max=-24 avg=-46.82 N=36226) CV=(56 min=0 max=63 avg=50.01 N=1035) ILev=(0 min=0 max=8 avg=2.67 N=3) RXQual=(1 min=0 
max=4 avg=0.69 N=994) SigVar=(2 min=0 max=63 avg=7.33 N=997) ChCoding=(0 min=0 max=3 avg=0.10 N=35940)#012#011 VR=121 VQ=121 stalled=0 mNumUpBlocksSinceAckNack=0 mtUpState=1#012#011 blocks: tota
l=696 unique=502 grants=1723

[kheimerl]

Need a version number on this.

On Wed, Apr 1, 2015 at 8:37 PM, Shaddi Hasan [email protected]
wrote:

We also have some mtCancels ending with cause=3101:

Apr 2 03:11:48 endaga openbts: NOTICE 16889:17202 2015-04-02T03:11:48.5 TBF.cpp:1125:mtCancel: 03:11:48.5,104632: 03:11:48.5@@@failed tbf cause=3101 TBF#210 mtMS= MS#3,TLLI=c0000001,80048002
mtDir=RLCDir::Up#012#011 channels: down=( 0:1 0:2 0:3) up=( 0:2,usf=1 0:3,usf=1)#12#011 mtState==TBFState::DataTransmit mtAttached=1 mtTFI=18 mtTlli=0xc0000001#012#011 mtMsgExpectedBits=0x20 m
tMsgAckBits=0x2 mtExpectedAckBSN[5]=104630#012#011 mtUnAckMode=0 OnCCCH=0 mtAssignCounter=1 N3101=21#012#011 msNumDataUSFGrants=1723 msAckNackUSFGrant=1715 msOldTlli=0x80048002 Pacch= idle=0#012
#11 TimingError=(1.15 min=-1.68 max=2.94 avg=0.88 N=36226) RSSI=(-61 min=-63 max=-24 avg=-46.82 N=36226) CV=(56 min=0 max=63 avg=50.01 N=1035) ILev=(0 min=0 max=8 avg=2.67 N=3) RXQual=(1 min=0
max=4 avg=0.69 N=994) SigVar=(2 min=0 max=63 avg=7.33 N=997) ChCoding=(0 min=0 max=3 avg=0.10 N=35940)#12#011 VR=121 VQ=121 stalled=0 mNumUpBlocksSinceAckNack=0 mtUpState=1#012#011 blocks: tota
l=696 unique=502 grants=1723

—
Reply to this email directly or view it on GitHub
#8 (comment).

[shaddi]

endaga:master.

[kheimerl]

Ok, so 5.0 isn't that hot either it seems.

On Wed, Apr 1, 2015 at 8:40 PM, Shaddi Hasan [email protected]
wrote:

endaga:master.

—
Reply to this email directly or view it on GitHub
#8 (comment).

[shaddi]

The causes are outlined here; 108 is Non-responsive and 3103 refers to the N3101 timer expiring. Both of these are uplink issues.

[kheimerl]

Power issue, resolved via config.