Joel,
I have played with sendsms and the results seem hit and miss. Sometimes OpenBTS core dumps. After a power cycle of my Nokia MS the first SMS is received OK e.g.
sendsms 234100516180928 SabreTek Hello
with a Wireshark trace showing:
GSM SMS 87 I, N(R)=0, N(S)=1(DTAP) (SMS) CP-DATA (RP) RP-DATA (Network to MS)
The next SMS does not get received and I see the following error in the OpenBTS logs:
GSML2LAPDm.cpp:925:sendMultiframeData: obj: 0x8b14078 attempt to send DATA on released LAPm channel
I would up the log level on GSM L2 and continue investigations here.
config Log.Level.GSML2LAPDm.cpp DEBUG
PS: You can clear the TMSI table using the following SQL in a file named "TMSI.sql":
BEGIN TRANSACTION; DELETE FROM TMSI_TABLE; COMMIT;
Then:
sqlite3 -init TMSI.sql /etc/OpenBTS/OpenBTS.db ".quit"
Regards Craig
Joel,
It could be because of bad Rx settings. What are your 'rxgain' and MS RSSI Target?
On Sun, Mar 24, 2013 at 7:06 PM, Joel Yabut joel.yabut@gmail.com wrote:
Sorry, its because our ettus are having no issues with this sms wherein we have this issue with the umtrx with the new build of open bts.
Sent from my Ipad
On Mar 24, 2013, at 2:23 PM, Alexander Chemeris alexander.chemeris@gmail.com wrote:
Joel,
These questions are actually out of topic for this mailing list. I recommend you to report this to the OpenBTS mailing list and attach relevant logs.
But my comments are below anyway:
On Sun, Mar 24, 2013 at 12:14 AM, Joel Yabut joel.yabut@gmail.com wrote:
Anyone having issues with sendsms using cli commands? I have been sending test 1 to 10 and i only received 5 out of 10.
I haven't used this command extensively, but it shouldn't be much different from sending an SMS through smqueue.
Also, it does not always send a welcome message to my phones when camping.
It sends welcome message only to phones which a cell sees for the first time, i.e. the ones which are not present at the TMSI table. Try clearing TMSI table before your experiments.
-- Regards, Alexander Chemeris. CEO, Fairwaves LLC / ООО УмРадио http://fairwaves.ru
Joel,
If you use an external LNA you could be saturating the input of LMS. It's hard to tell without checking your exact configuration with the proper tools. Anyway, try decreasing rxgain to minimum and see whether it works better.
On Sun, Mar 24, 2013 at 8:51 PM, Joel Yabut joel.yabut@gmail.com wrote:
Rxgain is at 20. I have an external lna installed. Noise is at -23 and target is at -15
Sent from my Ipad
On Mar 24, 2013, at 11:25 PM, Alexander Chemeris alexander.chemeris@gmail.com wrote:
Joel,
It could be because of bad Rx settings. What are your 'rxgain' and MS RSSI Target?
On Sun, Mar 24, 2013 at 7:06 PM, Joel Yabut joel.yabut@gmail.com wrote:
Sorry, its because our ettus are having no issues with this sms wherein we have this issue with the umtrx with the new build of open bts.
Sent from my Ipad
On Mar 24, 2013, at 2:23 PM, Alexander Chemeris alexander.chemeris@gmail.com wrote:
Joel,
These questions are actually out of topic for this mailing list. I recommend you to report this to the OpenBTS mailing list and attach relevant logs.
But my comments are below anyway:
On Sun, Mar 24, 2013 at 12:14 AM, Joel Yabut joel.yabut@gmail.com wrote:
Anyone having issues with sendsms using cli commands? I have been sending test 1 to 10 and i only received 5 out of 10.
I haven't used this command extensively, but it shouldn't be much different from sending an SMS through smqueue.
Also, it does not always send a welcome message to my phones when camping.
It sends welcome message only to phones which a cell sees for the first time, i.e. the ones which are not present at the TMSI table. Try clearing TMSI table before your experiments.
-- Regards, Alexander Chemeris. CEO, Fairwaves LLC / ООО УмРадио http://fairwaves.ru
-- Regards, Alexander Chemeris. CEO, Fairwaves LLC / ООО УмРадио http://fairwaves.ru
Joel,
On Mon, Mar 25, 2013 at 7:30 PM, Joel Yabut joel.yabut@gmail.com wrote:
I have a spectrum analyzer let me know what values to check.
Unfortunately, I don't have a ready answer for this on hands. May be someone with RF skills from the community could help you or extend our manual on how to tune parameters for a specific hardware.
On our side we could only offer you a paid support on this issue, as it includes remote debugging with hardware which is unknown to us (LNA, antenna, etc). Funds from the paid support are going to support development of UmSEL and other RF parts which are optimized to work with UmTRX.
Also, do you think this is hardware related and not software related?
Could be hardware setup related or software configuration related.
Just for the reference, we have a customer who achieve coverage of
1.2km with a 2W booster and things work fine for him.
-- Regards, Alexander Chemeris. CEO, Fairwaves LLC / ООО УмРадио http://fairwaves.ru
Joel,
On Mon, Mar 25, 2013 at 7:48 PM, Alexander Chemeris alexander.chemeris@gmail.com wrote:
Also, do you think this is hardware related and not software related?
Could be hardware setup related or software configuration related.
Good news - if you check out the latest 'umtrx' branch, the issue should be gone.
Ivan looked into the issue and was able to reproduce it on one of our test benches under some conditions. The real issue is that with a slow CPU or some heavy things like X11 present, processing which should be real-time in OpenBTS is not real-time anymore. And this creates issues like this at random points of time. The fix which Ivan committed today hides the problem for SMS, but doesn't fix the underlying problem. Ivan will try to look deeper in the coming days.
Your observation that everything worked well with USRP N and doesn't work with UmTRX is quite strange, because the CPU load with UmTRX is less than with USRP N. And we sue the same code except that we don't need resampling with UmTRX. The issue actually comes from the original OpenBTS code and was discussed some time ago at the OpenBTS mailing list (search for "Thoughts on the filler table and why it needs to go" thread). So I'd guess there are other changes which introduced this issue. May be you've changed your motherboard meantime, or added X11 or did something else which induced non-realtime behavior.
-- Regards, Alexander Chemeris. CEO, Fairwaves LLC / ООО УмРадио http://fairwaves.ru
-
Alexander Chemeris -
Craig Reading