Open 3.2.2 Crashes on Ubuntu 12.04 64 bit

[cagney]

I also get a new instance of the alsa plugin for each time the ais alarm is triggered. It does not happen when just testing the alarm sound.
I will enter this in the tracker.

Thomas

[redog]

Thanks for all the help so far.

I mulled over Dave's suggestion that "It is unusual for a user-space program to crash the OS in linux."
So I searched the web and found other instances of 12.04 64 bit crashing, main one I found related to chrome.
Suggestions were to upgrade kernel.

I just upgraded my kernel from 3.2 to 3.4.
So far 2 days with no crash.

Will keep you posted.
redog

[redog]

Originally Posted by redog

Thanks for all the help so far.

I mulled over Dave's suggestion that "It is unusual for a user-space program to crash the OS in linux."
So I searched the web and found other instances of 12.04 64 bit crashing, main one I found related to chrome.
Suggestions were to upgrade kernel.

I just upgraded my kernel from 3.2 to 3.4.
So far 2 days with no crash.

Will keep you posted.
redog

Took the boat out for a live try.
No lock ups with the kernel upgrade.

Dave
The stuttering track is still present!
So much so I wouldn't leave the wheel even for a short time while under AP.

"Set up"
GPSD priority 9 GPS antenna
UDP priority 6 AIS from backup computer
USB1 Apilot out only priority 5
Filter NMEA 5 sec
LAT LON in bottom of screen very erratic.

It got that bad at times it turned the boat around, probably trying to get to that fictitious waypoint.

Next removed UDP and tested.(1 source only)
Lets just say it went better e.g. no big U turns from AP.
LAT LON steadier but still jumpy.(2screen shots attached)
Note both screenshots are at full zoom.

Just had a thoughty!!
Should have tried AIS only on UDP, next time.

redog

[bdbcat]

redog...

Glad the kernel upgrade helped with the crashing issue. I wonder what the root cause was? We will probably never know.

On the stuttering tracks:
We really need to study an extended look at your NMEA data stream in detail. We can do this two ways"

1. Screen shots (several) of Connections->Debug NMEA window.
.or.
2. A recording of your runtime session using the VDR PlugIn. Recording should be long enough to encompass at least some of the stutters you show onscreen.

I suspect what is happening is that there are two or more sources of lat/lon in your stream, and they differ in decimal precision (2, 3, or 4 digits after the decimal point). I would look for anything other than 4 decimals, and filter that message type away. Or, correct it at source if possible. For instance, my Leica GPS allows setting of decimal precision in NMEA output.

Or, maybe something else....

Thanks
Dave

[redog]

Dave
Just got a strange one??????

With computers connected 1 source into main and second comp receiving via UDP.

Disconnect LAN cable.

2nd comp goes to red no source

Plug in house internet cable to 2nd comp and status goes green.
Showing the IP address of 1st comp in debug window.

Will work on getting some files for you for previous post.
Thanks
redog

[redog]

Originally Posted by redog

Dave
Just got a strange one??????

With computers connected 1 source into main and second comp receiving via UDP.

Disconnect LAN cable.

2nd comp goes to red no source

Plug in house internet cable to 2nd comp and status goes green.
Showing the IP address of 1st comp in debug window.

Will work on getting some files for you for previous post.
Thanks
redog

My bad....

Wireless was on 1st computer putting UDP through modem and thus to 2nd comp.

False alarm
redog

[redog]

Originally Posted by bdbcat

redog...

Glad the kernel upgrade helped with the crashing issue. I wonder what the root cause was? We will probably never know.

On the stuttering tracks:
We really need to study an extended look at your NMEA data stream in detail. We can do this two ways"

1. Screen shots (several) of Connections->Debug NMEA window.
.or.
2. A recording of your runtime session using the VDR PlugIn. Recording should be long enough to encompass at least some of the stutters you show onscreen.

I suspect what is happening is that there are two or more sources of lat/lon in your stream, and they differ in decimal precision (2, 3, or 4 digits after the decimal point). I would look for anything other than 4 decimals, and filter that message type away. Or, correct it at source if possible. For instance, my Leica GPS allows setting of decimal precision in NMEA output.

Or, maybe something else....

Thanks
Dave

Dave
I thought using priority settings would eliminate the second stream and only use it in the event of #1 stream failing!!!!

No I don't have the luxury of user settings at source, its just a puck.

redog

[nohal]

redog...
Yes it should. We just need to make sure it really does or if/why it does not...

Pavel

[redog]

Pavel
Thanks for clearing that up.
VDR recordings will follow next trip.
Sit tight could be a month though!
redog

[redog]

Dave/Pavel
Firstly after the success of the kernel upgrade I went 1 step further and upgraded to 13.04.

This version feels a lot more stable right from the start!!

I have 14 VDR files in total, each time I changed set up I started a new recording.

I had 1 spot where the boat turned around again under AP and O routing.(I saved a MOB waypoint)

Is there a way to put the VDR info beside the track info for the same period??

Or does the waypoint info include a date stamp to then look at the VDR data??

As I said I changed set up to try to find fault.
Starting with:
Comp1 GPS antenna USB using GPSD priority 9 all available sentences
USB output to AP and VHF
UDP receiving all available sentences from Comp 2

Comp2 AIS input USB using GPSD priority 9 all available sentences
UDP output to Comp1 all available.

I just tried to upload first 2 files, no go invalid file.
What is best format / process for you guys to view?

A strange thing on the return journey along the same stuttering track it did the same stuttering in the same area.

Attached is left track going right track coming home.

redog

[bdbcat]

redog...

You may rename the VDR files as {something}.pdf, and they should upload here OK.

Thanks
Dave

[redog]

Dave
Attached are the 2 VDR files relevant to the set up in last post.

I have more files but I will only upload them with a new post for each set up to try to save confusion.

One thing I noted
Display on VHF which is only acting as a receiver, is more erratic when connected to comp1 no matter what the set up is. When VHF is connected to comp2, course speed and position still fluctuate but are much smoother.
This seems to be the same even when the inputs are are changed around.
So my suggestion is that comp1 is handling the data differently to comp2.

Thanks
redog

[bdbcat]

redog....

OK, I lost the thread on which file is which computer.

However, looking just at $GPRMC, I extracted this much from from the top of nmea2.pdf.

$GPRMC,001055,A,2711.2373,S,15305.7502,E,6.7948,43.602,020813,,*0C
$GPRMC,001056,A,2711.2362,S,15305.7518,E,6.7948,43.602,020813,,*04
$GPRMC,001057,A,2711.2345,S,15305.7534,E,6.6845,41.960,020813,,*0A
$GPRMC,001058,A,2711.2332,S,15305.7545,E,6.7948,43.602,020813,,*07
$GPRMC,000912,A,2711.392,10,0.80,7.54,M,40.161,M,,*54
$GPRMC,001059,A,2711.2320,S,15305.7561,E,6.6845,41.960,020813,,*07
$GPRMC,001100,A,2711.2307,S,15305.7572,E,6.7948,43.602,020813,,*09
$GPRMC,001101,A,2711.2289,S,15305.7591,E,6.6845,41.960,020813,,*06
$GPRMC,001102,A,2711.2276,S,15305.7602,E,6.6845,41.960,020813,,*0C
$GPRMC,000913,A,2711.3946,S,15305.6351,E,6.8635,8.345,020813,,*39
$GPRMC,001103,A,2711.2264,S,15305.7618,E,6.6845,41.960,020813,,*05
$GPRMC,001104,A,2711.2246,S,15305.7629,E,6.6845,41.960,020813,,*00
$GPRMC,001105,A,2711.2234,S,15305.7645,E,6.6845,41.960,020813,,*0E
$GPRMC,001106,A,2711.2221,S,15305.7656,E,6.9529,42.374,020813,,*0F
$GPRMC,001107,A,2711.2210,S,15305.7673,E,6.6845,41.960,020813,,*0F
$GPRMC,001108,A,2711.2193,S,15305.7689,E,6.7948,43.602,020813,,*09
$GPRMC,001109,A,2711.2180,S,15305.7700,E,6.7948,43.602,020813,,*0A
$GPRMC,001110,A,2711.2168,S,15305.7716,E,6.7948,43.602,020813,,*03
$GPRMC,001111,A,2711.2150,S,15305.7727,E,6.7948,43.602,020813,,*0B
$GPRMC,001112,A,2711.2136,S,15305.7746,E,6.7948,43.602,020813,,*0F
$GPRMC,001113,A,2711.2125,S,15305.7762,E,6.7948,43.602,020813,,*0A
$GPRMC,001114,A,2711.2112,S,15305.7773,E,6.7948,43.602,020813,,*09
$GPRMC,001115,A,2711.2095,S,15305.7789,E,6.7948,43.602,020813,,*03
$GPRMC,001116,A,2711.2082,S,15305.7800,E,6.7948,43.602,020813,,*08
$GPRMC,001117,A,2711.2070,S,15305.7817,E,6.7948,43.602,020813,,*02
$GPRMC,001118,A,2711.2058,S,15305.7833,E,6.7948,43.602,020813,,*01
$GPRMC,001119,A,2711.2041,S,15305.7844,E,6.7948,43.602,020813,,*08
$GPRMC,001120,A,2711.2029,S,15305.7860,E,6.7948,43.602,020813,,*0A
$GPRMC,001121,A,2711.2015,S,15305.7879,E,6.7948,43.602,020813,,*0C
$GPRMC,001123,A,2711.1985,S,15305.7906,E,6.7948,43.602,020813,,*04
$GPRMC,001124,A,2711.1974,S,15305.7922,E,6.7948,43.602,020813,,*0B
$GPRMC,001125,A,2711.1961,S,15305.7933,E,6.7948,43.602,020813,,*0E
$GPRMC,001126,A,2711.1949,S,15305.7950,E,6.7948,43.602,020813,,*02
$GPRMC,000914,A,2711.3923,S,15305.6357,E,6.86345,020813,,*33
$GPRMC,001127,A,2711.1932,S,15305.7966,E,6.7948,43.602,020813,,*0A

As you can see, there is a spurious source of $GPRMC, three oddball sentences in the above snippet (912, 913, 914). Different sentence structure, though still valid. Out of sequence time hack. Clearly out-of-sequence lat/lon fields.

Query: Where is the bogus RMC coming from? This is your homework....

Could be a datum problem on one of the GPS receivers, since the bogus lat/lon is close, but not within nominal GPS specs.

Good luck
Dave

[redog]

Dave
I saw that RMC was possible problem so I dropped RMC from UDP supplied from comp2.
attached

Question how are you extracting all RMC into 1 document?
If I can do that looking will be easier.
I'm still learning linux.

redog

[bdbcat]

redog....

I'm sure there is a one-line linux incantation to do it, but I simple-mindedly just open the file in an editor and delete manually any line that does not start with $GPRMC. You could do that, too.

So, what is the latest attachment? Is it live, and did the position jump while it was being recorded?

Dave