Is Watchdog class ok? | Thales IoT Developer Community
June 17, 2016 - 8:44am, 6370 views
Hi,
I'm facing problems with instability with EHS6 Modules. When i activate the watchdog it is rebooting sporadic without any reason. When watchdog is deactivated it run's without any reboots for days/weeks..
My question: Could you imagine that there is a fault in this firmware?
My actual Firmware:
Cinterion
EHS6
REVISION 03.001
Pseudo code:
Watchdog2 wd = new Watchdog2();
wd.start(300);
do { wd.kick(); Thread.sleep(100); } while (true);
BR
Andy
Hello,
This should not happen without purpose.
Please write some more information about what the application is doing? Maybe one of the important threads is hanging or the execution took more time than expected and the watchdog was not kicked.
Maybe you are creating more than one instance and one is not kicked or it is not stopped. Have you tried to prepare the MIDlet which only creates and kicks the watchdog like in your example - is the module still rebooting?
Please check the detailed firware version with ati1 command.
Best regards,
Bartłomiej
Hello Bartłomiej,
there is only one instance of the watchdog. The application it opening a connection to our server over tcpip and is communicating over this connection from time to time(at least once in five minutes). While we are in connected state we are doing the wd.kick(), if no more connected for some reason we are doing a reboot with the watchdogs help by no more calling wd.kick(); When i comment out the wd.start(300) at start and the wd.kick() in the loop the app is running without any problems for a very long time but with the two lines more it can be, that after 5 seconds running it reboots but also for a few days it's weired and i used one usb output for std::out but there is no message before the module reboots (exception or what else)..
ATI1
Cinterion
EHS6
REVISION 03.001
A-REVISION 00.000.14
BR,
Andy
Hello,
This looks very strange especially that the module could be rebooted in 5 seconds before the timer expires. So maybe it would be a good test to try to reproduce it with the MIDlet that only uses watchdog and nothing else. Then I could also try with the same MIDlet.
The firmware version that you are using is not the latest so you could try to update before further analysis and tests.
What hardware are you using with these modules?
There's one more issue - the watchdog is doing a hard reset without caring for network deregistration and other cleaning operations. So it should not be the way of normal switching off because there some risk of data loss or **********. The recommended way is using a software shutdown, for example with AT^SMSO command. Are the modules new or were they working in the field for some time?
Regards,
Bartłomiej
Hello Bartłomiej,
Cinterion EHS6T
i've have now reduce my app to the minimu as you suggested and voila after 70 seconds uptime it rebooted.
Should i send you the source code of it?
BR
Andy
Hello,
Have you updated the firmware before the test?
Please send me the MIDlet.
How about the hardware - are you using some custome HW?
Regards,
Bartłomiej
Hello Bartłomiej,
no, because this (03.001 000014) is the latest firmware available from http://cinterion-extranet.force.com/
Is there another download portal to get a newer one?
Where to send to? I do not have an email address..
It is a norm EHS6T module..
BR Andy
Hello,
I've sent you the email.
This is the only portal. But you should also be able to get the software from your local technical sales.
So you are using the terminal version. This terminal has also an additional hardware watchdog which you can use. It can be configured on serial interface or i2c.
Is it possible that you have tied this watchdog and maybe configured some settings that could cause the reboots?
Hi,
no, i use only the watchdog class. I've sent you already the source code, so you can have a look at it.
BR
Andy
Hello,
I've sent you my comments via email. I'll try to also test it and possibly share more comments.
Regards,
Bartłomiej
Hello,
Have you any new information about this problem. We too are having some unexpected reboots and the watchdog is a possibly culprite. We have similiar issues in which the module will reboot without notice although there are plenty of kicks happening.
We plan on trying a module with the watchdog disabled to see if the reboots stops, then we would know for sure the problem was related to the watchdog.
I would like to know more about this problem and if it was resolved as it could be part of the issues we are having.
Thank you.
Pages