|Anonymous | Login | Signup for a new account||2017-08-24 01:17 CEST|
|My View | View Issues | Change Log | Roadmap | My Account|
|View Issue Details|
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0000183||easycwmp||[All Projects] Bug||public||2016-11-25 17:17||2016-11-30 10:06|
|Platform||VM||OS||Ubuntu Server||OS Version||16.04 LTS|
|Summary||0000183: ACS Retry timeout incorrect|
|Description||I was testing the retry and noticed that sometimes when watching the easycwmpd output on the console, there was no timeout. So I allowed it go on for 10 retries and could also see that the timeouts were too short when they are longer than 60 seconds. I've tried this a number of times and it always happens. I believe the problem could be the Periodic Interval timeout which I've left as 100 seconds. The maximium timeout I'm seeing is 70 seconds which with the timeout of 30 seconds curl timeout is the 100 seconds timeout. I'm also seeing timeout of zero seconds, see attached syslog and here are the results of my last test.|
Retry 1 - 8 seconds
Retry 2 - 15 seconds
Retry 3 - 0 (zero) seconds
Retry 4 - 60 seconds
Retry 5 - 0 (zero) seconds
Retry 6 - 70 seconds
Retry 7 - 70 seconds
Retry 8 - 70 seconds
Retry 9 - 70 seconds
Retry 10 70 seconds
I took the measurement between the lines "retry session" and "start session" in the syslog.
|Steps To Reproduce||Turn off ACS|
Start easycwmpd -f -b
After 10 Retries stop easycwmpd and collect syslog
|Tags||No tags attached.|
|Attached Files||cwmp-retry-error.txt [^] (9,878 bytes) 2016-11-25 17:17 [Show Content]|
|this because you are using a small periodic interval. The "start session" causes are the periodic events and not related to the retry. Please set a big periodic interval in the /etc/config/easycwmp and you will see the right timeout of retry interval.|
I don't see changing the retry interval to a larger one as a fix, it should be able to handle any length of periodic interval.
What should be done in the code is to cancel the periodic interval if the curl timeout has occurred.
Only start the periodic interval when the communications with ACS has finished successfully.
Then it wouldn't matter what the periodic interval was.
edited on: 2016-11-29 11:52
Please note it's not a bug! It's a normal behaviour and it's compliant with the standrad. And I suggested to set a big periodic interval in order to see the right retry timeout and not to fix it.
Sorry, but your both proposals are not compliant with the TR-069 standard. You can always customize your easycwmp version according to your need. But please note that they are out of the standard
Can you let me know where this is in the standard?
Page number would be helpful.
From the Standard: https://www.broadband-forum.org/technical/download/TR-069_Amendment-5.pdf [^]
Page 34 (126.96.36.199 Session Retry Policy):
"A CPE MUST retry a failed Session after waiting for an interval of time specified in Table 3 or when a new event occurs, whichever comes first."
|2016-11-25 17:17||dgr999||New Issue|
|2016-11-25 17:17||dgr999||File Added: cwmp-retry-error.txt|
|2016-11-29 10:13||mohamed.kallel||Note Added: 0000575|
|2016-11-29 11:12||dgr999||Note Added: 0000576|
|2016-11-29 11:49||mohamed.kallel||Note Added: 0000577|
|2016-11-29 11:52||mohamed.kallel||Note Edited: 0000577||View Revisions|
|2016-11-29 14:18||dgr999||Note Added: 0000578|
|2016-11-29 16:44||mohamed.kallel||Note Added: 0000579|
|2016-11-30 10:06||mohamed.kallel||Status||new => resolved|
|2016-11-30 10:06||mohamed.kallel||Resolution||open => fixed|
|2016-11-30 10:06||mohamed.kallel||Assigned To||=> mohamed.kallel|
|Copyright © 2000 - 2017 MantisBT Team|