I’m having some trouble with "Channel has faulted; channel=“default” retries=0 faultCode=“too_many_commits” faultMessage=“Too many commit iterations” in the new beta. The provision is not that big and worked in the previous version.
I seem to remember that I tweaked SESSION_TIMEOUT (maybe?) to deal with this in the previous version. Might have been something else, but the question is how to tweak my way out of this.
I have the same problem. A refresh of one parameter results in many many acsRequestName=“GetParameterNames” until channel has faulted.
I tried tweaking with
GENIEACS_MAX_COMMIT_ITERATIONS=64
GENIEACS_GET_PARAMETER_NAMES_DEPTH_THRESHOLD=3
but same result.
I went a bit over board and did this in genieacs.env on the beta:
GENIEACS_SESSION_TIMEOUT=60
GENIEACS_MAX_COMMIT_ITERATIONS=64
GENIEACS_CWMP_KEEP_ALIVE_TIMEOUT=30000
That seems to have eased the pain a bit. It still faults when I try to refresh the entire device-tree of a cpe, but at least I’m able to provision them without crashes.
Did that used to work in v1.1 without throwing a fault? Can you share traffic capture or debug log?
SESSION_TIMEOUT I believe shouldn’t any effect on this.
MAX_COMMIT_ITERATIONS definitely does. Have you tweaked that option in your v1.1 config before? I don’t see how it could have worked fine in v1.1 but not in v1.2 unless that’s a bug in v1.2.
MAX-COMMIT was tweaked before, but didn’t get transferred as I had the old in a docker container and the new beta installed regularly. I remembered it as I read the other answers in this thread.
It seems to be ok now, but it still freaks out if I try to refresh the entire device tree of a device. I’ll have to trial and error a bit to see what sticks.
I guess I can remove SESSION_TIMEOUT then?
With regards to this particular issue, yes.
I´ll send you the logs. Can you guide me how to enable logging for a device in 1.2?
Sure. From admin → config, set ‘cwmp.debug’ to:
id = '<your device ID>'
This will tell Genie to dump debug info to the file specified in your GENIEACS_DEBUG_FILE env var.
1 Like
Thank you. If you need anything else please let me know. Traces are on our pastebin:
genieacs-cwmp-accses.log:
https://paste.ccgmbh.net/paste/Waa5kLFS#f9hMV5za-YRAxgaN17mU5aWY21rk7JU3G+KkeyjR/e4
genieacs-debug.yaml:
https://paste.ccgmbh.net/paste/eWJ2om8a#obsn5CxG3D4hqWaD02uPXEtNgQnCmGFi8FIN0fYxhWk
So, have you fix this problem ?
I encounter this problem too.
Couldn’t reproduce the problem. What parameter were you attempting to refresh?
Hi, it was a summon. But it does not matter which parameter, it always end up in too many commit iterations.
Is there anything I can do to investigate the behaviour?
Can you reproduce it using genieacs-sim?
Didn´t use it so far, I´ll try.
can not reproduce it with genieacs-sim 
Is it possible to extract data model from CPE connected to version 1.1.3 for reference?
<00040E-FRITZ%21Box-00150C6AF983: No contact from device>
Could this be a connection issue? Timeout? When I do summon I get a “No contact from device” popup.
You can but I doubt it’ll help. Can you delete the device from Genie and try again? I’d like to see the logs from the very first contact.
Nevermind, I managed to reproduce the issue. Will fix soon. Thanks.
1 Like
Fixed now in latest master. Thanks for your help @sroehrig.
1 Like