New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Mesh interfaces occationally break on boot #94
Comments
Can you take a look at dmesg during your test for any possible kernel panics from the wireless driver? |
There was this crazy stack trace that only existed in the broken one.... That what you are looking for? I have a bunch of dmesg logs from various tests if it is a loose end.
|
Given the stack trace from dmesg, I believe this is the same issue with the driver freaking out when the radio's channel or other parameters are modified after the adhoc interface has been created (or possibly when an adhoc interface is present in combination with other VAPs). This has also possibly been a problem with the Linux client. I'm going to attempt to track down this bug once and for all, and either fix it or find out its exact triggers so that we can work around it effectively (I suspect that deleting and recreating the adhoc interface on any channel change may be an effective countermeasure). |
any progress yet? i just have same problem with code from trunk, can not On Tue, Jan 21, 2014 at 7:01 AM, Josh King notifications@github.com wrote:
|
@westbywest, have you seen anything like this before? |
Resolved in v1.1rc1. |
This intermittent bug does not cause any processes to fail but will stop the node from sending any mesh packets over the network. I believe that this is associated with many other bugs we have been seeing. This is most easily recreated by re-naming the mesh ssid in the basic user interface but does not seem to be associated with the name given or length. Though, with that said, and the rule of large numbers being what it is I have seen the first instance using a string that is greater than nine characters with a number as the last character and a dash in the middle. It is not required, nor does it always work... but, it works more consistently than others.
cc: @jheretic as this is most likely a issue with hotplug/netifd and the kernel as shown in logs further down.
This is most easily identified as broken by looking at httpinfo and checking the Destination Gateway. It will have all 0's on the netmask.
Here are some examples from various logs I have taken.
Best I can tell the kernel seems to be failing and not triggering a new scan to find a IBSS to join on instances where this occurs. See the "kern.info" messages in the "good" logfiles below. The bad logfiles are just the same area in the logfile. The first "BAD" logfile shows the last time that there are kernel messages in the bad logfiles. I assume this is where the kernel error actually exists. The final "GOOD" log section shows what I assume to be the set of commands that the kernel is missing in bas restarts.
The text was updated successfully, but these errors were encountered: