On Friday, February 7, 2020 4:18:19 PM CET Steve Newcomb wrote:
> On 2/7/20 9:51 AM, Simon Wunderlich wrote:
> > On Friday, February 7, 2020 3:13:47 PM CET Steve Newcomb wrote:
> >> @rpc152:/tmp/log# echo "$(logread)" | grep batman
> >> Thu Feb 6 15:21:13 2020 kern.warn kernel: [174193.938445] batman_adv:
> >> [Deprecated]: batctl (pid 22747) Use of debugfs file "nc_nodes".
> >> @rpc152:/tmp/log#
> >> 
> >> 
> >> What have I missed?
> > 
> > Hi Steve,
> > 
> > you can use "batctl log" to retrieve the log. It will not appear in your
> > logread.
> 
> Alas, that doesn't work either, and I don't know why:
> 
> root@rpc152:~# batctl log
> Error - no valid command or debug table specified: log
> Usage: batctl [options] command|debug table [parameters]
> options:
>          -h print this help (or 'batctl <command|debug table> -h' for
> the parameter help)
>          -v print version
> 
> commands:
>          meshif <netdev> aggregation|ag [0|1] display or modify
> aggregation setting
> ...
> 

Oops, you are right, we have actually removed that command in 2019.2. You can 
use one of the two following commands:

cat /sys/kernel/debug/batman_adv/bat0/log

(will be removed in the future when debugfs support is dropped

 trace-cmd stream -e batadv:batadv_dbg

> > When the problem happens, you can also check "iw wlan0 station dump"
> > and other
> > debug files (batctl n for neighbors) to find out if the WiFi layer is
> > still
> > working. It wouldn't be the first time that actually the WiFi chip or
> > driver
> > has a problem, not batman-adv.
> 
> I've seen that "batctl n" works, and "iw mesh0 station dump" works, too.
> 
> I am arranging for the nodes to send me such mail when things have gone
> awry, but prior to rebooting.  I've written a tiny mail queueing system
> that optionally uses nonvolatile memory for the queue.

By "works" you mean you get useful outputs where the timeout is not increasing 
or similar? can you still "batctl ping" to one of your neighbors?

Cheers,
      Simon