[Cerowrt-devel] [sqm-scripts] not started at boot?

[Sebastian Moeller]

Hi Alan,


On Mar 3, 2015, at 22:23 , Alan Jenkins <alan.christopher.jenkins at gmail.com> wrote:

> 
> On 3 March 2015 at 17:06, Sebastian Moeller <moeller0 at gmx.de> wrote:
> Hi Alan,
> 
> I would be delighted if you could try the etc/hotplug.d/iface/11-sqm and the updated usr/lib/sqm/run.sh from the attached files (I am not sure about the root to the files in the attached folder, but you will figure it out, I am sure ;) ). This should fix the reconnect when pop interface transiently go away issue. I am not sure whether this fixes the startup after reboot issue, so please feel free to test this as well.
> 
> Best Regards
>         Sebastian
> 
> Thanks Sebastian!
> 
> The scripts work for reboot & manual ifdown/ifup.  So I'd love to see this fix in openwrt CC.

	Excellent, now we have positive results from CC (you are running CC I believe) and cerowrt.

> 
> It still logs errors.  

	In the boot log? I tend to consider these not as errors but as valid information; as long as sqm-scripts is still being tested I want it to be verbose, so I have an easier time actually helping people… But once we agree it is stable I will try to make it a bit more quiet…

> Probably only from running the initscript first, it was just so noisy I thought it was still happening for every interface that came up :).   I used `uci set system. at system[0].log_size=200; uci commit` to prevent log overflow.

	Now that is useful, I always thought that this is a kernel default and can not be changed; so I learned something useful. But I take it you do not like to learn too intimate SQM details in your log files… Let’ revisit that once all real issues are solved, please.

Best Regards
	Sebastian

> 
> Full boot log attached.
> 
> Warm regards
> Alan
> 
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: log.txt
URL: <https://lists.bufferbloat.net/pipermail/cerowrt-devel/attachments/20150303/9bf2a84e/attachment-0002.txt>
-------------- next part --------------
>