Re: s6-rc-init hangs during stage2

From: Vasil Yordanov <vasil.yordanov_at_gmail.com>
Date: Mon, 6 Jun 2016 13:03:37 +0300

Hello Laurent,

Thanks for the fast response.

Yes, /run is not on a tmpfs. There is possibility to do this in docker I
will do this later.
But the problem occurs on a first ever boot of the container, so /run is
empty with sure, because I'm spawning a new container each time.
I will put the options to strace you recommend me, but will need some time
to reproduce the problem. Will write you soon.

Regards,

Vasil



On Mon, Jun 6, 2016 at 12:55 PM, Laurent Bercot <ska-supervision_at_skarnet.org
> wrote:

> On 06/06/2016 11:35, Vasil Yordanov wrote:
>
>> Docker, with ENTRYPOINT "/etc/s6-linux-init/init"
>> #if { s6-mount -nwt tmpfs -o mode=0755 tmpfs "/run" }
>>
>
> You have commented this line, which means /run is not a tmpfs anymore,
> but a subdirectory in your container filesystem.
> If this subdirectory is persistent in your container, then the
> /run/s6-rc:initial directory may not be empty; this could cause trouble.
>
> /run is expected to be empty at boot time, so if you're not making it a
> tmpfs, you should at least delete everything under /run before copying
> stuff into it.
>
> If it doesn't solve your issue, please use the "-v -f -s256" options to
> strace when you're stracing s6-rc-init, post the result on a site like
> pastebin, sprunge or tpaste, and post the URL here. (At least the -f
> option,
> whichwill help see what the s6-ftrigrd process is doing.)
>
> --
> Laurent
>
>
Received on Mon Jun 06 2016 - 10:03:37 UTC