[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: A few observations about systemd



On Tue, Jul 19, 2011 at 3:02 PM, Ian Jackson
<ijackson@chiark.greenend.org.uk> wrote:
> Tollef Fog Heen writes ("Re: A few observations about systemd"):
>> ]] Russell Coker
>> | But it does result in a system that doesn't work properly.
>>
>> Well, yes.  If init crashes, stuff generally don't work that well
>> afterwards. :-)
>
> That's why its best to have init contain as little code as possible,
> to minimise the chance that bugs cause init to crash.  Which was
> Bastien Roucaries's point, I think.

Exactly and udev, socket stuff and lot of stuff made more likely to
get this. pid == 1 should do only child rip, and respawn a systemd
process if it crash. And start a crash shell  if needed. Not rocket
science about ipc only a loop and two signal to catch:
- get SIGING: respawn systemd
- get SIGUSR2: spawn a sulogin shell
- check if systemd child die, respawn it if needed (rate limited)

All the funky stuff is done by a child of init. Remeber the contract
of init made by the kernel. In all the case you should reap children.
If they are some delay they are potential of dos.

Bastien

>
> Ian.
>
>
> --
> To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
> Archive: 20005.32880.591127.827897@chiark.greenend.org.uk">http://lists.debian.org/20005.32880.591127.827897@chiark.greenend.org.uk
>
>


Reply to: