Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 26 Nov 2009 08:36:15 -0800
From:      "Matthew Fleming" <matthew.fleming@isilon.com>
To:        "Ed Schouten" <ed@80386.nl>, "Julian Elischer" <julian@elischer.org>
Cc:        Rui Paulo <rpaulo@gmail.com>, Attilio Rao <attilio@freebsd.org>, gnn@freebsd.org, ticso@cicely.de, FreeBSD Arch <arch@freebsd.org>
Subject:   RE: [PATCH] Protect inetd and cron from being killedunder	high-pressure swapping
Message-ID:  <06D5F9F6F655AD4C92E28B662F7F853E021D4CFC@seaxch09.desktop.isilon.com>
References:  <3bbf2fe10911160702m3641b65cv15ac2942cbb023fd@mail.gmail.com>	<7i8we5xlbm.wl%gnn@neville-neil.com>	<30DCA579-53AD-49EA-A5A2-5A0796A285C1@gmail.com>	<20091120062637.GA49534@server.vk2pj.dyndns.org> <20091125144208.GZ20716@cicely7.cicely.de> <4B0D65FE.9020603@elischer.org> <CD5E7D77-6D17-40C3-97CB-4E0FA880F689@80386.nl>

next in thread | previous in thread | raw e-mail | index | archive | help

> we are considering whether a jail should have its own init...

Just FYI:

The AIX version of jails each had their own init process.  I wasnt =
directly working on that project/team, so I dont know all the details.  =
The only thing I remember was a few bugs my co-workers mentioned: first =
there were several places that assumed init was pid 1, and they all =
needed changing.  Second was I think some weird issues managing when the =
jails init was supposed to actually quit.  Last, when AIX wanted to send =
a signal to a processs parent like SIGCHILD, AIX needed to be careful =
which init to send it to for unparented processes, otherwise an init =
would get a signal for a process id it couldnt actually see.

Cheers,
matthew



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?06D5F9F6F655AD4C92E28B662F7F853E021D4CFC>