[Top][All Lists]

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

Re: Hurd logging. (was zalloc panic)

From: Marcus Brinkmann
Subject: Re: Hurd logging. (was zalloc panic)
Date: Sun, 3 Mar 2002 15:43:07 +0100
User-agent: Mutt/1.3.27i

On Thu, Feb 28, 2002 at 11:42:16PM -0700, Jon Arney wrote:
> As I said, I am not opposed to using syslogd and think starting there
> is a good idea.  If you believe it is feasible without causing
> recursive RPC calls, I'll believe you.  My only concern was that
> if, say, 'pfinet' needed to log something and syslog were configured
> to write to UDP ports, pfinet would have to write to a UDP port
> causing perhaps a recursive loop of log calls.  I realize that this
> is ultimately a pathological case, but it's worth considering.

If sending something over a socket causes the socket server to log a
message, then you have a serious problem anyway.

The same problem exists in a more minimal model: If writing something to a
file causes the filesystem server to log a message, then you are in big

You'll have to make sure those pathes are clean.  And in any case, logging
should be the exception rather than the default.  I disagree with our
earlier analysis that for example EINVAL from the auth server is unhelpful,
as this always means that the port died in transit (if it comes from the
auth server), and is something that can happen under normal circumstances
(for example, a client calls io_reauthenticate and then exits before it sends
auth_user_authenticate).  BTW: This error is then seen by the server, who
will just deal with it by failing with the io_reauthentication.  Which is
fine as the io port is probably dead by then as well.  The only case where
this could be seen by the user is if the user destroys the rendevous port
itself before the authentication (silly thing) or he uses another task's
port and that task destroys it (also silly).

BTW, just as your information, stdout and stderr can be captured either by
making the translator active:

settrans -a /mnt /hurd/ext2fs /dev/hda1 > /var/log/mnt.log 2>&1

or by using an intermediate shell with a passive translator:

settrans /mnt "/bin/sh /hurd/ext2fs /dev/hda1 > /var/log/mnt.log 2>&1"


`Rhubarb is no Egyptian god.' Debian http://www.debian.org brinkmd@debian.org
Marcus Brinkmann              GNU    http://www.gnu.org    marcus@gnu.org

reply via email to

[Prev in Thread] Current Thread [Next in Thread]