Re: cfrun: question about the output for unreachable hosts

[David Douthitt]

Moritz Bunkus wrote:
> Everything is fine. I'm just a bit confused about cfrun's output if a
> host is unrechable. Example:
>
> 0 root@ls-bs-si2:/home/mbunkus$ cfrun ls-bs-ws3
> cfrun(0):         .......... [ Hailing 
> ls-bs-ws3.bs.linet-services.de ] ..........
> cfrun:ls-bs-si2.bs.linet-services.de: Couldn't open a socket
> cfrun:ls-bs-si2.bs.linet-services.de: socket: No route to host
>
> ls-bs-ws3 is indeed down, but why does cfrun use ls-bs-si2... as the
> host name in this case? This gets even more confusing when I run cfrun
> without any argument and it iterates through all hosts in
> cfrun.hosts. For each unreachable host it just prints the two lines
> shown above, and I have no clue whatsoever which hosts are actually
> down.

The output has already been mostly explained, but I'd like to suggest a 
change in cfengine's output.

The latter two messages:

> cfrun:ls-bs-si2.bs.linet-services.de: Couldn't open a socket
> cfrun:ls-bs-si2.bs.linet-services.de: socket: No route to host

are quite likely the output of an error message from a system call. 
Thus, the messages (after ":") are generated by the system.

I would recommend that either they be replaced or supplemented by a 
message that could be output (in this case) like thus:

cfrun:ls-bs-si2.bs.linet-services.de: Host 
ls-bs-ws3.bs.linet-services.de seems to be unreachable

Part of the problem seems to stem from putting out low-level error 
messages without giving the user any high-level explanation for what 
seems to be happening.

My favorite system example is what happens when you try to add a route 
and the messages that come back...

--
David Douthitt
UNIX System Administrator
Linux+, LPIC-1, RHCE
HP-UX, Unixware, Linux, FreeBSD, OpenBSD
Member: ACM, USENIX/SAGE