> OK, I'd forgotten this request this afternoon; but, completed this
> remotely a few moments ago. I copied /cdrom/etc.lrp to /floppy and
> rebooted -- which logged the overflow errors.
>
> So, clearly, there is something unusual about /cdrom/etc.lrp . . .
>
> > Also, please provide more details about the IPChains errors, and any log
> > messages that might be relavent. They might shed some light on what's
> > actually going on...
>
> Apparently, these are not always present -- I don't remember seeing them
> today. Nor have they ever shown up in any logfile I've reviewed. It
> would appear that, whatever these maybe, they exist only on STDERR to
> console and are difficult to document ;<
>
>
> OK, since I haven't been able to get to the firewall nor figured out how
> to get files off of the firewall, today I went to the firewall, removed
> /floppy/etc.lrp and rebooted. Once it came back up -- including the
> subject errors -- I copied /cdrom/etc.lrp to /floppy and backed up
> (lrcfg) etc.lrp to same floppy. From a potato box, I extracted each
> into their own directory and did diff -r on each. The *only*
> differences were these:
>
> diff -r etc_cdrom/etc/hostname etc_lrcfg/etc/hostname
> diff -r etc_cdrom/etc/hosts etc_lrcfg/etc/hosts
> Binary files etc_cdrom/etc/ioctl.save and etc_lrcfg/etc/ioctl.save
> differ
> diff -r etc_cdrom/etc/motd etc_lrcfg/etc/motd
> diff -r etc_cdrom/etc/resolv.conf etc_lrcfg/etc/resolv.conf
>
> hostname, hosts, motd and resolv.conf differences are *all* accounted
> for by name assignments made by /etc/rcS.d/S39network during bootup.
>
> The total difference in size between each etc.lrp is 53 bytes!
>
> I'm not absolutely clear what is happening with ioctl.save? What is
> this file? How can it manifest itself in this problem?
>
> Or, since commenting OUT a couple names that are dynamically built to
> populate /etc/hosts, is there some timing and/or sequencing issues that
> could interfere? Are there ipchains rules generated *prior* to
> populating /etc/hosts? I admit, in twenty (20) years doing this stuff,
> I've never seen anything quite like this -- especially, since the errors
> do not appear to be indicative of any real problem and after generating
> /floppy/etc.lrp from the post-error environment, we cannot duplicate the
> errors ;<
>
> What do you think?
It sounds like a problem with name resolution. I'd bet something is using
hosts entries or trying to resolve names before the network scripts update
the files. You might try restoring the hostname, hosts, and resolv.conf
files individually, re-run the network scripts, and see if you get the
errors. This would at least help narrow down the search for problems.
To the best of my knowledge, there are no "race" conditions in the startup
scripts, which are executed sequentially. There can, however, be problems
with the sequencing. For instance, you don't want to launch a web-server
before another script has configured the interfaces & routing. I'd bet
there's something like this going on...
Charles Steinkuehler
http://lrp.steinkuehler.net
http://c0wz.steinkuehler.net (lrp.c0wz.com mirror)
_______________________________________________
Leaf-user mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/leaf-user
