Michael Burger wrote:
>
> To my knowledge, no...however, the easiest way to find out is to log
> in to the console, and run "ifconfig" to see if the interface has
> truly been taken down.
>
> However, as I stated at first, I do not believe that Linux actually
> does an "ifconfig eth0 down".
>
I don't think so either. The patch listed earlier in the thread might
worthwhile. I ran into this on a small network that I had adminsrative
control over all five machines (so i did it to myself) The site is
ninety miles away and after replacing a machine, I come back to the
office and notice that a cron job that talks to 4 of them was failing on
the ipaddress of the new box (should have been the same as the old box)
anyway there was also intermittant issues with another box too. I
pulled my hair out remotely for longer than it would have taken to drive
back there but learned alot in the process.
What apparently was occuring was as a packet came in for the conflicted
machine, the occasional arp request from the linux router I was vpn'ed
through would route packets to the first on that answered. I logged in
to one changed the password and ran the job and sure enough it failed.
Hmm I said and tried to login with the new password but could not.
hmm. I wen't and got a cup of coffee came back and logged in. Now I
was really confused. loggout and try to login nope tried old password
and viola logged in. It was around this time that I realised what I
must have done and changed the ipaddress to the correct one and
everything started working. That is when I decided that DHCP is my
friend and will setup the next network using it. After al lit was only
four machines right? How hard can that be to keep track of. Kicked my
ass for about five hours.
All this is a long winded way to say that the interface could indeed
appear to be down if the router is sending the packets to the wrong mac
address.
I have wondered why the kernel does not include at least a check for the
ipaddress space and apparently at least someone has written a patch to
do just that. I may indeed include it in my kernels from now on once I
find out why the kernel team does not include it.
Bret
_______________________________________________
Redhat-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-list
