Mailinglist Archive: opensuse (1188 mails)

< Previous Next >
Re: [opensuse] DHCP lease requests
  • From: Greg Freemyer <greg.freemyer@xxxxxxxxx>
  • Date: Fri, 6 Jan 2012 11:54:53 -0500
  • Message-id: <CAGpXXZJW7yNTb5ybKwPOOEjMoRTjQx=FQiaAbP+_uxt2ZAOi7w@mail.gmail.com>
On Fri, Jan 6, 2012 at 11:36 AM, Lew Wolfgang <wolfgang@xxxxxxxxxxxxxxx> wrote:
On 01/06/2012 03:25 AM, Per Jessen wrote:

Per Jessen wrote:

So I then issued a "/etc/init.d/network restart-all-dhcp-clients".
This option did a lot more as reflected in messages, but it still
didn't renew the DHCP lease.  It started dhcpcd but put it in
background, then dhcpcd reported "eth0 timed out".  At this point it
tried to "use old lease in /var/lib/dhcpcd/dhcpcd-eth0.info'", which
still contained the old lease info, and proceeded to reuse the old
lease without broadcasting for a new one.

It must have done a broadcast, otherwise why would it report "eth0
timed out"?  Following a time-out, it seems reasonable to attempt a
fall-back to the previous lease.

In the logs you posted, there is a 20 second span between broadcast and
time-out:

Jan  5 13:47:11 ironhead dhcpcd[6530]: eth0: broadcasting for a lease
Jan  5 13:47:27 ironhead ifup-dhcp:  eth0 DHCP4 continues in background
Jan  5 13:47:31 ironhead dhcpcd[6530]: eth0: timed out
Jan  5 13:47:31 ironhead dhcpcd[6530]: eth0: trying to use old lease in
`/var/lib/dhcpcd/dhcpcd-eth0.info'

Is there a problem with the DHCP server?  If you check the server logs,
you should see a complete trace of the negotation, but I guess the
server isn't responding, hence the time-out.




Ah, good point about the timeout!  The network is a large one, many
thousands
of hosts, and I don't have access to the server logs.  But I know who does
and
will check with him today.

I'll also try the switch in the opposite direction:  from the big network to
the
natted side of a small router.    There shouldn't be a timeout issue there.

Regards,
Lew

Lew,

If it's that big of a network , then it may just have a issue building
the spanning trees in the switches.

One of my clients has 80 switches in a very disorganized way. Thus
there are lots of routes from one place to another in there LAN.

When they bring up a new machine and do dhcp, it takes about 45
seconds to get the new IP. I'm convinced most of the time is the
switches playing games trying to figure out the most efficient way to
connect the newly introduced machine to the dhcp server.

Greg
--
To unsubscribe, e-mail: opensuse+unsubscribe@xxxxxxxxxxxx
To contact the owner, e-mail: opensuse+owner@xxxxxxxxxxxx

< Previous Next >
Follow Ups