I am running FC4 and whenever I bring up my eth0 I get the following messages:

SIOCGIFADDR: Cannot assign requested address
SIOCSIFBROADCAST: Cannot assign requested address
SIOCSIFBRDADDR: Cannot assign requested address
SIOCSIFFLAGS: Cannot assign requested address

I am using ip aliases via /etc/sysconfig/network-scripts/ifcfg-eth0-range0, but I don't think that's related (error messages show up before "Bringing up interface eth0-range0" message).

The odd thing is that the interface actually does seem to come up correctly, with all the aliases, so I'm not sure what the error is referring too. I know it is assigning the interface an ipv6 address as well, could this be the cause of the error?

One last thing, eth0 has a /29 subnet mask, but whenever I reboot the computer the routing table has the correct default gw, a correct /29 network route for the device, and a completely unnecessary /8 network route for that device. Any idea where the /8 address is coming from and how I can stop it?

Neither of these are really hurting anything, but they ShouldNotHappen(tm), so I'd like to track down the problem.

ifcfg-eth0, ifcfg-eth0-range0, and network (ips have been changed to protect the ignorant)

/etc/sysconfig/network-scripts/ifcfg-eth0:
DEVICE=eth0
BOOTPROTO=static
BROADCAST=64.10.20.39
IPADDR=64.10.20.34
NETMASK=255.255.255.248
NETWORK=64.10.20.32
ONBOOT=yes
TYPE=Ethernet

/etc/sysconfig/network-scripts/ifcfg-eth0-range0:
IPADDR_START=64.10.20.34
IPADDR_END=64.10.20.37
CLONENUM_START=0

/etc/sysconfig/network:
HOSTNAME=ns0.mydomain.com
NETWORKING=yes
GATEWAY=64.10.20.33
GATEWAYDEV=eth0
FORWARD_IPV4=no
NOZEROCONF=yes

Thank you for any insights in advance.

when did the problem occur.

if you can trace the problem starting to that point, then i'd take the card out, reboot, delete any references to it..then start over from scratch by installing the card.

good luck.

It's been happening since day 1 I believe. This is FC4 btw. Removing the network card is not really an option for two reasons; one it's built-in to the motherboard, but actually even more important is that this is a production machine - downtime is not allowed.

well you can do a 'word' search for the SIOCGIFADDR and see where that may play into the error.

are you using the sever version of fed or just 4?

and i do think ipv6 can be part of the problem.. possibly.

Sorry for the slow response, I was out of town.

Just using regular FC4, and yes I think ipv6 could be part of the problem. Is there a way I can speficy not to set up any ipv6 stuff? I'm not using it anyway.

OK, I figured out how to turn off ipv6, didn't make any difference though.

i know you said it is a production mach and can't be brought down...

however the ominous messages are surely signs of bad things to come.

i would figure out how to take the sys off line long enuff to reset your network.

i'd clear all the settings (config files deleted if nec) and start fresh with setting it up.

this is what i would do.

now if you wait you will definitely find the culprit..eventually.

however again you risk having more serious problems every day that passes... one because this 'shouldn't be happening', and two, we (you actually) should know what is going on here so it can be fixed.

again, good luck.. let us know (if i think of anything else, i will post

coupla more things:

*{EDITED} added sentence below:

double check your man pages and /usr/share/docs for assiging alias.. make sure your not pushing it beyond what they are meant for.

do a netstat -anp |more and see (in the begining where the addys are, if anything looks weird).

i suspect that the error message (as usual) may be pointing us in the right direction (ie. 'cannot assign addy..') .. well, duh if the addy is taken already.

does this make sense? lol

and oh, i didn't mean to sound 'mean' in the previous note about how you should have this figgured out before all hell breaks loose..

i was just talking like that because well, it is better to now, than wait for your boss to get pissed because the mach went bye byes.