IPTables
[root@s1 ~]# service iptables start
Flushing firewall rules: [ OK ]
Setting chains to policy ACCEPT: security raw nat mangle fi[FAILED]
Unloading iptables modules: [ OK ]
Applying iptables firewall rules: [ OK ]
[root@s1 ~]#
Can anyone help?
Thanks,
Michael
15 Replies
@Jay:
You've defined a non-existent table in your iptables configuration. There's no iptable-table named "security". Those rules should be moved into filter, nat, or mangle (most likely filter).
How do I remove it?
Terry
"The issue is that the "Latest 2.6 Paravirt" kernel has a "security" chain and iptables doesn't know how to handle it. Usually switching to the "Latest 2.6 Stable" kernel resolves the issue without any further tweaking of the iptables init script (it often just ignores that chain and starts normally). Our builds team is indeed aware of this problem, however I do not have an ETA on if/when it will be resolved.
It is perfectly fine to continue using our "Latest 2.6 Stable" kernel – this kernel was actually the default selection for CentOS deployments until recently. No applications, with the exception of iptables, will operate differently when using the stable kernel."
So not really an answer if you can use the paravirt kernel without a problem. If you find out the answer, please post.
Note that if you execute an iptables-save while using the paravirt kernel, it will save a security chain in the /etc/sysconfig/iptables file so upon start-up with the 2.6 stable kernel, iptables will try to load a security chain and will really fail.
Terry
Haven't tried it out yet. Not sure if it will screw anything else up. Anyone found a fix for this yet or has tried this out?
Thanks,
Terry
mv iptables ~/iptables.bak
wget
chmod +x iptables
rm -rf 12023.txt
Now, "iptables" should now start successfully:
service iptables restart
EDIT: I don't have this error with the latest paravirt kernel 3
I tried with the latest paravirt 3 and I still get the same error.
Terry
@troublshootr:
I tried with the latest paravirt 3 and I still get the same error.
It is a bug in CentOS, not in the kernel itself, so I wouldn't anticipate newer kernels changing much.
@hoopycat:
@troublshootr:I tried with the latest paravirt 3 and I still get the same error.
It is a bug in CentOS, not in the kernel itself, so I wouldn't anticipate newer kernels changing much.
I don't consider it a bug in CentOS since it happen only with linode kernel and some other one.
@hoopycat:
So the bug doesn't happen if you download the latest mainline kernel from kernel.org, compile it using a reasonably-similar configuration (e.g. from /proc/config.gz on a Linode), and boot with it on normal hardware? If it doesn't happen, I will retract my statement just as soon as I finish eating my hat.
I don't have tested it with similar configuration of the linode one so I can't answer.
Any chance you could repost the solution to this problem as I am only just starting out with Linux/HP-UX (migrating from Windows) administration and the pastebin links are dead?