Hi to all, I'd like to inform about %subject%. I have recently upgraded tens of C4 machines to 2.6.9-78.0.5.EL kernel. Two of machines are having symptoms being said in Centos3131, BZ456664. There are two workarounds, change /etc/init.d/iptables sed -i.bak -e 's/_tables$/_CONNTRACKNEW/;s/_conntrack$/_tables/;s/_CONNTRACKNEW$/_conntrack/' /etc/init.d/iptables or run a "newer" kernel. I have rebuild RH kernel on Centos and published it here http://fs12.vsb.cz/hrb33/pub/Centos3131/
I guess CentOS team should release warning to users about that issue and inform them to not upgrade to recent kernel on remote machines without remote management. I will explain why. If user upgrades remote machine to 2.6.9-78.0.5.EL kernel, initiates reboot finally, machine is going to hang on iptables unload. If the machine is remote one, without remote reboot, switching power supply, remote management (ILO, Drac, ...) machine is complete dead. The only way is to send someone to reset the machine. Regards, David Hrbáč