Hi All
HP Proliant DL120 G7 - 16GB RAM - 2x SSD HD Raid 1 - Intel 4 port server NIC card - 2 internal onboard NIC's
Centos 7 x64

Saturday I ran yum update. This updated the kernel to kernel-3.10.0-229.1.2.el7.x86_64 from kernel-3.10.0-123.20.1.el7.x86_64.

The first thing that got me worried was the fact that only 1 of my 6 NIC's kept the same position. The MAC addresses were allocated diffetent device names.

Secondly I noticed that my DLINK VPN router (pptp coneection) to a UK address lost connection. After many hours of iptable and Wireshark investigation I rebooted onto the old kernel-3.10.0-123.20.1.el7.x86_64. The NIC's then went back to the original order and the VPN connected.

I have multiple PPTP VPN's to various companies and each of these had stopped working on the new kernel-3.10.0-229.1.2.el7.x86_64.

I have a suspicion that for some reason the kernel is blocking the GRE packets. I was able to connect to the UK VPN using L2TP from the PC, the DLINK router only supports PPTP.

Am I the only person having bad effects with the new kernel-3.10.0-229.1.2.el7.x86_64 I'm hoping not.

So the nic's got reordered - well, that could very well cause your problems. Have you confirmed that the configuration for each nic is correct?
Your VPN tunnels are handled by a DLINK VPN router, it will not be affected when you update some computer in the network.

Hi,
If your server is firewall with NAT, try to charge nf_nat_pptp and nf_nat_proto_gre kernel modules. My problem was resolved with this.

Regards.

Hi tuxLinuxbr
I'm not sure what you mean by change the nf_nat_pptp and nf_nat_proto_gre kernel modules. Can you expand a little on that for me please, please.

The solution for was to load the nf_conntrack_pptp module from within IPTABLES as suggested by jourquinb in this Fedora forum http://forums.fedoraforum.org/showthread.php?t=302794. It seems that this module is no longer loaded as default by the kernel.
/sbin/modprobe nf_conntrack_pptp

The last Kernel I was able to use on my Lenonvo desktop was 3.10.0-123.20.1.el7.x86_64. Since then the system hangs as soon as it has loaded. This is the ABRT error message I get for this each and every time I can only use 3.10.0-123.20.1.el7.x86_64.

id 2dcd7224b3e6555c746bfab35458c5093685531f
reason: BUG: soft lockup - CPU#0 stuck for 24s! [kworker/0:2:13686]
time: Wed 04 Nov 2015 08:48:11 PM PST
cmdline: BOOT_IMAGE=/vmlinuz-3.10.0-229.20.1.el7.x86_64 root=/dev/mapper/cl_localhost-root ro crashkernel=auto rd.lvm.lv=cl_localhost/swap vconsole.font=latarcyrheb-sun16 vconsole.keymap=us rd.lvm.lv=cl_localhost/root rhgb quiet noapic LANG=en_CA.UTF-8
Directory: /var/spool/abrt/oops-2015-11-04-20:48:11-13691-0
Reported: cannot be reported

Two updates and it still is not usable except with 3.10.0-123.20.1.el7.x86_64. Is there anything being done -- fixes -- patches or just stick with older kernels?