Eth0 won't start up and unable to change kernel

March 12, 2015 2.7k views
Control Panels Logging Configuration Management DigitalOcean CentOS

My CentOS droplet rebooted this morning and now networking won't start up. I get the following error:

FATAL:Could not load /lib/modules/2.6.32-358.6.2.el6.x86_64/modules.dep: No such file or directory

Sure enough, there is no /lib/modules/2.6.32-358.6.2.el6.x86_64 directory. It appears that kernel isn't available on the server. The available kernels are:

  • 2.6.32-431.5.1.el6.x86_64
  • 2.6.32-431.11.2.el6.x86_64
  • 2.6.32-431.17.1.el6.x86_64
  • 2.6.32-431.20.3.el6.x86_64
  • 2.6.32-504.3.3.el6.x86_64

I shut down the droplet (via "poweroff") and went into DO's control panel to change the kernel. I tried each of these kernels but the droplet won't even boot up. The only way I can get it to boot up is by using my original kernel (2.6.32-358.6.2). The funny thing is -- this kernel doesn't appear in /boot/ and there isn't a directory for it in /lib/modules/.

So I have a few questions:

  1. Why does it power up with 2.6.32-358.6.2 (which is not on the server)?
  2. Why won't it power up with those other kernels (which are available on the server)?
  3. Is there some sort of log file so I can see what happens when it fails to boot up with one of the other kernels?
8 Answers

Please open a ticket with our support team so they can look into this for you.

I did open a ticket. But it took almost 6 hours to get a reply to it. Pretty disappointing.

In the end, DO Support fixed this for me. They manually installed the kernel.

They were able to fix it quickly once they started working on it. It should be helpful for others to know that DO Support is able to fix this manually.

I had the same problem after doing a yum update and reboot. Luckily, I took a snapshot before the updates, as I have burned enough on Linux updates. I reverted to my snapshot, and the problem appears to be that the kernel kernel-2.6.32-504.12.2.el6.x8664 update removes (it says as much) 2.6.32-358.6.2.el6. I think the real problem is then that the control DO panel needs to be set to the latest kernel as well, but I can't see a DO listing for 2.6.32-504.12.2.el6.x8664. So, you can't set the DO control panel to the latest kernel, and I believe you get this message if the DO control panel doesn't match.

I suppose when DO gets to updating the list of kernels in the control panel (or they current my understanding here), I will avoid the kernel update.

I tried to change from a 64 bit version to a 32 bit because of reasons and I couldn't get connected after that. Server said it was running, but no one was home. Thank cthulhu I had made a snapshot of the server before I did this. I just reverted back and everything is fine. That could have been a nasty mess if I had not made that snapshot before hand. Remember to always backup, backup, backup.

Hello everybody,
I've just upgraded Centos with yum upgrade and I spotted exactly the same problem - the eth0 interface does not start because of some strange errors. I'm currently recovering the machine using my last snapshot (created just before update) but it seems to be a critical issue that has not been fixed yet (although the thread is more than one month old!). @toepoke Did they fixed only your machine? Do you have the ticket number?


Yes, DO Support fixed this for me. My ticket number was 556467.

Thanks, I opened new ticket so that they can fix it before I make next upgrade :)

Have another answer? Share your knowledge.