Greetings

After a recent substantial update which included a kernel upgrade from

3.10.0-957.27.2.el7.x86_64

to

3.10.0-1062.1.1.el7.x86_64

I rebooted to 1062 and it seems to boot up OK but there is no connection to the outside world via eth0. loopback works. eth0 seems up ok. Anyway...

This is on a live server, so I didn't have much time to poke and prod. So I booted the previous kernel and public service returned... for now.

Which brings me to my question - what can I do to check what the problem may have been without rebooting into 1062 and feeling the pressure of uptime monitors messaging me and my phone going off with people wanting to know why XYZ isn't working.

I don't know if I've been lucky all these years, but this is the first time a kernel upgrade hasn't been seamless.

So far there don't seem to be any conversations about the newest kernel giving problems to anyone else, so it seems my config is "special" (oh yay...)

---
Added: Hmmm, this looks like it is in the ballpark :\
http://scientificlinuxforum.org/inde...=0&#entry21821

And:
There we go. I'm not alone. It's a good feeling
https://bugs.centos.org/view.php?id=16413

Have you a realtek network card? Is it old?

I had issues like that a long time ago. It turned out that the realtek didn't like the interrupt it was on, threw a hissy fit and wouldn't work. It wanted irq 11 but half the time the video would get there first and it would get irq 12. Normally these days, things are much more tolerant of various irqs.

I bodged my way out of it somehow (I forget) but I'm sure you're capable of checking that and sorting it.