Corrupt initramfs file?

[garyw]

Hi All,

I hit a weird issue yesterday. I rebooted one of our CentOS 6.5 servers and it promptly kernel panicked with a RAMDISK incomplete write! error.
Booting into rescue mode I could see that the initramfs-2.6.32-431.el6.x86_64.img file was half the size it should be (attached here, I've renamed the broken file to .backup).

Copying the file from another linux box and rebooting fixed the issue but I'd love to know why that file got damaged/corrupted in the first place. The /boot volume has 396MB free space so I don't believe it to be a space issue.

Thanks for any suggestions!

recovery.PNG (7.86 KiB) Viewed 2063 times

[tunk]

I don't know what caused it, but I think I would see if there's any disk problems
and if there's been any (unscheduled) reboots. Also, CentOS 6.5 has six years of
accumulated security problems; it's advisable to update.

Edit: You could also boot from a live CD and run a file system check.

[TrevorH]

Apparently you did something to regenerate all initramfs files yesterday and that was what corrupted this one. I'd check your root .bash_history and see what was done.

Also, as alluded to by tunk, you are way way out of date and severely insecure. The kernel you are complaining about not booting is from 6.5 which dates from 2013. The newer -573 kernel you have listed there is also ancient but marginally less so: it dates from 6.7 in 2015. This means that at a minimum you are missing 5 years of security fixes and possibly as much as 7.

You should be running yum update quite a lot more regularly than every 5 years. Both those versions have enough holes in them to drive a truck through so I really hope this is not exposed to the internet and that you really trust every single one of your users who has access to any internal network connected to that machine.

Oh, and lastly, all of CentOS 6 goes EOL in November this year and there will be no more updates and no support. You should be planning your migration to a supported release ASAP.

[garyw]

Also, as alluded to by tunk, you are way way out of date and severely insecure. The kernel you are complaining about not booting is from 6.5 which dates from 2013. The newer -573 kernel you have listed there is also ancient but marginally less so: it dates from 6.7 in 2015. This means that at a minimum you are missing 5 years of security fixes and possibly as much as 7.

You should be running yum update quite a lot more regularly than every 5 years. Both those versions have enough holes in them to drive a truck through so I really hope this is not exposed to the internet and that you really trust every single one of your users who has access to any internal network connected to that machine.

Oh, and lastly, all of CentOS 6 goes EOL in November this year and there will be no more updates and no support. You should be planning your migration to a supported release ASAP.

No arguments from me with any of that. This isn't my system so I have next to no say. The people that own it have plans to migrate the system but I am not privy to those plans.