An interrupted Yum update has left me with a production server that won’t reboot. An ordinary reboot command via SSH and command line, leaves it hanging and unresponsive. Using the digitalocean recovery console to reboot gives an error of “Error invalid signature..... press any key to continue”.

Diagnosing the cause is tough since this happens before logs are being written. It is possible to reboot successfully from the recovery console by issuing a “shutdown -r now” and then selecting a kernel (all work) from the grub menu.

Things I’ve tried are:

1) Reinstalling grub as detailed here: https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html/system_administrators_guide/sec-reinstalling_grub_2

2) Created a new initrd: https://wiki.centos.org/TipsAndTricks/CreateNewInitrd

There are full backups of the server and LAMP apps on it so I could just move sites to a new droplet created from a pre-Yum interruption but since it’s a running system, I’d prefer to repair.

Any suggestions would be deeply appreciated.