Improve error handling for corrupt checkpoints

[jmintb]

Issue: #4025

Description

This PR improves handling of corrupt checkpoints by logging the error and skipping the corrupt partition instead of panicking.

I am not sure if this is what you had in mind so this is a draft PR.

Questions:

1. Is there any bookkeeping I am missing when skipping a partition?
2. What kind of information would you like to see in the error message? maybe the partition ID?
3. Do you have any thoughts on testing this?

Let me know what you think and I will apply the final version to the remaining panic cases :)

How was this PR tested?

Not tested yet :(