Sha1 mismatch after hundreds of successful deployments

Having an issue… We have a fixlet that has been deployed for months. Suddenly we are seeing this error.

And the odd thing is, the file it is pointing to, has the correct SHA1

This has been an active deployment for months and has successfully deployed to 459 machines.

What am I missing?

I redeployed it and it appears to be working but I would like to know what may have caused this.

Did you copy the Action, or take a new action from the source fixlet/task? I wonder whether the fixlet was changed / hash updated to the new value?

I redeployed from the source fixlet.

Before I did anything, I checked the file, no changes, file had the same date as before, the sha1 matches what was in the fixlet.

I checked the created and last modified date of the fixlet, both had my user name, they were from the original deployment in Sept.

Feels to me like the cached file became corrupted?

Yeah I think that could be possible. At the root and relay, the downloads are cached in a directory based on action ID, so if one of those files was modified the relay might serve the modified version to clients (with the clients rejecting them due to the hash mismatch)

There are a dozen or more relay being used by the systems it failed on so the corruption must have been at the root.

The re-deployment seemed to work. They are all pending restart now.