I frequently see a situation where a file will show up as corrupted (upper-case name in download folder). When I manually disable the primary server and re-download the same corrupted file entirely from the backup server it is 100% correct. So it appears as if some articles were corrupted on the primary server and were not downloaded again from the backup. I think the backup servers are currently only used for missing articles and not for fixing corrupted primary server articles.
Can we add an optional feature to try re-downloading corrupted articles from backup servers?
