Posted 17 May 2019 - 05:58 AM
It's been ages since I was in charge of any Windows servers but I do recall this error and believe it had more to do with fragmented files, not necessarily the entire volume. Large files, like perhaps db.have, that are constantly written to can become heavily fragmented in short order, even when newly written/rewritten. The amount of free space on the volume will also adversely affect fragmentation.
I'm not entirely sure what the remedy is aside from defragmenting on a regular basis. But even then, the error might occur at different intervals depending on database activity and the above mentioned free space.
If you _really_ want to go whole hog with a solution, I'd recommend switching to a Linux server and not running Perforce on Windows at all. This is one of at least several problems you'll eventually run into with Windows/NTFS. It might feel unholy if you're a Windows development shop, to run a Linux server, but you won't regret it (probably.) There are some considerations for line endings and upper/lowercase preservation on the underlying filesystem (recommend XFS or ext4), but I've been supporting Mac, Windows and Linux users with exclusively Linux Perforce servers for over a decade and it's served me quite well.
Staff SCM Engineer, Perforce Administrator
Dolby Laboratories, Inc.
1275 Market St.
San Francisco, CA 94103, USA