Here's another one - Solaris had a 2GB file size limit (this was before ZFS). Which isn't as crazy as it sounds now - hard drives were 9GB at the time. So ordinarily this wasn't a problem, but when the first Harry Potter movie came out, harrypotter.com (which was being served off the same server as spacejam.com) was the most popular website on the internet and the web server log would hit the 2GB limit every couple hours and we would have to frantically move it somewhere and restart the process.
Most likely the main Oracle server which had a disk array attached. These days it's easy to forget how tricky it could be to get basic things like this working. rsync was in its infancy and I doubt we were using it. As I recall most servers didn't have ssh installed; telnet was standard. Lots of tar piped over rsh to get files from A to B.
7-8 years ago I've encountered a full-on rsh internal network in a fairly big datacenter. No rsync, rcp. It turned out that the sysadmins were too lazy to set up ssh keys on all the servers, and their manager was skilled at deflecting issues.
Here's another one - Solaris had a 2GB file size limit (this was before ZFS). Which isn't as crazy as it sounds now - hard drives were 9GB at the time. So ordinarily this wasn't a problem, but when the first Harry Potter movie came out, harrypotter.com (which was being served off the same server as spacejam.com) was the most popular website on the internet and the web server log would hit the 2GB limit every couple hours and we would have to frantically move it somewhere and restart the process.