If CloudFlare does not support zips and exes they should not be interrupting download streams when they do a cache refresh—but that is what seems to be happening.
Technically I interpret “do not support” as “do not handle correctly”. The ZIP and EXE files from DAZ are passing through CloudFlare but CloudFlare has no Content-Type parameter in the HTTP header for these files which may be part of the problem.
But I can see now that DAZ has directed downloads back to their own servers again so maybe things are back to normal now.
If one could manage to get the whole file down before the cache went for a refresh then there were no problems. Using a ultra high speed connection I was able to bring down several hundred megabytes of content. I timed it like this: Start a download (any file), as soon as it fails start the download of the file I wanted. Using this method I was able to bring down files >200MB.
From my research, it was definitely CouldFlare. Also, CloudFlare has been having issues of their own the last week. So unsupported content+cache refresh intervals set too close together+internal network problems = poor user experience.
BTW, services like CloudFlare are supposed to keep users from seeing “page not found” errors from bad connections. If one is using a caching service and users are getting pages like those reported, the failure is on the part of the cache provider.
Kendall
- old-school mainframe sysprog retooled to a unix (Aix) sysadmin, so I’ve no clue on how you’re doing the html diagnosis. Could you point me to a viable tutorial? (And apologies to the rest for the off-topic here!)
