I'm looking into a different report about "Internal apreq error" in "Apache2::Request::upload" due to a complaint received from a student and found 2 forum chains.
- https://webwork.maa.org/moodle/mod/forum/discuss.php?d=2673
- https://webwork.maa.org/moodle/mod/forum/discuss.php?d=183
The first chain reports that it was fixed by the ISP which was blocking some of the communication and once the stopped the blocking the problem disappeared.
However, I suspect that in most cases the problem is probably pretty rare and intermittent and caused by network latency issues which occur more or less at random, as there do not seem to be frequent reports of such issues anywhere.
I found a small number of such error messages in the Apache error.log and every one seen in the several months (the log file I have) is preceded by a line reporting "ap_get_brigade failed during prefetch".
[Log records edited to hide IP address details and full course paths]
Most of the log records (over 85%) are for a timeout, ex:
- [Fri Jan 24 11:38:39.242984 2020] [:error] [pid 319] (70007)The timeout specified has expired: [client ip_address:port] ap_get_brigade failed during prefetch, referer: https://webwork.technion.ac.il/webwork2/courseID/...
But two other formats were seen:
- [Thu Jan 30 18:47:22.658777 2020] [:error] [pid 3517] (70008)Partial results are valid but processing is incomplete: [client ip_address:port] ap_get_brigade failed during prefetch, referer: https://webwork.technion.ac.il/webwork2/html2xml
- [Sat Feb 08 12:49:49.159787 2020] [:error] [pid 2582] (104)Connection reset by peer: [client ip_address:port] ap_get_brigade failed during prefetch, referer: https://webwork.technion.ac.il/webwork2/html2xml