A "piped log program '(null)' failed unexpectedly" error message for each configured piped log is reported when you do an apache restart. Looking at the source, this is expected on a restart. Logging it gives the impression something is breaking. Attached will be a small patch to prevent this from being logged on a restart.
Created attachment 7328 [details] Remove piped log program '(null)' failed unexpectedly error on apache restart
FYI... this can happen at apachectl stop too*... another PR, another patch which should close this hole as well :) incidentally, part of this other patch is some infrastructure to get your marvelous cgid restart patch working with prefork, which in turn will let us merge it to stable *whether any of this happens seems to be timing related... number of piped loggers, OS, MPM, each are inputs to a function I cannot understand that determines if this bogosity occurs
Fix committed to 2.1-dev. Hopefully it can be committed to stable branch before too terribly long.