In #apache on FreeNode IRC, mod_unique_id is one of the most common problems. Thanks to some distributions (ahem. gentoo.), that commonly do not configure DNS correctly, and enable mod_unique_id by default, several times a day people come in saying apache will not start for them.
I see three ways to solve this issue: 1) Make the error we spit out more verbose when DNS is broken. 2) Continue running, turning off mod_unique_id. 3) Complain to upstream vendors.(don't enable mod_unique_id by default!)
Most of the people with this problem have no reason to be running mod_unique_id, but their distro has enabled it, so I believe the best method is to spit out info about it in the error log, but to keep Apache running.
writing to the error log isn't completely quiet, but could very well go unnoticed, which could then lead to applications failing, possibly much later than the server start/restart...
this matters not to somebody who doesn't need mod_unique_id in the first place, but more to people with reasonable if not failsafe DNS setup who hit a glitch at web server restart time
let folks decide which camp they are in when the server fails to start because some feature which has been enabled won't possibly work; maybe a discussion in the documentation and a pointer to it in the error message would be the right solution
(learning to yank an unnecessary module or fix system DNS setup are both reasonable chores)
