[exim] Bad file descriptor

I'm not sure what I did, but I broke my exim install. It was working
fine, till suddenly it wasn't. I was having a problem where it
couldn't setuid and gid, but that was because the binary lost it's
setuid bit, again, I have no idea how/why.

My current problem is when mail is received, it is not delivered
unless explicitly do exim -M <message-id>. If I don't it just sits in
the queue, not-frozen. This occurs only for local deliveries, relayed
mail goes right through. There are no errors in the
/var/log/exim/main.log, /var/log/exim/panic.log, but there is the
below error in /var/log/maillog, a log file I've just noticed
existed..

May 2 16:17:29 slicer exim[15304]: 2007-05-02 16:17:29
1HjLBp-0003xI-Tv failed to write to main log: length=114 result=-1
errno=9 (Bad file descriptor)
May 2 16:17:29 slicer exim[15304]: write failed on panic log:
length=117 result=-1 errno=9 (Bad file descriptor)
May 2 16:17:29 slicer exim[15305]: 2007-05-02 16:17:29
1HjLD0-0003xo-VR failed to write to main log: length=114 result=-1
errno=9 (Bad file descriptor)
May 2 16:17:29 slicer exim[15305]: write failed on panic log:
length=117 result=-1 errno=9 (Bad file descriptor)
May 2 16:17:29 slicer exim[15301]: 2007-05-02 16:17:29 failed to
write to main log: length=45 result=-1 errno=9 (Bad file descriptor)
May 2 16:17:29 slicer exim[15301]: write failed on panic log:
length=99 result=-1 errno=9 (Bad file descriptor)


The error occurs on startup, and maybe periodically while exim is
running, but I'm not sure about that.

As much as I enjoy having to coax messages through, a solution to this
would be helpful.

I'm using CentOS with the exim from the base repository, version 4.43-1.RHEL4.5.

Thanks,

Andrew