[3074] in testers
/etc/named.local; /etc/vfstab; desync
daemon@ATHENA.MIT.EDU (Albert Dvornik)
Sat Apr 26 00:15:35 1997
To: testers@MIT.EDU
From: Albert Dvornik <bert@MIT.EDU>
Date: Sat, 26 Apr 1997 00:15:26 EDT
The first thing I noticed today while playing with si is that
"standard" named.local hacks no longer work with the new named,
because it actually paid attentionh when it was told named.localhost
(which $INCLUDEs named.local) describes the domain "localhost", and
will ignore all records outside of that domain. This is sort of
documented by a paragraph in named.localhost; however, it should be
documented in the release documentation, since I believe this is the
first occurence of a "clever" named on Athena.
Next, I re-noticed an old issue that I don't think I never mentioned
before. On each boot, Athena Solaris machines complain about trying
to mount ufs twice. This is because the partition marked "usr" in the
partition table gets automatically mounted before mountall is invoked,
and that partition is also listed as mount-at-boot "yes" in
/etc/vfstab, so mountall tries again. I think the warning message is
aesthetically displeasing. I have turned mount-at-boot for /usr in
vfstab to "no" ages ago, with no problems at all; can we make the
change in the release too?
Finally, and most seriously, auto_update had trouble when I tried
playing with it because it couldn't find /srvd/etc/athena/desync,
which seems to be missing in 8.1. This is caused by what I presume to
be a typo in athena/etc/desync/Makefile.in:
${INSTALL_PROGRAM} -m 555 desync ${DESTDIR}${lsbindir}/rep
This cause desync to be installed as /srvd/etc/athena/rep , since this
doesn't already exist as a subdirectory. Since the man page is still
installed as desync.8, I'd wager this was a misplaced cut-and-paste or
something. =) _In related news, I think there may be another typo in
configure.in for desync, whych says:
test -z "$lsbindir" && lbindir='${sbindir}'
Since lbindir doesn't occur anywhere else in configure, this may want
to be setting "lsbindir". Then again, I may be confused.
--bert
