- backout master map lookup changes made to beta3.
- change default master map from /etc/auto.master to auto.master so that we
always use nsswitch to locate master map.
- change default installed master map to include "+auto.master" to pickup
NIS master map (all bz# 193831 again).
- add patch to remove extra debug print.
- add patch to
- fix memory alloc error in nis lookup module.
- add "_" to "." mapname translation to nis lookup module.
- add patch to add owner pid to mount list struct.
- add patch to disable NFSv4 when probing hosts (at least foe now).
- add patch to fix white space handling in replicated server selection
code.
- Add fix for premature return when waiting for lock file.
- Update copyright declaration for reentrant-syslog source.
- Add patch for configure option to disable locking during mount. But don't
disable locking by default.
- Add ability to handle automount schema used in Sun directory server.
- Quell compiler warning about getsockopt parameter.
- Quell compiler warning about yp_order parameter.
- allow -net instead of /etc/auto.net
- Fix a buffer overflow with large key lengths
- Don't allow autofs to unlink files, only to remove directories
- change to the upstream reentrant syslog patch from the band-aid deferred
syslog patch.
- Get rid of the init script patch that hard-coded the release to redhat.
This should be handled properly by all red hat distros.
process. If the mount completed first, the parent would not expire the
stale entry, leaving it first on the list. This causes map updates to
not be recognized (well, worse, they are recognized after the first
expire, but not subsequent ones). Fixes a regression, bug #137026
(rhel3 bug).
- Update our auto.net to reflect the changes that went into 4.1.4_beta2.
This fixes a problem seen by at least one customer where a malformed
entry appeared first in the multimount list, thus causing the entire
multimount to be ignored. This new auto.net places that entry at the
end, purely by luck, but it fixes the problem in this one case.
(such as those used for /net) could be processed in the wrong order,
resulting in directories not showing up in a multimount tree. The fix
is to process these directories in order, shortest to longer path.
auto_master. Addresses bz #130079
- When using an auto.smb map we now remove the leading ':' from the path
which caused mount to fail in the past. Addresses bz #147492
- Autofs now checks /etc/nsswitch.conf to determine in what order files &
nis are checked when looking up autofs submount maps which don't
specify a maptype. Addresses IT #57612.
bz #138606
- Return first entry when there are duplicate keys in a map. Addresses bz
#140108.
- Propagate custom map variables to submounts. Fixes bz #143074.
- Create a sysconfig variable to control whether we source only one master
map (the way sun does), or source all maps found (which is the default
for backwards compatibility). Addresses bz #143126.
- Revised version of the get_best_mount patch. (#146887) cfeist@redhat.com
The previous patch introduced a regression. Non-replicated mounts would
not have the white space stripped from the entry and the mount would
fail.
- Handle comment characters in the middle of the automount line in
/etc/nsswitch.conf. Addresses bz #127457.
