Red Hat Bugzilla – Bug 216796
automount has no method to re-read resolver settings (resolv.conf)
Last modified: 2008-01-08 11:48:33 EST
Description of problem:
There is no way to make automount reread /etc/resolv.conf values short of
restarting it. At least a -HUP signal should facilitate this.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Update /etc/resolv.conf with new name server records (or from dhcp release)
2. Reload/restart nscd
3. pkill -HUP automount
automount processes still use old name server(s), even after -HUPing as
confirmed by strace
Changes to DNS resolver (commonly in use) ought to be realized, at least upon
signalling where restarting is not possible.
Restarting automount jobs require unmounting all managed filesystems, which
often isn't possible in distributed environments.
Created attachment 141856 [details]
Patch to automount.c which (naively) runs res_init() at map (re)read (assumes DNS employed)
This seems like a reasonable approach to me. Did you test it?
(In reply to comment #2)
> This seems like a reasonable approach to me. Did you test it?
Seems sensible to me as well.
I'll put this on the list for v5 as well.
Only thought would be whether this would affect autofs users not using DNS, as
this now pulls in libresolv directly instead of relying on nss via libc.
Otherwise seems to work for us.
OK, so this needs to be tested in an environment that uses nis for hostname lookups.
This is fixed in package version autofs-4.1.3-205.
This request was evaluated by Red Hat Product Management for inclusion in a Red
Hat Enterprise Linux maintenance release. Product Management has requested
further review of this request by Red Hat Engineering, for potential
inclusion in a Red Hat Enterprise Linux Update release for currently deployed
products. This request is not yet committed for inclusion in an Update
I cannot reproduce the problem that was reported by you. Can you give more
specific information about your test configuration? Here is what I did:
create /etc/resolv.conf with a single name server entry:
Start the automounter with a single map for /test, /etc/auto.test:
ls /test/foo succeeds as expected. Then, I change the server line in
and I drop all packets destined for the old server
iptables -A OUTPUT -d 10.12.32.11 -j DROP
And I trigger another lookup, and it succeeds. So, how exactly did you get this
Eric, I'm going to have to remove this change given that I can't reproduce the
This request was previously evaluated by Red Hat Product Management
for inclusion in the current Red Hat Enterprise Linux release, but
Red Hat was unable to resolve it in time. This request will be
reviewed for a future Red Hat Enterprise Linux release.
I'm still waiting for a reproducer. Unless one is provided, I'm going to close
this bug WONTFIX.
Given that I can't reproduce the problem and there's been no response from the
reporter for some time, I'm closing this bugzilla.