Bug 211083

okay. btw. "service named reload" should work for you.

Could you attach content of /var/log/messages when named fails to restart?

There's not a lot;

Oct 17 10:33:48 echo named[1751]: rejected command channel message from
127.0.0.1#35066

The reload option does work however.

Could you please attach your /etc/named.conf?

Created attachment 138682 [details]
named.conf

My named.conf. Should just be the default.

I am experiencing the same problem on my FC5 box after updating to the latest
version of bind packages. And I also have to manually restart named due to bug
#206604. I also have the default named.conf, the same one as Simon.

Could you please run "/usr/sbin/rndc stop" and attach an output here?

Created attachment 139526 [details]
Output of  "/usr/sbin/rndc stop" command

Here is the output of 'sudo /usr/sbin/rndc stop' on my FC5 box.
I hope it helps.

(In reply to comment #7)
> Created an attachment (id=139526) [edit]
> Output of  "/usr/sbin/rndc stop" command

I don't see any problem here. What is return code of the "/usr/sbin/rndc stop"
command? ($? shell variable). It should be zero...

Have you got any messages on console?

The return code of command "sudo /usr/sbin/rndc stop" is zero and I don't get
any messages at the console (as expected).

However, running "sudo /sbin/service named stop" returns error code 1 and I get
the "Stopping named: [FAILED]" error message at the console.

Please check the latest test update for FC5 (bind-9.3.3-0.2.rc2.fc5).

Hi!

In the mean time, I updated my system to FC6 which now has bind-9.3.3-4.fc6
(from updates), but it still has the same problem with stopping named.

There seems to be a mistake, because updates-testing repo for FC6 contains the
packages of bind for FC5 (see
http://download.fedora.redhat.com/pub/fedora/linux/core/updates/testing/6/i386/).
So, I think I should wait for a new test update for FC6 (perhaps version
30:9.3.3-6 from cvs?)

Update for FC6 is on the way.

I updated bind to version 9.3.3-6 today and now I get the following error:

[tadej@tlinux-stable ~]$ sudo /sbin/service named restart
Password:
Stopping named: ..................................................no response,
killing with -TERM                                          [  OK  ]
Starting named: named: already running                     [FAILED]
[tadej@tlinux-stable ~]$ sudo /sbin/service named reload
Reloading named:                                           [FAILED]
[tadej@tlinux-stable ~]$ sudo /sbin/service named start
Starting named:                                            [  OK  ]
[tadej@tlinux-stable ~]$

Stopping named takes a very long time, but it has to be killed with -TERM, then
immediately starting named also fails. Manually staring afterwards works OK.

The relevant entries in /var/log/messages:
Nov  7 17:52:49 localhost named[1523]: rejected command channel message from
127.0.0.1#43260
Nov  7 17:54:32 localhost named[1523]: shutting down
Nov  7 17:54:32 localhost named[1523]: stopping command channel on 127.0.0.1#953
Nov  7 17:54:32 localhost named[1523]: exiting
Nov  7 17:55:10 localhost named[2511]: starting BIND 9.3.3rc3 -u named -D
Nov  7 17:55:10 localhost named[2511]: found 1 CPU, using 1 worker thread
Nov  7 17:55:10 localhost named[2511]: loading configuration from '/etc/named.conf'
Nov  7 17:55:10 localhost named[2511]: listening on IPv4 interface lo, 127.0.0.1#53
Nov  7 17:55:10 localhost named[2511]: listening on IPv4 interface eth0,
192.168.0.50#53
Nov  7 17:55:10 localhost named[2511]: command channel listening on 127.0.0.1#953
Nov  7 17:55:10 localhost named[2511]: zone 0.in-addr.arpa/IN: loaded serial 42
Nov  7 17:55:10 localhost named[2511]: zone 0.0.127.in-addr.arpa/IN: loaded
serial 1997022700
Nov  7 17:55:10 localhost named[2511]: zone 255.in-addr.arpa/IN: loaded serial 42
Nov  7 17:55:10 localhost named[2511]: zone
0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.ip6.arpa/IN:
loaded serial 1997022700
Nov  7 17:55:10 localhost named[2511]: zone localdomain/IN: loaded serial 42
Nov  7 17:55:10 localhost named[2511]: zone localhost/IN: loaded serial 42
Nov  7 17:55:10 localhost named[2511]: D-BUS dhcdbd subscription disabled.
Nov  7 17:55:10 localhost named[2511]: D-BUS service enabled.
Nov  7 17:55:10 localhost named[2511]: running
Nov  7 17:55:35 localhost NetworkManager: <WARNING>      add_ip4_config_to_named
(): Could not set forwarders for zone '.'.  Error: 'Did not receive a reply.
Possible causes include: the remote application did not send a reply, the
message bus security policy blocked the reply, the reply timeout expired, or the
network connection was broken.'.

The first message seems to indicate there is still a problem with permissions
for stopping named (I'm not using SELinux, if that has anything to do with it).

Could you please check it as a root? (not under sudo)

Hmm, something strange is going on here.

If I try 'service named restart' right after Fedora boots up, it seems to hang
trying to stop named. The same thing happens for root and sudo.

If I execute 'service named reload' (as sudo) it finishes immediately and the
subsequent executions of command 'service named restart' work immediatelly (both
for root and sudo).

I don't know what is 'special' about the first attempt to stop (and restart) the
named service, because subsequent attempts (after waiting for a very long time
for named to stop or executing 'service named reload' in between to 'fix' the
issue) to restart named work flawlessly.

Note: I have to restart named after startup due to bug in NetworkManager:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=206604 (I'm waiting for
your patch to be applied to NetworkManager).

I had similar problem, and solved by adding:

        inet ::1 allow { localhost; } keys { rndckey; };

to /etc/named.conf, insite the controls directive. Note the IPv6 address. This
seems to be caused due to ::1 listed as the localhost address in /etc/hosts.

It should be fixed in bind-9.3.3-0.1.rc3.fc6 / bind-9.3.3-0.1.rc3.fc7.