CTBD don't start when create soft link on CentOS 7

[Martin Schwenke]

On Wed, 14 Jun 2017 13:50:21 +0700, GiangCoi Mr via samba-technical
<samba-technical at lists.samba.org> wrote:

> Hi team.
> 
> I have a issue when I configure CTDB with Samba 4 on CenOS 7. This is my
> configuration:
> 
> ----------------------------------------------
> vim /data/lock/ctdb
> 
> CTDB_RECOVERY_LOCK=/data/lock/lockfile
> #CIFS only
> CTDB_PUBLIC_INTERFACE=ens32
> CTDB_PUBLIC_ADDRESSES=/etc/ctdb/public_addresses
> CTDB_MANAGES_SAMBA=yes
> #CIFS only
> #
> CTDB_NODES=/etc/ctdb/nodes
> -----------------------------------------------
> vim /data/lock/nodes
> 172.17.4.10
> 172.17.4.20
> -----------------------------------------------
> vim /data/lock/public_addresses
> 10.10.5.92/24 ens32
> -----------------------------------------------
> 
> I create soft link on file01 and file02
> 
> ln -s /data/lock/nodes /etc/ctdb/nodes
> ln -s /data/lock/public_addresses /etc/ctdb/public_addresses
> 
> after I start ctdb, but it's not running, this is my log.ctdb
> 
> 2017/06/14 12:27:25.690819 [ 1068]: CTDB starting on node
> 2017/06/14 12:27:25.712588 [ 1068]: Recovery lock file set to "". Disabling
> recovery lock checking
> 2017/06/14 12:27:25.712826 [ 1068]: Failed to read nodes file
> "/etc/ctdb/nodes"
> 2017/06/14 12:27:25.712836 [ 1068]: Failed to load nodes file
> "/etc/ctdb/nodes"
> 2017/06/14 12:35:33.627848 [ 2614]: CTDB starting on node
> 2017/06/14 12:35:33.627916 [ 2614]: Recovery lock file set to "". Disabling
> recovery lock checking
> 2017/06/14 12:35:33.628818 [ 2614]: Failed to read nodes file
> "/etc/ctdb/nodes"
> 2017/06/14 12:35:33.628836 [ 2614]: Failed to load nodes file
> "/etc/ctdb/nodes"
> 2017/06/14 13:10:46.830114 [ 2657]: CTDB starting on node
> 
> But if I copy file from /data/lock/nodes to /etc/ctdb, it's running but
> have not virtual IP. please help me to fix it. Thanks so much

Symlinking the nodes and public_addresses files into the cluster
filesystem works for me on RHEL 6.9 with SELinux disabled.

2 possibilities I can think of:

* SELinux is enabled and is not allowing ctdbd to read the
  configuration file from your cluster filesystem

* systemd is somehow being "helpful"  ;-)

peace & happiness,
martin