Description of problem: During setting up CTDB, we need to create /etc/ctdb/nodes file on all the nodes that is used as Samba servers and add the ips of these nodes to the file. E.g.: $cat /etc/ctdb/nodes 10.16.157.0 10.16.157.3 10.16.157.6 10.16.157.9 But if there is any white space after the IP addresses, ctdb fails to parse it and goes to unhealthy state. e.g: 10.16.157.0<white space> 10.16.157.3<white space> I can see below in /var/log/log.ctdb 2014/07/04 15:24:19.244686 [ 2697]: Starting CTDBD (Version 2.5.3) as PID: 2697 2014/07/04 15:24:19.558658 [ 2697]: common/ctdb_util.c:455 Failed to translate 10.70.43.93 into sin_addr 2014/07/04 15:24:19.558718 [ 2697]: tcp/tcp_connect.c:118 Unparsable address : 10.70.43.93 . 2014/07/04 15:24:19.559087 [ 2697]: Freeze priority 1 2014/07/04 15:24:19.561197 [ 2697]: Freeze priority 2 2014/07/04 15:24:19.562124 [ 2697]: Freeze priority 3 2014/07/04 15:24:19.587347 [ 2697]: common/ctdb_util.c:455 Failed to translate 10.70.43.93 into sin_addr 2014/07/04 15:24:19.587389 [ 2697]: tcp/tcp_connect.c:118 Unparsable address : 10.70.43.93 . 2014/07/04 15:24:19.834621 [ 2697]: common/ctdb_util.c:455 Failed to translate 10.70.43.93 into sin_addr 2014/07/04 15:24:19.834673 [ 2697]: server/ctdb_recover.c:145 Failed to parse 10.70.43.93 into a sockaddr Version-Release number of selected component (if applicable): ctdb2.5-2.5.3-5.el6rhs.x86_64.rpm glusterfs-3.6.0.22-1.el6rhs.x86_64.rpm How reproducible: Always Steps to Reproduce: Follow below documentation and while putting ips in /etc/ctdb/nodes, put white space after one of the ips https://documentation-devel.engineering.redhat.com/site/documentation/en-US/Red_Hat_Storage/3/html/Administration_Guide/sect-Configuring_Automated_IP_Failover_for_NFS_and_SMB.html#Setting_Up_CTDB Actual results: In "ctdb status" the heath of the cluster was unhealthy Expected results: CTDB should be intelligent enough to ignore the white space. Additional info:
Created attachment 914706 [details] ctdb logs
The issue is still seen with: ctdb2.5-2.5.4-1.el6rhs.x86_64 2015/05/04 06:33:29.910685 [24680]: common/system_util.c:154 Failed to translate 10.16.157.78 into sin_addr 2015/05/04 06:33:29.910708 [24680]: tcp/tcp_connect.c:118 Unparsable address : 10.16.157.78 . 2015/05/04 06:33:29.910728 [24680]: Unable to bind to any of the node addresses - giving up 2015/05/04 06:33:29.910741 [24680]: tcp/tcp_init.c:67 Failed to start listening on the CTDB socket
Patch exists by now in upstream samba master. Need backport to ctdb 2.5.
Verified the BZ on RHEL7. Waiting for RHEL6 build.
On RHEL6 verified the bz with following steps: While doing the setup for CTDB , introduce a whitespace after the IP entry in /etc/ctdb/nodes files and save the config an do ctdb start, Expected result: CTDB should start successfully. Tried with multiple spaces , it works fine. Verified on RHEL6 and RHEL7 with builds as follows: ctdb-4.2.4-8.el6rhs.x86_64 : RHEL6 ctdb-4.2.4-6.el7rhgs.x86_64 : RHEL7
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHBA-2016-0324.html