Unsolved
This post is more than 5 years old
44 Posts
0
1782
nsrexecd on NetWorker 8.2 RHEL 7 Client erroneously binds to ipv6 address
Hi Experts
Our Linux Engineers built up an Environment for Docker-Containers. On Default Docker is buildung up an IPv6 network for its communication needs.
I'm now unable to backup this Environment and found out that nsrexecd has bound its Services to ipv6.
vlx00198 ~]# netstat -tulpen | grep nsr
tcp6 0 0 :::8682 :::* LISTEN 0 37771 5109/nsrexecd
tcp6 0 0 :::7937 :::* LISTEN 0 38007 5109/nsrexecd
tcp6 0 0 :::7938 :::* LISTEN 0 37987 5109/nsrexecd
tcp6 0 0 :::8905 :::* LISTEN 0 37768 5109/nsrexecd
udp6 0 0 :::7938 :::* 0 37985 5109/nsrexecd
During Startup NetWorker has to decide which IP to use for its services.
- How does it tell which one?
- How can I influence this process?
Thanks for any hints!
Regards
Christian
Kleinenbroich
86 Posts
0
November 25th, 2016 07:00
Hi Cristian,
as far as I remember, the /etc/hosts plays an important role to decide which IP to use for its services: NW queries the /etc/hosts file and the first entry for the NetWorker server is used as name for the NetWorker sevrer (also listed as name in the NSR/NSRLA field (can be accessed via nsradmin on the NW server)). If the NW server has several IPs: From the first row with an IP of its own, the fist host name is used as NW server name. This name is send to the NW clients, storage nodes and so on.
One more thing (from the NW Installation Guide):
- The hosts file on each NetWorker host has an entry that associates the IPv6 loopback interface (::1) with the localhost. Add the IPv6 loopback interface entry before the IPv4 loopback entry (127.0.0.1 localhost)
For example:
::1 localhost
127.0.0.1 localhost.localdomain localhost
- The IPv6 loopback entry must remain in the hosts file when the host is operating in a pure IPv4, pure IPv6, or dual stack configuration.
Regards
Michael