Just something basic to test :-)
This command works right?
# nslookup agent.example.com
If not does?
# ping agent.example.com
Where is this occurring?
Is it an existing server that has suddenly stopped working or a new build that is haveing the issue?
Is that message just standardout or is it out of the agents log file?
nslookup agent.example.com and ping work. I am run out of places where agent can get hostname with ].
sorry. can't reproduce problem today.
What’s in the /usr/lib/rsc/secure file?
We don't set hostname in /usr/lib/rsc/secure
# cat secure
I don't think that problem was in BladeLogic agent configuration. Agent is usually installed using silent install file and if I remember correctly person who installs agent does not have to enter hostname for agent.
I've checked uname, hostname, /etc/resolv.conf, dns lookup, ip reverse lookup and CN in certificate.pem and couldn't find the place that has domain name with ].
Why agent.example.com is not 127.0.0.1? is it not supposed to be the localhost?
"agent.example.com" is not localhost. It is some real name and IP address. I wouldn't scramble IP address information, if it was 127.0.0.1
Have you got other servers running AIX agents?
I know there is an issue on BL 8 AIX agents which prevents agents from starting up if there is a file (I believe in the agent home directory) called open_ssl.conf.
Don't know if this applies to BL 7.x agents, but if that file is there, try and rename it and try again
other agents talk to bladelogic appserver. This problem happened only on one agent. I've tried to diagnose, but I am not AIX guy (generic Unix/Linux background) and maybe I haven't looked at all possible places where AIX can store domain information.
One thing you can do to troubleshoot is open the /etc/rc.d/init.d/rscd (or /etc/rc.tcpip) and execute all the steps manually.
Basically this is about setting the environment then running the rscd binary.
It may give you more information
(BTW: agents don’t talk to the appserver. It’s the other way round)
PS: I still don’t understand why the agent should try and resolve that remote hostname when starting…
agent.example.com is the hostname of the server the agent is installed on, correct?
you can set the hostname in the secure file if you want to bind the agent to a specific ip, but you are not doing that.
1 of 1 people found this helpful
I had the same problem few minutes ago, and I have a solution.
In /etc/hosts short name must be equal to FQD Name for example:
10.10.3.4 host2 host2.domain
You can check it by commands:
hostname - returns short name
hostname -f - returns fqdn
In my case some one put something like this:
127.0.0.1 host222 localhost.domain localhost
So I added line
and corrected line
127.0.0.1 localhost localhost.domain localhost
Now the agent starts without problems.
On linux the hostname of the box (usually set in /etc/sysconfig/network) must also have an entry in /etc/hosts I believe.
your problem seems to occurs because your hostname is not well parametered : run hostname, then uname -n , you should have 2 different results. Correct this and it should work
i've got exactly the same problem on a HPUX 11.11:
root@foo:/exec/products/bladelogic/NSH/bin# ./rscd: Unknown host: 'foo' 'foo'
root@foo:/etc# cat hosts
127.0.0.1 localhost loopback
10.170.65.9 foo foo.domain.com
root@foo:/etc# cat hostname
Does anyone has an issue ?