Operating System - HP-UX
1748051 Members
4905 Online
108757 Solutions
New Discussion

Re: serviceguard commands not working ,but applications are ok

 
Unix-team3
Visitor

serviceguard commands not working ,but applications are ok

Hello ,

None of the cluster commands are working , 

# cmviewcl
Failed to load data from configuration database.
Failed to load data from configuration database.
#

But the applications which are part of cluster are still up and running . 

It seems only the cluster commands are affected .

Should we restart inetd , ? will it impact running applications ? 

Any support is much appreciated.

 

 

 

 

 

4 REPLIES 4
KishJ
HPE Pro

Re: serviceguard commands not working ,but applications are ok

Greetings,

 

First of all, we need to identify why SG commands are hanging. Restarting inetd will help if the cause has to do with communication. 

I would suggest adding "-l" to the following line in /etc/inetd.conf file

 

auth stream tcp6 wait bin /usr/lbin/identd identd -l

 

And then, "inetd -c" will re-read the file. 

Please check system logs for more details on what is casing the failure.

 

You may also check for previous commands that may still be running rendering other commands to wait on them.

 

The following syntax will be useful to list processes in the order of parent-child

 

# export UNIX95=1; ps -AHflx

 

On a side note. here is an excerpt that will be helpful for future use:

 

If the -T option to identd is available on your system, you should set it to 120 (-T120); this ensures that
a connection inadvertently left open will be closed after two minutes. In this case, the identd entry
in /etc/inetd.conf should look like this:
auth stream tcp6 wait bin /usr/lbin/identd identd -T120
Check the man page for identd to determine whether the -T option is supported for your version of
identd

 

Hope it helps.


I am a HPE Employee

Accept or Kudo

KishJ
HPE Pro

Re: serviceguard commands not working ,but applications are ok

Greetings,

I hope my suggestions helped. If the SG commands continue to hang even after killing the old(read pending) "cm" commands, then it is better to stop the current inetd and restart using:

 

# inetd -k

# inetd -l

 

Please let us know how it goes.


I am a HPE Employee

Accept or Kudo

Unix-team3
Visitor

Re: serviceguard commands not working ,but applications are ok

Hello ,

   Thanks for the  suggestions . 

Query:        

As the Database and applications part of cluster packages  are currently running , will this restart of inetd make any impact to the these running databases and applications . ?

Also the following entry is there for identd:

auth stream tcp6 wait bin /usr/lbin/identd identd -t120

 

As per man page for identd ,

WARNINGS
The -w and -t options are currently not supported on HP-UX.

We have t120 , but i think it should be  T120 .

 

 

KishJ
HPE Pro

Re: serviceguard commands not working ,but applications are ok

Greetings,

 

Stop and restart of inetd service will not impact the existing connections. It's behavior is well explained in the man page, here is an excerpt:

 

      The inetd daemon starts servers for both stream and datagram type
      services.  For stream services, inetd listens for connection requests
      on Internet stream sockets.  When a connection is requested for one of
      its sockets, inetd decides which service the socket will support,
      forks a process, invokes an appropriate server for the connection, and
      passes the connected socket to the server as stdin and stdout.  Then
      inetd returns to listening for connection requests.

      For datagram services, inetd waits for activity on Internet datagram
      sockets.  When an incoming datagram is detected, inetd forks a
      process, invokes an appropriate server, and passes the socket to the
      server as stdin and stdout.  Then inetd waits, ignoring activity on
      that datagram socket, until the server exits.

 

I hope it helps. Please do not forget to post the results/outcome.


I am a HPE Employee

Accept or Kudo