/sbin/init.d processes again present after reboot

[federico_3]

Hi to all,

Running ps -ef| grep oracle, i see :

root 1573 142 0 Nov 19 console 0:00 /sbin/rc4.d/S850oracle /sbin/rc4
.d/S850oracle start

root 1738 1573 0 Nov 19 console 0:00 remsh casapp -n /sbin/init.d/ora
cle_cmwh start

As you can see, it seems that the remsh doesn't terminate.
I have gone to casapp and looked for processes that refer to /sbin/init.d/oracle_cmwh but all seems OK and system casapp was up and running at the time the remsh command started.

On casapp i found:
root 2135 834 0 Nov 26 ? 0:00 remshd
roort 2136 2135 0 Nov 26 ? 0;00

and the oracle_cmwh is ......


Thanks

Federico

[John Palmer]

Hi federico,

This one was discussed at length some time ago, do a search in the forums for the full details.

What causes the problem is that the standard input and output of the remsh process are inherited by any process that it forks. So if you for instance start an oracle database (or any daemon type process) then the remsh will 'hang' until that cheild process terminates.

Try redirecting standard input of remsh and the child processes to /dev/null and standard output/error to either a relevant log file or /dev/null.

Hope this helps,
John

[Vincente Fernandes]

Oracle init process are hanging because remsh command either didn't terminate or the scripts which is running didn't complete. The remsh command is getting hunged on the other system. You also need to have a look at the /sbin/init.d/oracle_cmwh init script.
You can try running from machine 1 the following:
#remsh machine2 "bdf or /usr/sbin/lanscan" or any normal HP-UX command and if that works fine then the actual problem is with your oracle script.
Is it the first time it happened?

[Dan Hetzel]

Hi Frederico,

On your 'casapp' server, the zombie process most probably has a child still running (or hanging)

In your attached file, there could be a problem in the event that file /etc/rc.config.d/oracle_cmwh is missing.

As the error part of your 'if' doesn't terminate the script, the next 'if' statement is executed with a test for an undefined variable ($ORACLE_START), so the test will fail with a message "sh: ORACLE_START: Parameter not set."

Same for the 'stop' case.

Is file /etc/rc.config.d/oracle_cmwh existing ??

Best regards,

Dan

[federico_3]

The strange thing is that from another machine i start another Concurrent manager on the same casapp by using the remsh command in /sbin/init.d/oracle_cm ( the difference between the oracle_cm and oraclke_cmwh scripts is only the starting of the listener FNDFS, the rest is the same ).How come on the same machine do we have different bahaviours? Is it , maybe, the problem connected to the starting of listener FNDFS?Oracle is UP and running ( also on casapp where all seems OK a part this return code that does not return )but on the console of the machine in which starts the remsh to casapp the Oracle startup is blocked waiting for the return code from casapp.

Dan,

the file you are referring is present and the problem is after the starting of listener because it has been started.

federico

[Dan Hetzel]

Frederico,

Even if the file is present in this case, it would be a good idea to modify your script, just in case... ;-)

You're suspecting the problem to occur after the start of the listener because it's started.
Are you sure it's not during the startup of your listener ?

Could you try to stop the listener and restart it using exactly the same command as in your script.

(logged as root)
# su - oracWH -c "lsnrctl start listener_FNDFS"

Is it starting OK ? Are you getting your shell prompt immediately after, without intervention?
Or is there something in the .profile of user oracWH which prevents executing, like a prompt?

Best regards,

Dan

[federico_3]

Hey guys ,

the problem is connected to the command :
su - oracWH -c ""lsnrctl start listener_FNDFS"

I tried this command from other two machines and using another listener and i got the same problem:the listener starts but the remsh hangs.

It seems that the remsh and the starting of listener give problems( i tried to visualize the status of the listener using the remsh and it works)

how come?


Federico

[John Palmer · Accepted solution]

Hi,

Further to my previous post, have you tried:-

su - oracWH -c "lsnrctl start listener_FNDFS 2>&1"

I presume that the "" in your last post was a typo?

Regards,
John

[federico_3]

John,
your explanation seems to be the solution to the problem but only tomorrow i'll check it.
Now i would like you to provide me further info in order to let me know better the problem.
Can you explain better what is happening.. ?Sorry , but it's a strange thing that i have never seen ...


thank you very much

Federico