INIT: Command is respawning too rapidly. (2057 Views)
Reply
Trusted Contributor
Posts: 420
Registered: ‎09-19-2011
Message 1 of 9 (2,057 Views)

INIT: Command is respawning too rapidly.

Hello Experts,

 

I am getting an error on my rx6600 server running hp-ux 11.23. The error is

 

INIT: Command is respawning too rapidly.
Will try again in 5 minutes.
Check for possible errors.
id:  h1 "/sbin/init.d/init.cssd run >/dev/null 2>&1 </dev/null"

 

I am getting this error only on console not through ssh or telnet. No error logs in syslog or dmesg. The server is working fine. 1 disk got failed on this disk but after replacing it also getting same error.

 

Any help will be very much appreciated.

 

Regards

Syed

Honored Contributor
Posts: 6,271
Registered: ‎12-02-2001
Message 2 of 9 (2,055 Views)

Re: INIT: Command is respawning too rapidly.

There is a line in /etc/inittab that tells init to run the listed command and restart it if it stops for any reason. Now this command keeps dying almost immediately, and after multiple restart attempts, init's safety mechanism is limiting the restart attempts so that the system won't waste all its power trying to restart something that is apparently not going to work.

 

The /etc/inittab line would look like this:

h1:<runlevel numbers>:respawn:/sbin/init.d/init.cssd run >/dev/null 2>&1 </dev/null

The <runlevel numbers> part is probably some combination of numbers 2, 3, 4 and 5.

 

Googling for "init.cssd" reveals this might be related to Cluster Synchronization Service Daemon, a component of an Oracle Cluster or an Oracle RAC database.

Is the system supposed to be running something like that? If not, this might be a remnant of some previous software installation that has not been completely uninstalled, and you can safely remove the line from /etc/inittab.

 

Since the command in inittab is directing all its output to /dev/null, the only way to see any error messages it might produce is to run it manually without the I/O redirections. So please run this command:

/sbin/init.d/init.cssd run

It will probably produce an error message, which might offer some more clues about what is going wrong with it. 

 

If this system should be running an Oracle Cluster or a RAC database, you and your DBA probably should start troubleshooting it. (The database might currently be running just fine, but the failure of the Cluster Synchronization Service might make the cluster unable to recover automatically if something else fails.)

MK
Trusted Contributor
Posts: 420
Registered: ‎09-19-2011
Message 3 of 9 (2,039 Views)

Re: INIT: Command is respawning too rapidly.

Hi MK,

 

Many thanks for such a detail reply. Sorry for late response as I was not availaible.

 

I tried to find the error message by running the command:

/sbin/init.d/init.cssd run

 

But not able to find any O/P. Please help

 

Regards

Syed

 

 

 

Honored Contributor
Posts: 6,271
Registered: ‎12-02-2001
Message 4 of 9 (2,035 Views)

Re: INIT: Command is respawning too rapidly.

Hmm... further googling indicates init.cssd is used with Oracle ASM too. We have some servers with ASM, so I took a look at /sbin/init.d/init.cssd.

 

Apparently any logs would be written to $ORA_CRS_HOME/log/$HOST/cssd directory.

$ORA_CRS_HOME is the Oracle binary installation directory: if you don't know what it is in your environment, you should ask your DataBase Administrator. $HOST is simply the hostname of your server.

 

Within that directory, you should find some log files: in our environment, there was "cssdOUT.log" and "ocssd.log".

 

If you cannot find them, please run this command to find the correct pathname:

find / -name ocssd.log -type f

 

MK
Trusted Contributor
Posts: 420
Registered: ‎09-19-2011
Message 5 of 9 (2,028 Views)

Re: INIT: Command is respawning too rapidly.

Hi MK,

 

Thanks for the reply.I will try to work on this with my DBA. Have one more query. We have RAID mirroring on all disk. From the day I am getting this error on my console I am also getting Amber LED on my 2 disk. From syslog, dmesg, diskinfo etc.. no error regarding the disk.

Is this amber LED because of this error?

Any help will be very much appreciated.

 

regards

Syed

Valued Contributor
Posts: 597
Registered: ‎05-24-2004
Message 6 of 9 (2,018 Views)

Re: INIT: Command is respawning too rapidly.

I met such issue when there was file system problem, maybe you could check out whether there was disk or file system error.

 

regards

Stephen

Trusted Contributor
Posts: 420
Registered: ‎09-19-2011
Message 7 of 9 (2,012 Views)

Re: INIT: Command is respawning too rapidly.

Hi Mk,

 

Thanks for the reply. There is no directory under oracle like ORA_CRS_HOME. Also when i try to find cssd with comand

 

find / -name ocssd.log -type f

 

it gives no output.

 

As per your earlier support since this server is running simple Oracle10g no RAC or Oracle Cluster I have commented out the line in /etc/inittab

## h1:3:respawn:/sbin/init.d/init.cssd run >/dev/null 2>&1 </dev/nul

Will this change effect the application or server anyway?

 

Thanks for your continous support.

 

regards

Syed

Trusted Contributor
Posts: 420
Registered: ‎09-19-2011
Message 8 of 9 (2,011 Views)

Re: INIT: Command is respawning too rapidly.

Hi Stephen,

 

Thanks for the reply. Where to check the error regarding file system or disk as I found no error in syslog or dmesg or diskinfo.

 

Regards

Syed

Honored Contributor
Posts: 6,271
Registered: ‎12-02-2001
Message 9 of 9 (2,000 Views)

Re: INIT: Command is respawning too rapidly.

ORA_CRS_HOME is an environment variable. If your oracle user account (usually named "oracle" but may be different) has this variable set, the variable will identify the correct directory.

 

Assuming your oracle user account is named "oracle", these commands should display the environment variable if it exists.

su - oracle
echo $ORA_CRS_HOME
exit

 

If the variable does not exist, you will get no output from the "echo $ORA_CRS_HOME" command. That probably means the product was already removed. If the variable exists but the directory does not exist, it probably means the same thing: the previous sysadmin has already uninstalled the Oracle product (RAC, Cluster or ASM) that caused the /etc/inittab line to be added, but for some reason forgot to remove all the parts.

 

The absence of the ocssd.log file also suggests to me that the CSSD software component has probably already been removed.

 

In that case, commenting out the /etc/inittab line is safe: you're just completing an earlier software uninstallation that was done improperly.

 

The "INIT: Command is respawning too rapidly" message indicates the /sbin/init.d/init.cssd script already cannot do its job properly: if the CSSD was required by Oracle or some application, any possible harm to applications is already done. By commenting out the line, you simply stop the system from uselessly trying to start CSSD, as you already know it does not work any more.

 

After changing /etc/inittab, you should run:

init q

 to make the init process re-read the inittab file. After reading the modified inittab, the init process will stop trying to restart /sbin/init.d/init.cssd, so the "command is respawning too rapidly" error messages will no longer appear on your console and in /var/adm/syslog/syslog.log.

MK
The opinions expressed above are the personal opinions of the authors, not of HP. By using this site, you accept the Terms of Use and Rules of Participation.