Databases
Showing results for 
Search instead for 
Do you mean 

oracle started on its own gettime

SOLVED
Go to Solution
Highlighted
Honored Contributor [Founder]

oracle started on its own gettime

Hi one of our server restarted oracle because of this error I think its because of UNIX was unable to tell time to oracle that is why it restart.
please give your comments and solution.

Wed Jan 25 09:18:53 2006
Errors in file /in/oracle/rdbms/log/lgwr_530072.trc:
ORA-07211: slgcs: gettimeofday error, unable to get wall clock.
Compaq Tru64 UNIX Error: 83: System call timed out

Wed Jan 25 09:18:53 2006
LGWR: terminating instance due to error 7211
Instance terminated by LGWR, pid = 530072
Wed Jan 25 09:19:00 2006
Shutting down instance (abort)

thanks
Kapil
I am in this small bowl, I wane see the real world......
8 REPLIES
Honored Contributor [Founder]

Re: oracle started on its own gettime

Check your syslog log messages and binary errorlog. Was the system responsive at that time? What was the load?

If you have access to metalink, try searching there also.
Por que hacerlo dificil si es posible hacerlo facil? - Why do it the hard way, when you can do it the easy way?
Honored Contributor [Founder]

Re: oracle started on its own gettime

Hi,

please give us your os version and patch version of your machine. Have you checked with HP for hardware clock errors?

greetings,

Michael
Honored Contributor [Founder]

Re: oracle started on its own gettime

Hi Micheal
Version of OS is
Compaq Tru64 UNIX V5.1B (Rev. 2650);
NHD Rev(V6.0) installed
Still we have not logged any thing to HP
Kapil
I am in this small bowl, I wane see the real world......
Honored Contributor [Founder]

Re: oracle started on its own gettime

Are you running an NTP service?
Could there have been a NTP client/server or generic network problem at the time of the Oracle start? Check the (ntp) log for problems around the time of the oracle problems?

Look for entries in the daemon.log file like:

Xxx dd dd:dd:dd xxxxxx xntpd[dddd]: synchronisation lost
Xxx dd dd:dd:dd xxxxxx xntpd[dddd]: synchronized to d.d.d.d, stratum=dd

Check: ntpq -p
Check: ntp -v

Btw... You can get rid of the actual gettimeofday calls with:

# mknod /dev/timedev c 15 0
# chmod 644 /dev/timedev
and then restart Oracle.
Oracle will then just read a memory cell with the time, instead of using the system service. This may be a significant performance boost when running production with timed_statistics on, and may well serve as a workaround for this problem (WAG).

Cheers!
Hein.
Honored Contributor [Founder]

Re: oracle started on its own gettime

Hi all
I have a explation for this from HP:

It is possible for getttimeofday() to time out when, for example, xntpd
is adjusting the clock. Oracle in the past has released patches which
cause Oracle to retry gettimeofday() if it times out.
If you are using timed_statistics on Oracle you can enable a more system resource-efficient method of
accessing the system time by creating a special device that will allow
Oracle to avoid the getttimeofday() call:
mknod /dev/timedev c 15 0
chmod +r /dev/timedev
Restart Oracle after creating this node and Oracle will use /dev/timedev
for most of its system time checks. This has less overhead than
gettimeofday().

And I have heard this is famus oracle bug which has been removed in oracle9i.
Thanks
I am in this small bowl, I wane see the real world......
Honored Contributor [Founder]

Re: oracle started on its own gettime

Hmm, that sounds vaguely like what I wrote in the reply before that. :-) :-)
Honored Contributor [Founder]

Re: oracle started on its own gettime

Hi Kapil,

I have noticed that you have not assigned any points to any of the answers to your questions. Please have a look at:
http://forums1.itrc.hp.com/service/forums/helptips.do?#33

thanks,

Michael
Honored Contributor [Founder]

Re: oracle started on its own gettime

Problem resolved CLosing!!
I am in this small bowl, I wane see the real world......