RE: ** impact of time clock changes on Running Oracle DB

  • From: "Goulet, Dick" <DGoulet@xxxxxxxx>
  • To: "Ramesh FL" <karai.ramesh@xxxxxxxxx>
  • Date: Fri, 25 Feb 2005 09:54:04 -0500

Joshi,

        Yes, Mladen is correct.  Dates are stored as 7 bit binaries and
sysdate is just a call to the system clock for the current date and
time.  The spring change is of little consequence since we move forward
one hour, basically 01:59:59 moves one second forward to 03:00:00.  It's
the fall change that can cause a problem since 01:59:59 moves one second
forward to 01:00:00.  Most third part vendors who rely on time,
SmartTime comes to mind since I've inherited it, recommend shutting down
their applications for one hour during the fall change.  It's a pain,
but then it only happens twice a year.  The good part is that time for
the database is not tied to any clock, but to the SCN which marches
onward with impunity.  Course I've never tried to do a point in time
recovery of a database that crashed in the middle of that fateful fall
hour.  That might be interesting!!

-----Original Message-----
From: Ramesh FL [mailto:karai.ramesh@xxxxxxxxx]=20
Sent: Thursday, February 24, 2005 8:54 PM
To: Goulet, Dick
Cc: ajoshi977@xxxxxxxxx; oracle-l@xxxxxxxxxxxxx
Subject: Re: ** impact of time clock changes on Running Oracle DB

Dick,

Will I be right in assuming that while table with columns of type
"timestamp with local Timezone"  will not be affected but tables with
"date" columns and hence the application be affected? Even in 10g
there must be some tables with column type "date" and so it may have
some impact (like Mladen implies).

Thanks.


On Thu, 24 Feb 2005 17:07:53 -0500, Goulet, Dick <DGoulet@xxxxxxxx>
wrote:
> Joshi,
>=20
>        I'm going to make the assumption that your running Oracle 9i or
> 10g.  Maybe that's a bad assumption, but you can correct me if I'm
> wrong.  Reason that I mention this is that if your using "timestamp
with
> local timezone"  as your date data type then there is no time change
as
> far as the database is concerned.  This is because Oracle converts the
> local date/time to GMT, or Greenwich Mean Time or ZULU, before storing
> it and GMT does not change at all.  Otherwise you should see
absolutely
> no problems.  Been happening to my servers twice a year for 13 years
> now.
>=20
> -----Original Message-----
> From: A Joshi [mailto:ajoshi977@xxxxxxxxx]=3D20
> Sent: Thursday, February 24, 2005 3:38 PM
> To: oracle-l@xxxxxxxxxxxxx
> Subject: ** impact of time clock changes on Running Oracle DB
>=20
> Hi,
>  I would like to know the impact of changing the time on my UNIX
> machine. how would it impact a ORACLE database running on the server.
I
> think Oracle takes its sysdate from the UNIX. Thanks for your =
help.=3D20
>
--
//www.freelists.org/webpage/oracle-l

Other related posts: