VMS Timezone/TDF Question

[CDahl]

Hello all,

On 4/1/07 (date of the old DST change) our Oracle sysdate reported +1 hour ahead of the current VMS system time. We found the Time Differential Factor (TDF) was set to -6h instead of the correct -5h offset.

I found that Oracle's sysdate function uses TDF to calculate it's time, but shouldn't the Oracle sysdate have been 1 hour behind instead of 1 hour ahead??

Our AUTO_DLIGHT_SAV is set to "0" so no automatic change should have occurred. Setting the TDF correctly to -5h corrected the problem but sure doesn't seem like this could happen.

Version Info: We are running OpenVMS 7.3-1 and Oracle 8.1.6.

Anyone…Anyone…Bueller…Bueller??

[Richard Whalen]

The calculation is based upon the value of the SYS$TIMEZONE_RULE logical. If you have applied the patch it will have a value of "EST5EDT4,M3.2.0/02,M11.1.0/02", if you haven't applies the patch it will have a value of "EST5EDT4,M4.1.0/02,M10.5.0/02".
(The example I gave is for the eastern timezone, the appropriate values get substituted for the first string in other timezones.)

[Hoff]

Fix your TZs, and check your Oracle version as Oracle has/had its own DST, check your Java as it has/had its own DST, etc...

Details on this DST effort -- everything that has surfaced out of this particular little adventure -- are included at:
http://64.223.189.234/node/72

And DST-related materials are listed at:
http://64.223.189.234/taxonomy/term/57

Stephen Hoffman
HoffmanLabs LLC