Hi Norm,

After querying numerous people that have tested the DST dilemma re business
time calculations the explanation that seems to make the most sense boils
down to this.  And I apologize if all the words are not exactly right but I
was having a hard time getting my arms around it too.  I couldn't test
because we already have our dev server at v7 and did not have plans to do
production before 3/11.  So I'm relying on other's information.

Apparently with v5.x a 'library' was added that in effect defines the DST
start and end dates.  The focus was mainly on v6 and v7 with a scan
reference to v5 since v5 is no longer 'supported'.  It appears this library
was not in the AR Server before v5 but I do not have a confirmation on that.


Since that 'library' is there all business time calculations at some point
reference it to determine how it should calculate.  Since there's no patch
for v5 we are forced to upgrade.

My belief is that since this is an extraordinary situation a patch for v5
should be provided.  There are quite a few people still on it since v7 is
relatively new.  I understand the need to keep a certain level of support in
control but this is not the norm and preparation time has been minimal.

It doesn't matter if you have applied the appropriate patches to
the workstations and the server and the database.  This library is internal
to the AR Server and will play a role.

I have one critical calculation that is of concern.  There are other calcs
but if they are off an hour for a few weeks everyone will live through it.
The reason we haven't finished our upgrade on the production server is a
resource issue here.  Well, it's worse now with all the systems that need
something done to them!  A patch would be so much easier.

Please bmc, how much work could it be for you to do a v5 patch?  There are
allot of customers out here that would be grateful.  It would provide a
great deal of good will.

Thanks,
Susan

Server:  ARS 5.1.2 Patch 1428
OS:  Windows NT 5.0  2CPU's 4G Memory
Database:  Oracle 9i2
User:  ARS 5.1.2 Patch 1316
User OS:  XP, NT, Win 2000
Admin:  ARS 5.1.2 Patch 1289
Crystal that created reports:  9


Susan Palmer
ShopperTrak
200 W Monroe St  11th Floor
Chicago, IL  60606
Office:  312-529-5325
Cell:     302-502-7687
[EMAIL PROTECTED]



On 2/27/07, Kaiser Norm E CIV USAF 96 CG/SCWOE <[EMAIL PROTECTED]>
wrote:

**

Hi all:



I'm still trying to wrap my head around all the DST ramifications for
5.1.2, and I guess I'm thinking it would help if I had a technical white
paper or other document that specifies exactly how Remedy 5.1.2 calculates
time and time conversions.



Here's my thinking:



The Remedy server stores all time values as Unix time, which is the total
of seconds since 1 January 1970 GMT.  Time values, then, get stored in a
number field in the database (as opposed to a date/time field).
Accordingly, if a user passes a date and time in a search query, Remedy must
convert the date and time supplied by the user to the equivalent Unix time.
It must do this by first adding or subtracting the appropriate number of
hours based on the time zone and then possibly add an hour for DST.



If you run such a query, which piece of Remedy does this conversion before
the query is passed to the underlying database? Is it the server or the
client? Does the client do the time conversion before the query is passed to
the server or does the client just pass the query to the server as-is and
the server does the time conversion?



If the server does the time conversion, is it saying, "OK, I got a time
value in this query I'm to execute.  So let me convert the time to something
I truly understand.  So let's see now…what time zone am I in…and are we
observing daylight savings time?" I assume, then, that the server queries
the operating system for the timezone??? And does it query the operating
system for whether or not the time zone is currently observing DST? It
can't, in my mind, otherwise there wouldn't be a bug.  It must be
calculating whether or not DST is being observed itself based on its own
internal date/time algorithm? Yes?



Does anyone know the answers to these issues or know of a whitepaper that
definitively describes how Remedy calculates time?



Thanks,

Norm
__20060125_______________________This posting was submitted with HTML in
it___

_______________________________________________________________________________
UNSUBSCRIBE or access ARSlist Archives at www.arslist.org ARSlist:"Where the Answers 
Are"

Reply via email to