Hi All,
Hope anyone seen this before and could give me some hints or explanation.

I encounter a strange timing behaviour on a LINUX,ASE. It happens, that
the getdate() returns "wrong" values.
Not sure if it's really a wrong time value but the time jumps for and
back, back and for. At least it lokks like.


1> select @@version
2> go




---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Adaptive Server Enterprise/12.5.0.2/EBF 10753 ESD#1/P/Linux
Intel/Linux 2.4.7-10smp i686/rel12502/1867/32-bit/OPT/Sun Dec 15
07:34:10 2002


The simple "counter" I use is:

declare @count int
select @count=1
while (@count) < 10
BEGIN
select "count: ",@count," = ",convert(varchar(25), getdate(), 109)
select @count = @count+1
waitfor delay "00:00:01"
END

and the output like:

count: 1 = Aug 14 2003 5:24:12:576P
count: 2 = Aug 14 2003 5:24:13:563P
count: 3 = Aug 14 2003 5:24:17:456P
count: 4 = Aug 14 2003 5:24:15:563P
count: 5 = Aug 14 2003 5:24:16:566P
*count: 6 = Aug 14 2003 5:24:26:386P <==
^^^^
count: 7 = Aug 14 2003 5:24:27:720P
count: 8 = Aug 14 2003 5:24:29:053P
*count: 9 = Aug 14 2003 5:24:20:566P <==
^^^^

but getdate() from count 6 was called before getdate() of count 9. But
why is the result the opposite? Same for count 3.

If I do a similar time-loop with Perl
for(my $i = 1; $i <= 100; $i++) {
print localtime() ."\n";
sleep 1;
}

everything looks okay.
So I really have no idea how this could happen.

Thanx in advance
Joerg

Just a additional remark.

The timing problem is already visible in the SYBASE log at startup:
....
00:00000:00000:2003/08/14 15:21:44.71 kernel Checked out license ASE_S
00:00000:00000:2003/08/14 15:21:44.71 kernel Adaptive Server Enterpris
00:00000:00000:2003/08/14 15:21:36.43 kernel Using config area from pr
00:00000:00000:2003/08/14 15:21:36.48 kernel Internal run-time model s
00:00000:00000:2003/08/14 15:21:36.49 kernel Using 1024 file descripto
00:00000:00000:2003/08/14 15:21:44.88 kernel Adaptive Server Enterpris
00:00000:00000:2003/08/14 15:21:44.88 kernel Confidential property of
....
00:00000:00001:2003/08/14 15:21:45.84 kernel Component Integration Ser
00:00000:00001:2003/08/14 15:21:45.85 kernel Component Integration Ser
00:00000:00001:2003/08/14 15:21:45.85 server Opening Master Database .
00:00000:00001:2003/08/14 15:21:39.40 server Loading SQL Server's defa
00:00000:00001:2003/08/14 15:21:39.41 server Loaded default Unilib con
00:00000:00001:2003/08/14 15:21:39.42 server Recovering database 'mast
00:00000:00001:2003/08/14 15:21:39.44 server Redo pass of recovery has
00:00000:00001:2003/08/14 15:21:39.50 server Checking external objects
00:00000:00001:2003/08/14 15:21:39.52 server Database 'master' is now
00:00000:00001:2003/08/14 15:21:37.55 server The transaction log in th
....


Thanx in advance
Joerg

Joerg Gempe wrote:

> I encounter a strange timing behaviour on a LINUX,ASE. It happens, that
> the getdate() returns "wrong" values.
> Not sure if it's really a wrong time value but the time jumps for and
> back, back and for. At least it lokks like.

ASE gets its date from the system. So I would suggest directing
your efforts there. Check patches etc. or ask on the ASE Linux
mailing list or Linux newsgroup(s).

-am © 2003

Joerg Gempe <j.gempe@epages.de> wrote in
news:bhiksp$3ug$1@branwen.iks-jena.de:

> Any other hints?

Unfortunately as another poster has stated, this is not an ASE issue
but an O/S issue.
--
Pablo Sanchez, Blueoak Database Engineering
http://www.blueoakdb.com