I think we found the conflict. The two systems that fail to run the install program
also have Cygwin SSHD. If we remove cygwin from the PATH, then
the install goes fine.
So some kind of conflict with (at least some versions of) cygwin was introduced after
7.5.0.8, as that version and earli
Hi Daniel,
Your command did not help.
We have 6 systems we have tried, all via WTS.
-- 2 fail, the other 4 work. All are same service pack / patches of win2k.
-- ALL Work with 7.5.0.8 but these two srevers fail with 7.5.0.11 and later.
-- We completely removed MaxDB / SAPDB from regist
We have never had this problem before, so we assume a recent change with the SDBINST
program.
When you try to do a install on a Windows 2000 Server (SP4 with latest Windows Update
patches)... the SDBINST does not properly launch
the 'second window' and never lets you do a install.
It is poss
Environment: Win32 dotNet 1.1, Windows 2000 Server Service Pack 4 with latest patches
and latest MDAC. One server running MaxDB
7.5.0.8, the other server running ASP.NET dotNet 1.1 as a ODBC client. connection
pooling turned on (always check this after driver install or
upgrade, it always
Is SQLDBC out of beta, supported interface?
Is there a C / C++ example program to demonstrate how to do a simple SELECT? Prefer
for win32 / Microsoft compiler.
Anyone have code to share?
Thank you.
Stephen Gutknecht
Full time traveler
currently in Dallas, Texas USA
--
MaxDB Disc
Howdy all,
Thanks to Daniel's great tips, wasn't too bad. We patched xserver to skip the DNS
lookup. The xserver parameter processing is rather different on win32 vs unix in the
source, but we did our best to copy the 'existing code
style' in how we implemented the change.
The code for the
Hello Gerd,
First thing to consider:
Have you run UPDATE STATISTICS on both systems before doing your tests?
In my experience... if MaxDB has out of date statistics, it will sometimes think the
table is small and a table scan will seem easier :)
Stephen Gutknecht
On Tue, 6 Apr 2004 17:3
Howdy all,
On Wiin32 (Windows XP or Windows 2000) with latest MaxDB 7.5.0.8 driver and server...
two systems (one server, one client). Using the
Unicode ODBC driver.
If I do these two queries from the same connection:
EXPLAIN SELECT field FROM TABLE
SELECT * FROM SHOW
I get my explain out
I feel like I"m getting bogged down in explaining my desire for a -F parameter on the
Windows version of xserver. So I'm going to take another
approach.
Offlist or onlist, can someone help me get a build environment together so I can hack
on the xserver code myself?
I see I can start with
I appreciate your confirmation of the problem.
A host file has not proven enough in case of reverse-DNS lookups that MaxDB / SAPDB
does. The same behavior has hit us wen reverse DNS was misconfigured or failed.
Especially if
you are doing a partial subnet network (/28 /29 /30, etc.) in many s
We were again hit with a production outage due to DNS dependence within MaxDB. This
time we spent 1 day setting up two test servers with
Win2K and a third running Linux. All running latest 7.5.0.8.
I reported this problem over a year ago, and no response to my two posts on the topic.
http://
I will chime in...
First off, we have seen no difference between 7.3 / 7.4 / 7.5 versions. The driver is
independent. I believe you can even use a 7.5 ODBC
driver against an older xserver/kern. I'm a bit confused why so many people on the
list aren't upgrading to 7.5? Older releases are no
I'll preface this: I'm not trying to offend, trying to help :)
No application program (userland) should be able to reboot the system (without
explicit intention). This sounds like a hardware or operating
system problem. RAM and disk controller hardware would be suspect first. OS disk
drivers
Ok, ignore this previous post :) ALTER TABLE had been used to add a couple fields to
the old set1 table and the order of fields was no longer
the same. The error returned was due to datatype mismatch on the fields in the table
:)
Specifying the exact fields in the INSERT and SELECT corrects
Using MaxDB 7.5.0.8 on Windows 2000 the following fails:
INSERT INTO tempset1 SELECT * FROM set1 WHERE timestampfield > '2004-02-28 23:30:00'
Where the statement:
SELECT * FROM set1 WHERE timestampfield > '2004-02-28 23:30:00'
Works fine in the same context. Error returned:
[MySQL MaxDB
Can someone please point me to specifics on these features?
The MaxDB press release for 7.5.00.06 mentions them:
http://www.mysql.com/press/release_2003_35.html
""Other new features introduced in MaxDB 7.5.00.06 are snapshots, archive tables and
MaxDB replication.""
What is a "snapshot" and w
On Tue, 2 Mar 2004 09:17 , 'Schildberg, Steffen' <[EMAIL PROTECTED]> sent:
>Hi Stephen,
>
>> Please do! It would be great to have the ability to snapshot
>> production data only specific tables without down time or
>> having to do a full backup. Right
>> now, the only method I know of would
On Tue, 9 Mar 2004 20:01 , 'Becker, Holger' <[EMAIL PROTECTED]> sent:
>Hi,
>
>after some discussion the only reason we could found for this exclusive
>locks was that they should guarantee that the result of a long running
>update statistic command could be inserted into the system table withou
On Tue, 2 Mar 2004 16:43 , 'Schildberg, Steffen' <[EMAIL PROTECTED]> sent:
>> It is DBEXTRACT that concerns me most and is partly
>> responsible for misleading me :) DBEXTRACT creates a
>> catalog script that will do things such as:
>>
>> USE USER a
>> CREATE TABLE a1
>> USE USER B
>>
On Mon, 1 Mar 2004 11:15 , 'Schildberg, Steffen' <[EMAIL PROTECTED]> sent:
>> I found that using tableextract PAGES of all tables I get
>> 4.7GB. Is this because the indexes are left behind? 12GB vs
>> 4.7GB is pretty big difference.
>>
>Yep, could be. It depends of course on the indexes yo
Hello Steffen,
Any progress on this topic?
It is DBEXTRACT that concerns me most and is partly responsible for misleading me :)
DBEXTRACT creates a catalog script that will do things
such as:
USE USER a
CREATE TABLE a1
USE USER B
CREATE TABLE b1
instead of:
CREATE TABLE a.a1
CR
Hi,
As I have mentioned on the list on the past... UPDATE STATISTICS requires lots of slow
disk I/O and it locks the tables. Some of our tables
are large and the lock can last for long periods (30 minutes or longer for a single
table). Yes, this is on professional 15,000 RPM SCSI hard
dri
I am feeling like loadercli is not getting attention it needs. I've posted several
topics related to it and wasted a lot of time due to design
problems I have reported. Is no one at SAP / MySQL actively taking this topic and
responding on the list?
Interesting discovery I made regarding load
Hi Sven,
We are runing all 7.5.0.8 on our systms and I've been seeing similar problems. what
is also odd is that the 'rejected bad connect packet'
sounds aweful similar to an error I was seeing with DBMCLI / SQLCLI - and these don't
even use the ODBC?
In production we have: Win2k 7.5.0.8 odb
Hi,
We seem to have discovered that loadercli tableextract locks the table it is
extracting.
we can do a backup disk to disk, but tablextract disk to disk causes all the clients
to halt.
Is there a way to tell loadercli to not do this and act more like the backup?
Thank you.
Stephen Gut
Hi,
We are working with our newly aquired 8GB Sun Solaris Sparc system. We figured once
the DATA_CACHE was populated that most of our
SELECT statements would scream... as we are only working with 10GB of data...
Solaris shows that disk writes are being performed during our SELECT statements
Steffen,
Ahh, progress!
= Windows XP output START
C:\>loadercli -n 192.168.0.150
Loader Server Protocol File:'c:\program files\sdb\data\wrk\loader.prt'
loadercli>USE USER test test SERVERDB TST
ERR -25004
A connection to the database could not be established: 1 = database no
Ok, starting this time from bottom up 'testing MaxDB' instead of 'assuming it
works'... I found the following:
loadercli running on Windows XP talking to a Solaris 8 does NOT behave the same as
loadercli running native on Solaris!
1. Install MaxDB 7.5.0.8 on Windows XP, Install MaxDB 7.5.0.8
Hi Steffen
I started over with PAGES. This time doing only one table (TABLEEXTRACT) instead of
tyring to do my entire 12GB database. The entire
database takes like 8 hours :)
Using only the Windows XP 7.5.0.8 loadercli: the export went fine, the import went
fine using PAGES format.
HOWEVE
Ok, PAGES format failed... so I'm trying "RECORDS UCS2" mode.
1. When I try to run loadercli from Windows XP using the Solaris 64 as a destination
node... I often get unxplained errors with authentication
or other. Loadercli will sometimes say protocol error. However, there are times
where
I can not get the Solaris Sparc system to import the data created with loadercli
DBEXTRACT PAGES. This is with the same version of MaxDB
(7.5.0.8) on both Windows XP and Solaris 8. I have tried it remote (using Windows XP
loadercli with node parameter) and local (using scp to
copy file to s
I believe I already know the answer to this, consider it a feature request :)
I would like to see if there is a way to make loadercli or backup work so that data is
streamed directly from one MaxDB server into another :)
Even if the data is Endian reversed (sparc vs. i386)? In other words,
Hello,
Using these commands:
DBEXTRACT CATALOG OUTSTREAM FILE 'extract-catalog'
DATA OUTSTREAM FILE 'extract-data' PAGES
I am creating an extract from Windows 2000 / Windows XP. Intention is to import into
a Solaris 64-bit system. What concerns me is the use
of the "PAGES" pa
Hi,
We have a windows 2000 server, service pack 4. 2GB of RAM. MaxDB 7.5.0BUILD
005-121-054-885.
The MaxDB server stopped responding, even to local dbmcli commands. I did a db_cold
and db_warm and it resumed normal operation.
Reviewing the kernel error log, during the error time per
Hi,
Does anyone have experience to share about trying to get DATA_CACHE as high as
possible on Solaris 8 (SunOS 5.8)? I have a Sun Blade 1000 with
8GB of RAM (64bit Sparc). No other application is hosted on this system, only MaxDB
for a single database instance.
I have no prior Solaris ex
35 matches
Mail list logo
