headers
Beermann, Albert | 2 May 2005 10:56
Picon

Problem with loader ???

Hello Everybody !

to sync to databases(same databasename, same tablename)
 on different servers i have a programm that does:
1. catalogextract table a on system a
   drop table a on system b
   catalogload on system b

2. dataextract table a on system a
   dataload table a on system b

everything works as expected !

but when i trie to load a table with a varchar coloumn the following error appears:

err -25319
missing data for specified column NL

i can see a value for nl in the instream file ????
the position is after the varchar field infotext ???

catalog load:
CREATE TABLE "BWALLE"."BSARTIKELSTAMM"
(
 "BSARTIKELSTAMMID"               Char (10) ASCII    NOT NULL,
 "ARTIKELNR"               Char (10) ASCII    NOT NULL    DEFAULT ' ',
 "NAME1"               Char (35) ASCII    NOT NULL    DEFAULT ' ',
 "BSARTIKELGRUPPEID"               Char (10) ASCII    NOT NULL    DEFAULT ' ',
 "EPREIS"               Fixed (12,4)    NOT NULL    DEFAULT         0.0000,
 "INFOTEXT"               Varchar (2000) ASCII    NOT NULL    DEFAULT ' ',
(Continue reading)

Permalink | Reply | 0 comments |
headers
Beermann, Albert | 2 May 2005 12:57
Picon

update statistics ???

Hello Everybody !

After creating a lot of tables and loading data from another server i
want to do an update statistics.

I tried dbmcli -d xxx -u dba,xyz -n xxxx
> update statistics *

err: unkown command update

Did i miss something ???

Any help welcomed
Albert

                              ''''' 
                             ''''''''' 
                            (0 0) 
 +---------oOO-----------(_)------------------------------+ 
 | Tel: 0541/5841-868                                  | 
 | Fax: 0541/5841-869                                 | 
 | Mail: mailto:a.beermann <at> piepenbrock.de  |  
 | Internet:  http://www.piepenbrock.de <http://www.piepenbrock.de/>
| 
 +--------------------------------------oOO----------------+ 
                          |__|__| 
                            ||  || 
                        ooO Ooo
(Continue reading)

Permalink | Reply | 0 comments |
headers
Knappke, Christian | 2 May 2005 13:45
Picon
Favicon

RE: update statistics ???


> After creating a lot of tables and loading data from another server i
> want to do an update statistics.
>  
> I tried dbmcli -d xxx -u dba,xyz -n xxxx
> > update statistics *
>  
> err: unkown command update
>  
> Did i miss something ???

have you checked this?

$ dbmcli help | grep update
sql_updatestat           [<table or column spec>]
sql_updatestat_per_systemtable

HTH
Christian
-- 
#include <std_disclaimer.h>
/* The opinions stated above are my own and not
   necessarily those of my employer. */

--

-- 
MaxDB Discussion Mailing List
For list archives: http://lists.mysql.com/maxdb
To unsubscribe:    http://lists.mysql.com/maxdb?unsub=gcdm-maxdb <at> m.gmane.org
(Continue reading)

Permalink | Reply | 1 comment |
headers
Fabrice Bourdel | 2 May 2005 23:40
Picon
Favicon

RE: how to automate backup and restore with a filename that vary


> -----Message d'origine-----
> De : Hübschen, Peter [mailto:peter.huebschen <at> voit.de]
> Envoyé : vendredi 29 avril 2005 13:47
> À : 'Fabrice Bourdel'; maxdb
> Objet : AW: how to automate backup and restore with a filename that vary
>
>
> Hello Fabrice,
>
> I can't answer your second question, but maybe I can help with your other
> question, even if I make it a slightly different.
> To backup my database I call dbmcli and it will create a file i.e.
> "mybackupfile".
> Afterwards I do a 'mv' to rename my file with the current date as
> suffix, so
> it will not be overwritten on the next backup.
> Here is my mv-command (maybe it helps):
>
> mv /<my>/<path>/mybackupfile /<my>/<name>/mybackupfile.`date +%Y%m%d`

thanks for your response and example !

> Afterwards it looks like this:
>
> mybackupfile.20050428
>
> So I have added the current date to my backupfile.
> To import it on Windows maybe you can go the other way around: rename file
> to mybackupfile and then import it. Just my 2 cents.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Fabrice Bourdel | 2 May 2005 23:42
Picon
Favicon

RE: how to automate backup and restore with a filename that vary


> -----Message d'origine-----
> De : Heinrich, Tilo [mailto:tilo.heinrich <at> sap.com]
> Envoye : vendredi 29 avril 2005 20:23
> A : maxdb
> Objet : RE: how to automate backup and restore with a filename that vary
> 
> 
> Hi Fabrice,
> 
> If you write a script anyway, just avoid the -i option of dbmcli:
> 
> #!/bin/sh
> #
> SUFFIX=`date +%y%m%d%H%M`
> 
> 
> dbmcli -n machine -d DB -u XXX,XXX << EOF
> db_admin
> util_connect
> medium_put Master H:\MasterBackup\MASTER_COM_$SUFFIX FILE
> backup_start Master DATA
> db_online
> EOF
> 
> dbmcli -d DB -u YYY,YYY << EOF
> db_admin
> util_connect
> medium_put Master /pathtodatabase/MASTER_COM_$SUFFIX FILE DATA
> db_activate RECOVER Master DATA
(Continue reading)

Permalink | Reply | 501 comments |
headers
David Brady | 3 May 2005 18:54

JOIN_OPERATOR_IMPLEMENTATION performance...


Would love to get some feedback on this...

I've been using an old slow desktop (Linux; P2 766; 512MB) for a 
production database server - running MaxDB 7.5.0.8.  Performance on my 
most complicated queries has actually been surprisingly good.

Finally got the budget for a new server, and picked up a dual 3GHZ Xenon 
with more memory than I need.  Installed MaxDB 7.5.0.23 on it... moved 
the database from the old server to the new, and was shocked to discover 
that our complicated queries were taking an order of magnitude longer on 
this new server.  Instead of the 5 to 10 seconds I was used to, the 
queries were taking 100 to 200 seconds.  (I did update stats, btw...)

I remember reading something about a join param: 
JOIN_OPERATOR_IMPLEMENTATION.  The default was Yes, and I changed it to 
No.  After the change, performance was back where I expected - now 3 to 
4 seconds for the same queries.

I'm wondering - in what situations will JOIN_OPERATOR_IMPLEMENTATION=Yes 
help improve performance?  I'm guessing that it must, otherwise you 
folks wouldn't have set that as a default...

Cheers,
David

--

-- 
MaxDB Discussion Mailing List
For list archives: http://lists.mysql.com/maxdb
To unsubscribe:    http://lists.mysql.com/maxdb?unsub=gcdm-maxdb <at> m.gmane.org
(Continue reading)

Permalink | Reply | 0 comments |
headers
John Singleton | 3 May 2005 22:18
Picon
Gravatar

LOG Function....

Hello all,

Quick question: Should this work?

>select LOG(2, 65536) from dual

Whenever I execute this I get: 

* -5016: Missing delimiter: ) SQLSTATE: 42000

Am I doing something wrong? I have to be missing something...

Thanks!

JLS

PS: I'm using 7.5.00.19...

--

-- 
MaxDB Discussion Mailing List
For list archives: http://lists.mysql.com/maxdb
To unsubscribe:    http://lists.mysql.com/maxdb?unsub=gcdm-maxdb <at> m.gmane.org
Permalink | Reply | 0 comments |
headers
王 旭 | 4 May 2005 06:03
Picon
Favicon

RE: a question about connection via ado.net

Mr Thomas,

   This question have bother me near upon one month.Could you give me a 
small example project?Maybe it include some insert operation,query 
operation,delete operation,alter operation.

   very thanks.  

>From: "Koetter, Thomas Theodor" <thomas.theodor.koetter <at> sap.com>
>To: "? ?" <liulian0588 <at> hotmail.com>, <maxdb <at> lists.mysql.com>
>Subject: RE: a question about connection via ado.net
>Date: Fri, 29 Apr 2005 15:05:29 +0200
>
>
>
> > -----Original Message-----
> > From: liulian0588 <at> hotmail.com [mailto:liulian0588 <at> hotmail.com]
> > Sent: Freitag, 29. April 2005 11:36
> > To: maxdb <at> lists.mysql.com
> > Subject: a question about connection via ado.net
> >
> > I know maxdb support ado.net connection.But my test is fail.
> > I connect maxdb via odbc.net.
> > I config odbcconnectionstring and odbcdataadapter.
> > But i can't connect.
>
>
>If you did something like
>
>	create user myuser password mypassword
(Continue reading)

Permalink | Reply | 0 comments |
headers
Raimund Jacob | 4 May 2005 08:52
Picon

deadlock with long queries

hello *!

i have a problem with a live system that serves as a backend for some 
website. i use java with at most 40 parallel connections to connect to a 
7.5.0.24 on a linux machine (MAXUSERTASKS is 70 to leave some headroom).

there are two major tables in the DB: one with about 600 000 rows, the 
other one with about 6.4 million. during normal daily operation 
everything works fine. but at night some statistics are calculated and 
some rather complicated queries on those two tables are run (all using 
only one connection). the entire process usually takes about 10 minutes.

now, sometimes the DB completely deadlocks during the 
statistics-calculation. the 40 parallel conections pile up (they are 
lazily allocated) because each one gets no response from the DB. the 
load on the machine drops to 0 and nothing happens until i restart the 
kernel and the application. the SESSION_TIMEOUT is set to 0 if that 
makes a difference.

i increased MAXLOCKS so that it is way above what the 'maximum set' 
shown on the Activity page of DBMGui. but there are still collisions and 
table locks. the isolation level used is TRANSACTION_REPEATABLE_READ 
which is rather high, i guess.

there are 1.5 GB of permanent data and 2GB of data cache, so 
I/O-Activity should be pretty low. nothing shows up in knldiag{.err}. 
well, i see this:

Could not create thread: 'Compiled with Linuxthreads support, but no 
Linuxthreads implementation detected', rc = 0
(Continue reading)

Permalink | Reply | 0 comments |
headers
Zabach, Elke | 4 May 2005 09:47
Picon
Favicon

AW: LOG Function....

John Singleton wrote:
> 
> Hello all,
> 
> Quick question: Should this work?
> 
> >select LOG(2, 65536) from dual
> 
> Whenever I execute this I get:
> 
> * -5016: Missing delimiter: ) SQLSTATE: 42000
> 
> Am I doing something wrong? I have to be missing something...
> 

Sometimes I hate JDBC/ODBC and its 'private' definitions:
SQL defines LOG in a way you used it.
ODBC/JDBC only accepts it with exactly one parameter.
I assume, you used SQLStudio, which internally uses ODBC/JDBC, causing
the kernel to check syntax of functions against the ODBC/JDBC-way,
causing this error.

Sorry.
Elke
SAP Labs Berlin

> Thanks!
> 
> JLS
>
(Continue reading)

Permalink | Reply | 0 comments |

Gmane