sqlrelay-discussion — Discussion of topics related to SQL Relay

Re: [Sqlrelay-discussion] memory corruption, apache segfault

[Ionut C. <ion...@av...>]

Yes, still happening.

Should i try CVS?

Ionut

----- Original Message -----=20
From: "David Muse" <dav...@fi...>
To: <sql...@li...>
Sent: Wednesday, September 14, 2005 4:27 AM
Subject: Re: [Sqlrelay-discussion] memory corruption, apache segfault


| Well, I said I'd look into this problem, then I misplaced the email
| until today :/
|
| Are you still having this problem?
|
| David Muse
| dav...@fi...
|
|
| On Fri, 2005-08-19 at 07:48 +0300, Ionut Ciocirlan wrote:
| > Hi,
| >
| > Nasty problem I ran into:
| >
| > I'm running sqlrelay-0.36.4 compiled against oracle instant=20
client-10.1.0.4,
| > on fedora core 3, glibc-2.3.5; apache-2.0.53 / php-4.3.11.
| >
| > Trying to fetch text containing several utf-8 characters (namely=20
diacritics)
| > from clobs sometimes crashes apache, with glibc memory errors.
| >
| > NLS_LANG is .AL32UTF8 for both Oracle and SQL Relay; data posted into=
=20
Apache
| > is in UTF-8.
| >
| > Apache crashes with various messages:
| >
| > *** glibc detected *** malloc(): memory corruption: 0x094d6f20 ***
| > [notice] child pid 4005 exit signal Aborted (6)
| >
| > *** glibc detected *** free(): invalid next size (fast): 0x0a174a80 *=
**
| > [notice] child pid 15277 exit signal Aborted (6)
| >
| > *** glibc detected *** double free or corruption (!prev): 0x094d0ee0 =
***
| > [notice] child pid 4004 exit signal Aborted (6)
| >
| > *** glibc detected *** double free or corruption (out): 0x0948c598 **=
*
| > [notice] child pid 4062 exit signal Aborted (6)
| > [notice] child pid 4009 exit signal Segmentation fault (11)
| >
| > The most common error is the first (memory corruption). On almost all
| > occasions, the client browser gets a zero-sized reply. I'd say it's=20
random,
| > but on fast reruns it happens ~ 5 out of 6 times. Sometimes (rarely) =
i=20
get
| > no crash, but apache hangs without replying, and the browser hangs=20
waiting
| > for an answer.
| >
| > Reproduce this by inserting characters 'z =C4=83=C3=AE=C5=9F=C5=A3=C3=
=A2 y =C4=82=C3=8E=C5=9E=C5=A2=C3=82 x' in a clob=20
column.
| > If it doesn't reproduce add this strign several more times and mix it=
=20
with
| > various other text.
| >
| > So, this happens only on clob columns (never on char/varchar):
| >
| > Fetching (illegally?) in non-lob mode crashes at sqlrcur_sendQuery:
| >
| > $query =3D 'select column from table where id =3D xx';
| > sqlrcur_sendQuery($cur,$query);
| > ---- crash
| >
| > Clob fetching crashes at sqlrcur_executeQuery:
| >
| > $query =3D 'begin
| > select column into :outvar from table where where id =3D xx;
| > end;'
| > sqlrcur_prepareQuery($cur, $query);
| > sqlrcur_defineOutputBindClob($cur, 'outvar');
| > sqlrcur_executeQuery($cur);
| > ---- crash
| >
| >
| > The memory corruption issues seem to be pretty real: it reflects main=
ly=20
in a
| > flash communication server running on the same machine, that after=20
several
| > such events starts behaving erratically. Sshd and samba get erratic t=
oo
| > (connections lost etc.). Also, after several hours of testing/debuggi=
ng,=20
my
| > kernel Oopsed. And just now network connectivity went into oblivion. =
The
| > only solution to restoring normal operation seems to be rebooting.
| >
| > Conclusions so far:
| > - not a machine problem; memtested the hell out of the ram;
| > - not an oracle/oci/glibc problem; i recompiled php's oci8 extension=20
against
| > the same oracle instant client libs, and i get the expected output, p=
lus=20
no
| > crashes;
| > - i also recompiled sqlrelay against the 9.1 oracle installation; sam=
e
| > results;
| > - might be a problem that appeared since 0.34 (we did some extensive=20
testing
| > back then, and it all worked out fine); i wanted to test now on 0.34=20
too,
| > but for some reason it now fails to compile right from the start
| >
| > I hope this will be easy to reproduce / repair.....
| >
| > Regards,
| > Ionut
| >
| >
| >
| > -------------------------------------------------------
| > SF.Net email is Sponsored by the Better Software Conference & EXPO
| > September 19-22, 2005 * San Francisco, CA * Development Lifecycle=20
Practices
| > Agile & Plan-Driven Development * Managing Projects & Teams * Testing=
 &=20
QA
| > Security * Process Improvement & Measurement *=20
http://www.sqe.com/bsce5sf
| > _______________________________________________
| > Sqlrelay-discussion mailing list
| > Sql...@li...
| > https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
| >
|
|
|
| -------------------------------------------------------
| SF.Net email is sponsored by:
| Tame your development challenges with Apache's Geronimo App Server.=20
Download
| it for free - -and be entered to win a 42" plasma tv or your very own
| Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
| _______________________________________________
| Sqlrelay-discussion mailing list
| Sql...@li...
| https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion=20

Re: [Sqlrelay-discussion] memory corruption, apache segfault

[David M. <dav...@fi...>]

Well, I said I'd look into this problem, then I misplaced the email
until today :/

Are you still having this problem?

David Muse
dav...@fi...


On Fri, 2005-08-19 at 07:48 +0300, Ionut Ciocirlan wrote:
> Hi,
> 
> Nasty problem I ran into:
> 
> I'm running sqlrelay-0.36.4 compiled against oracle instant client-10.1.0.4,
> on fedora core 3, glibc-2.3.5; apache-2.0.53 / php-4.3.11.
> 
> Trying to fetch text containing several utf-8 characters (namely diacritics)
> from clobs sometimes crashes apache, with glibc memory errors.
> 
> NLS_LANG is .AL32UTF8 for both Oracle and SQL Relay; data posted into Apache
> is in UTF-8.
> 
> Apache crashes with various messages:
> 
> *** glibc detected *** malloc(): memory corruption: 0x094d6f20 ***
> [notice] child pid 4005 exit signal Aborted (6)
> 
> *** glibc detected *** free(): invalid next size (fast): 0x0a174a80 ***
> [notice] child pid 15277 exit signal Aborted (6)
> 
> *** glibc detected *** double free or corruption (!prev): 0x094d0ee0 ***
> [notice] child pid 4004 exit signal Aborted (6)
> 
> *** glibc detected *** double free or corruption (out): 0x0948c598 ***
> [notice] child pid 4062 exit signal Aborted (6)
> [notice] child pid 4009 exit signal Segmentation fault (11)
> 
> The most common error is the first (memory corruption). On almost all
> occasions, the client browser gets a zero-sized reply. I'd say it's random,
> but on fast reruns it happens ~ 5 out of 6 times. Sometimes (rarely) i get
> no crash, but apache hangs without replying, and the browser hangs waiting
> for an answer.
> 
> Reproduce this by inserting characters 'z ăîşţâ y ĂÎŞŢÂ x' in a clob column.
> If it doesn't reproduce add this strign several more times and mix it with
> various other text.
> 
> So, this happens only on clob columns (never on char/varchar):
> 
> Fetching (illegally?) in non-lob mode crashes at sqlrcur_sendQuery:
> 
> $query = 'select column from table where id = xx';
> sqlrcur_sendQuery($cur,$query);
> ---- crash
> 
> Clob fetching crashes at sqlrcur_executeQuery:
> 
> $query = 'begin
> select column into :outvar from table where where id = xx;
> end;'
> sqlrcur_prepareQuery($cur, $query);
> sqlrcur_defineOutputBindClob($cur, 'outvar');
> sqlrcur_executeQuery($cur);
> ---- crash
> 
> 
> The memory corruption issues seem to be pretty real: it reflects mainly in a
> flash communication server running on the same machine, that after several
> such events starts behaving erratically. Sshd and samba get erratic too
> (connections lost etc.). Also, after several hours of testing/debugging, my
> kernel Oopsed. And just now network connectivity went into oblivion. The
> only solution to restoring normal operation seems to be rebooting.
> 
> Conclusions so far:
> - not a machine problem; memtested the hell out of the ram;
> - not an oracle/oci/glibc problem; i recompiled php's oci8 extension against
> the same oracle instant client libs, and i get the expected output, plus no
> crashes;
> - i also recompiled sqlrelay against the 9.1 oracle installation; same
> results;
> - might be a problem that appeared since 0.34 (we did some extensive testing
> back then, and it all worked out fine); i wanted to test now on 0.34 too,
> but for some reason it now fails to compile right from the start
> 
> I hope this will be easy to reproduce / repair.....
> 
> Regards,
> Ionut
> 
> 
> 
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> Sqlrelay-discussion mailing list
> Sql...@li...
> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
> 

[Sqlrelay-discussion] Debian 3.1 - SP w/ set nocount off Doesn't Return Results

[Ben G. <me...@be...>]

Another strange sqlrsh issue on this new Debian 3.1 box: sqlrsh and 
sqlrelay's Ruby module won't return stored procedure results if set 
nocount is off.

The stored procedure:
    CREATE PROCEDURE spTest AS
    set nocount off
    select 'here'

Results
    0> exec {db}.{user}.spTest;
        Rows Returned   : 0
        Fields Returned : 0
        System time     : 20000

If I change "set nocount off" to "set nocount on", things work 
fine--"here" is returned. Interestingly, if I execute the same sp on a 
Debian 3.0, both the "nocount on" and "nocount off" versions work 
properly (return "here").

Any ideas?

Have a great day!
Ben

Re: [Sqlrelay-discussion] Error messages are not visible

[Maciej W. <mac...@co...>]

> Try it from sqlrsh and see if it works.  This, for example works for me:
I'm not able to check it now (I have no sqlrelay, oracle environment here), 
but I'll check this later.


> You should be able to run the function and get it's result without using
> a select, like this:
Yes, I know it works this way but the point is that in our apps we already
have some 'select xxx from dual' statements and I'm afraid about them. The 
problem is not only with "cannot perform a DML
operation inside a query" but AFAIR with other exceptions too. 

I can write script that will change inproper (dual)
statements to correct ones and tell programmers not to use such things
but it'll be nice if sqlrelay will work with this.

-- 
Maciej Wisniowski




Firstworks/4access wrote:

>I wonder if the problem is specific to the python API.  Try it from
>sqlrsh and see if it works.  This, for example works for me:
>
>0> create table testtable (col1 varchar2(20));         Rows Returned   :
>0
>        Fields Returned : 0
>        System time     : 400000
>
>0> create or replace function test(var1 in varchar) return varchar is
>begin insert into testtable values (var1); return 'hello'; end;;
>        Rows Returned   : 0
>        Fields Returned : 0
>        System time     : 410000
>
>0> select test('hi') from dual; ORA-14551: cannot perform a DML
>operation inside a query
>ORA-06512: at "TESTUSER.TEST", line 1
>
>Attempted Query:
>select test('hi') from dual
>
>        Rows Returned   : 0
>        Fields Returned : 0
>        System time     : 420000
>
>0>
>
>
>By "works", I mean, I get the error.
>
>You should be able to run the function and get it's result without using
>a select, like this:
>
>from SQLRelay import PySQLRClient
>from SQLRelay import PySQLRDB
>con=PySQLRClient.sqlrconnection('localhost',9000,'','test','test',0,1)
>cur=PySQLRClient.sqlrcursor(con)
>res = cur.prepareQuery('begin :out=PCK_TESTY_STEROWNIKA.PROC_3(:in1, :in2); end;')
>cur.inputBind('in1', 'jkkkak', 10, 0)
>cur.inputBind('in2', 1, 10,2)
>cur.defineOutputBind('out',20)
>res = cur.executeQuery()
>out = cur.getOutputBind('out')
>
>Dave
>  
>

[Sqlrelay-discussion] sqlrsh not exiting after completing a script

[Ben G. <me...@be...>]

Hi,

On this new Debian 3.1 box, sqlrsh won't automatically exit after 
running a SQL script. If I run sqlrsh -id (id) file.sql, I will see 
output from the script but no command prompt. Pressing Ctrl-C gets me 
back to the command prompt. On a Debian 3.0 box, sqlrsh will run scripts 
then exit automatically (returning to a command prompt) as it should.. 
Do you know why it does not automatically exit on my Debian 3.1 setup?

Ben

[Sqlrelay-discussion] Re: Invalid Column Name When Using Double Quotes Around an Update..Set Value

[Ben G. <me...@be...>]

Hi,

Ah! You are right. Setting FreeTDS's TDS version setting for that server 
down from 7.0 to 4.2 changed the default quoted_identifier setting which 
fixed the problem.

 >   > (sqlrelay > Sybase > SQL Server 7)
 > Sybase?

Yes. Our current set up uses a Sybase driver to connect sqlrelay to Sql 
Server. I think its a Sybase ASE driver.

Thank you for your help,
Ben


From: Maciej Wisniowski <maciej@co...> (2005-09-09 13:31)
 I think it's not SQLRelay problem. Check QUOTED_IDENTIFIER setting in
 your SQL Server 7 it can be:
 SET QUOTED_IDENTIFIER ON or SET QUOTED_IDENTIFIER OFF and it's about
 quotes handling
 I think.
 
  > (sqlrelay > Sybase > SQL Server 7)
 Sybase?
 

[Sqlrelay-discussion] Re: Connecting from Ruby to FreeTDS to Microsoft SQL Server 7.0

[Ben G. <me...@be...>]

A clean wipe of the hard drive fixed this problem. :-) I guess one of my 
many experiments had messed things up. Thank you for your help.

====
From: Ben Gribaudo <me@be...>
 Re: Connecting from Ruby to FreeTDS to Microsoft SQL Server 7.0   
2005-08-31 13:01  
 Howdy,
 
 Dave, thank you for your response. I just saw it today.
 
 In the mean time, I've removed the code I complied as best I could and
 tried using only Debian packages obtained from apt-get.
 
 Using a script similar to the one I previously posted (with or without a
 call to endSession()), I get:
 "/usr/local/lib/site_ruby/1.8/i386-linux/sqlrelay.so:
 /usr/local/lib/site_ruby/1.8/i386-linux/sqlrelay.so: undefined symbol:
 __ti13genericsocket -
 /usr/local/lib/site_ruby/1.8/i386-linux/sqlrelay.so (LoadError)"
 
 Would any of you know what's wrong? I tried Googling on this but didn't
 find help there.
 
 Thank you
 Ben

 

Re: [Sqlrelay-discussion] Error messages are not visible

[Firstworks/4access <dav...@fi...>]

I wonder if the problem is specific to the python API.  Try it from
sqlrsh and see if it works.  This, for example works for me:

0> create table testtable (col1 varchar2(20));         Rows Returned   :
0
        Fields Returned : 0
        System time     : 400000

0> create or replace function test(var1 in varchar) return varchar is
begin insert into testtable values (var1); return 'hello'; end;;
        Rows Returned   : 0
        Fields Returned : 0
        System time     : 410000

0> select test('hi') from dual; ORA-14551: cannot perform a DML
operation inside a query
ORA-06512: at "TESTUSER.TEST", line 1

Attempted Query:
select test('hi') from dual

        Rows Returned   : 0
        Fields Returned : 0
        System time     : 420000

0>


By "works", I mean, I get the error.

You should be able to run the function and get it's result without using
a select, like this:

from SQLRelay import PySQLRClient
from SQLRelay import PySQLRDB
con=PySQLRClient.sqlrconnection('localhost',9000,'','test','test',0,1)
cur=PySQLRClient.sqlrcursor(con)
res = cur.prepareQuery('begin :out=PCK_TESTY_STEROWNIKA.PROC_3(:in1, :in2); end;')
cur.inputBind('in1', 'jkkkak', 10, 0)
cur.inputBind('in2', 1, 10,2)
cur.defineOutputBind('out',20)
res = cur.executeQuery()
out = cur.getOutputBind('out')

Dave


On Thu, 2005-09-08 at 14:08 +0200, Maciej Wisniowski wrote:
> Hi!
> 
> When I try to execute Oracle function from python with code like this:
> 
> # 
> ----------------------------------------------------------------------------------------------
> from SQLRelay import PySQLRClient
> from SQLRelay import PySQLRDB
> con=PySQLRClient.sqlrconnection('localhost',9000,'','test','test',0,1)
> cur=PySQLRClient.sqlrcursor(con)
> res = cur.prepareQuery('select PCK_TESTY_STEROWNIKA.PROC_3(:in1, :in2) 
> from dual')
> cur.inputBind('in1', 'jkkkak', 10, 0)
> cur.inputBind('in2', 1, 10,2)
> res = cur.executeQuery()
>    
> print 'res:', res, cur.errorMessage()
> print 'val:', cur.getField(0,0)
> print 'end'
> # 
> ----------------------------------------------------------------------------------------------
> 
> then if there is a DML operation in my function Oracle returns error like:
> 
> Ora 14551 Cannot Perform A Dml Operation Inside A Query
> 
> that says it's not possible to use this function with select xxx from
> dual clause.
> 
> The problem is that SQLRelay doesn't return any errors.
> 
> Output is simply:
> 
> res: 1 None
> val: None
> end
> 
> I'm not sure whether other errors from functions called in this way are
> visible... but it seems to be a bug. We're using SQLRelay 0.36.4
> 

Re: [Sqlrelay-discussion] SQLRelay and MySQL Cluster

[Firstworks/4access <dav...@fi...>]

Sorry, SQL Relay can't currently do that.  I've gotten lots of requests
for that feature, so I'll bump it's priority up.

David Muse
dav...@fi...

On Mon, 2005-09-12 at 17:35 +0800, Wang Chunsheng wrote:
> if it can do so,it's really  perfect!
> 
> 
> 
> 
> 2005/9/9, Sandeep Chayapathi <csa...@cy...>:
> > Hi all,
> > 
> > I have a mysql cluster, with single write server and multiple read
> > servers. All commits/inserts to the write servers are replicated on to
> > the read servers via the mysql master/slave setup.
> > 
> > I plan to use SQLRelay to pool the connection. Is it possible to setup
> > SqlRelay to always connect to a single server for all commits/inserts
> > and to connect to a cluster of servers for all reads/selects ?
> > 
> > - Sandeep
> > 
> > 
> > 
> > -------------------------------------------------------
> > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> > Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> > Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> > _______________________________________________
> > Sqlrelay-discussion mailing list
> > Sql...@li...
> > https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
> >
> 
> 
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> Sqlrelay-discussion mailing list
> Sql...@li...
> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
> 

Re: [Sqlrelay-discussion] SQLRelay and MySQL Cluster

[Wang C. <chu...@gm...>]

if it can do so,it's really  perfect!




2005/9/9, Sandeep Chayapathi <csa...@cy...>:
> Hi all,
>=20
> I have a mysql cluster, with single write server and multiple read
> servers. All commits/inserts to the write servers are replicated on to
> the read servers via the mysql master/slave setup.
>=20
> I plan to use SQLRelay to pool the connection. Is it possible to setup
> SqlRelay to always connect to a single server for all commits/inserts
> and to connect to a cluster of servers for all reads/selects ?
>=20
> - Sandeep
>=20
>=20
>=20
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practic=
es
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & Q=
A
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> Sqlrelay-discussion mailing list
> Sql...@li...
> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
>

Re: [Sqlrelay-discussion] Invalid Column Name When Using Double Quotes Around an Update..Set Value

[Maciej W. <mac...@co...>]

I think it's not SQLRelay problem. Check QUOTED_IDENTIFIER setting in 
your SQL Server 7 it can be:
SET QUOTED_IDENTIFIER ON or SET QUOTED_IDENTIFIER OFF and it's about 
quotes handling
I think.

 > (sqlrelay > Sybase > SQL Server 7)
Sybase?

-- 
Maciej Wisniowski

Ben Gribaudo wrote:

> Hello,
>
> I'm still working on this Debian 3.1 box (sqlrelay > freetds > SQL 
> Server 7) and have encountered a strange error: An update query with 
> single quotes around the value works but using double quotes results 
> in sqlrsh reporting an invalid column name error. The column it says 
> it can't find is the *value* of the set statement. Stranger still, I 
> can copy-and-paste the double-quote query to a Debian 3.0 box 
> (sqlrelay > Sybase > SQL Server 7) and it works fine. Any ideas?
>
> Ben
>
> ==== Using double quotes around the value ====
> 0> update entity set name="Gribaudo3" where idnumber=183276;
> Server message:
>        severity(16)
>        number(207)
>        state(3)
>        line(1)
> Server Name:
> OBNTDEV01
> Procedure Name:
>
> Error:  Invalid column name 'Gribaudo3'.  <---- Notice that this is 
> the value to be set, not the column name.
>
> Attempted Query:
> update entity set name="Gribaudo3" where idnumber=183276
>        Rows Returned   : 0
>        Fields Returned : 0
>        System time     : 20000
>
> ==== Using single quotes ====
> 0> update entity set name='Gribaudo3' where idnumber=183276;
>
>
>        Rows Returned   : 0
>        Fields Returned : 0
>        System time     : 30000
>
>
>
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle 
> Practices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing 
> & QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> Sqlrelay-discussion mailing list
> Sql...@li...
> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
>

[Sqlrelay-discussion] Invalid Column Name When Using Double Quotes Around an Update..Set Value

[Ben G. <me...@be...>]

Hello,

I'm still working on this Debian 3.1 box (sqlrelay > freetds > SQL 
Server 7) and have encountered a strange error: An update query with 
single quotes around the value works but using double quotes results in 
sqlrsh reporting an invalid column name error. The column it says it 
can't find is the *value* of the set statement. Stranger still, I can 
copy-and-paste the double-quote query to a Debian 3.0 box (sqlrelay > 
Sybase > SQL Server 7) and it works fine. Any ideas?

Ben

==== Using double quotes around the value ====
0> update entity set name="Gribaudo3" where idnumber=183276;
Server message:
        severity(16)
        number(207)
        state(3)
        line(1)
Server Name:
OBNTDEV01
Procedure Name:

Error:  Invalid column name 'Gribaudo3'.  <---- Notice that this is the 
value to be set, not the column name.

Attempted Query:
update entity set name="Gribaudo3" where idnumber=183276
        Rows Returned   : 0
        Fields Returned : 0
        System time     : 20000

==== Using single quotes ====
0> update entity set name='Gribaudo3' where idnumber=183276;


        Rows Returned   : 0
        Fields Returned : 0
        System time     : 30000

[Sqlrelay-discussion] SQLRelay and MySQL Cluster

[Sandeep C. <csa...@cy...>]

Hi all,

I have a mysql cluster, with single write server and multiple read
servers. All commits/inserts to the write servers are replicated on to
the read servers via the mysql master/slave setup.

I plan to use SQLRelay to pool the connection. Is it possible to setup
SqlRelay to always connect to a single server for all commits/inserts
and to connect to a cluster of servers for all reads/selects ?

- Sandeep

[Sqlrelay-discussion] Error messages are not visible

[Maciej W. <mac...@co...>]

Hi!

When I try to execute Oracle function from python with code like this:

# 
----------------------------------------------------------------------------------------------
from SQLRelay import PySQLRClient
from SQLRelay import PySQLRDB
con=PySQLRClient.sqlrconnection('localhost',9000,'','test','test',0,1)
cur=PySQLRClient.sqlrcursor(con)
res = cur.prepareQuery('select PCK_TESTY_STEROWNIKA.PROC_3(:in1, :in2) 
from dual')
cur.inputBind('in1', 'jkkkak', 10, 0)
cur.inputBind('in2', 1, 10,2)
res = cur.executeQuery()
   
print 'res:', res, cur.errorMessage()
print 'val:', cur.getField(0,0)
print 'end'
# 
----------------------------------------------------------------------------------------------

then if there is a DML operation in my function Oracle returns error like:

Ora 14551 Cannot Perform A Dml Operation Inside A Query

that says it's not possible to use this function with select xxx from
dual clause.

The problem is that SQLRelay doesn't return any errors.

Output is simply:

res: 1 None
val: None
end

I'm not sure whether other errors from functions called in this way are
visible... but it seems to be a bug. We're using SQLRelay 0.36.4

-- 
Maciej Wisniowski

Re: [Sqlrelay-discussion] Error in Oracle procedure and strange SQLRelay behaviour

[Maciej W. <mac...@co...>]

Hi!

> I've seen errors like that when 2
> versions of SQL Relay are installed on the same machine and the Python
> API was compiled against 1 but ends up being run against another at run
> time.


I did more tests and example below causes same error 
as described in previous mail. I've 
checked this on two different servers with sqlrelay 0.36.4 
installed. We've cleaned system and installed sqlrelay again
and the problem persists so I don't think it's because of
different sqlrelay versions.

#------------------------------------------------------------------
PROCEDURE PROC_2(a_varchar_out OUT zm_varchar%TYPE,                 
                 a_number_out OUT zm_number%TYPE,
                 a_varchar_in IN zm_varchar%TYPE,
                 a_number_in IN zm_number%TYPE)
IS    
    a_number NUMBER (10,2);
    a_varchar VARCHAR2(10);
BEGIN     
    a_varchar_out := a_varchar_in;
    a_number_out := a_number_in;
END;
#------------------------------------------------------------------
 
types are:
  zm_number NUMBER(10,2);
  zm_varchar VARCHAR2(10);

I call it like this from Python:

#------------------------------------------------------------------
from SQLRelay import PySQLRClient
from SQLRelay import PySQLRDB
 
con=PySQLRClient.sqlrconnection('localhost',9000,'','test','test',0,1)
cur=PySQLRClient.sqlrcursor(con)

res = cur.prepareQuery('begin PCK_TESTY_STEROWNIKA.PROC_2(:out1, :out2, :in1, :in2); end;')
 
cur.inputBind('p1', 'xxxx', 0, 0)
cur.inputBind('p2', None, 0,0)    #ERROR APPEARS WHEN None IS BINDED

cur.defineOutputBind('out1', 100)
cur.defineOutputBind('out2', 100)
 
res = cur.executeQuery()

print 'res:', res, cur.errorMessage()
print 'out1:', cur.getOutputBind('out1')
print 'out2:', cur.getOutputBind('out2')
print 'end'
#------------------------------------------------------------------

-- 
Maciej Wisniowski

Re: [Sqlrelay-discussion] SQLRelay hangs with incorrect connection string

[Maciej W. <mac...@co...>]

And now I'm receiving "Authentication error" :)

Thanks

-- 
Maciej Wisniowski

Firstworks/4access napisał(a):

>Looks like there's a bug in the sqlr-listener.
>
>If you edit src/listener/sqlrlistener.C and look for the clientSession()
>function, there's a bit of code like this in the function:
>
>		// snooze before and after returning an
>		// authentication error to discourage
>		// brute-force password attacks
>		snooze::macrosnooze(2);
>		clientsock->write((unsigned short)ERROR);
>		snooze::macrosnooze(2);
>
>(or something similar)
>
>insert a call to flushWriteBuffer(clientsock)
>directly after the call to clientsock->write() and
>it should fix the problem.
>
>Dave
>dav...@fi...
>
>On Wed, 2005-08-31 at 10:53 +0200, Maciej Wisniowski wrote:
>  
>
>>Hi
>>
>>I've just ecountered one problem with connection parameters to SQLRelay
>>
>>. In sqlrelay.conf I have:
>>
>><instance id="xadm" port="9004"
>>               socket="/tmp/xadm.socket"
>>               dbase="oracle7" connections="3" maxconnections="5"
>>               maxqueuelength="0"
>>               growby="1" ttl="60" endofsession="commit"         
>>               sessiontimeout="600" runasuser="nobody"
>>               runasgroup="nobody" cursors="5"
>>               authtier="listener" handoff="pass"
>>               debug="None">
>><users>
>>    <user user="test" password="test"/>
>>    </users>
>>    <connections>
>>        <connection connectionid="xadm" 
>>string="user=xadm;password=xadm;oracle_sid=devz" metric="1"/>
>>    </connections>
>></instance>
>>
>>
>>when using script like this:
>>
>>// ------------------------------------------------------------
>>from SQLRelay import PySQLRClient
>>from SQLRelay import PySQLRDB
>>
>>con=PySQLRDB.connect('localhost',9004,'','test','test',0,1)
>>cur=PySQLRClient.sqlrcursor(con)
>>res = cur.sendQuery('select * from xxx')
>>// ------------------------------------------------------------
>>
>>everything is OK (ORA exception is thrown (table xxx not known)
>>and res == 0)
>>but changing connection string to:
>>
>>con=PySQLRDB.connect('localhost',9004,'','test','testx',0,1)
>>
>>(here is incorrect password: testx) causes script to
>>hang waiting for response...
>>
>>This is PythonAPI and SQLRelay 0.36.4.
>>Is this a bug or my config file or script is wrong?
>>
>>    
>>
>
>
>
>-------------------------------------------------------
>SF.Net email is Sponsored by the Better Software Conference & EXPO
>September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
>Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
>Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
>_______________________________________________
>Sqlrelay-discussion mailing list
>Sql...@li...
>https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
>
>  
>

Re: [Sqlrelay-discussion] 0.37pre2

[Maciej W. <mac...@co...>]

We've just tried to compile this version and we
ecountered problems with make - something in tcl files:

make[4]: Wchodzę katalog `/home/zope/sqlrelay/sqlrelay-0.37/src/api/tcl'
/bin/sh ../../../libtool --mode=compile g++ -Wall -pipe  -DHAVE_CONFIG  
-DUSE_TCL_STUBS -I/usr/include -I../../.. -I../../../src/api/c++/include 
-I/usr/local/firstworks/include -I/usr/kerberos/include -c sqlrelayCmd.C 
-o sqlrelayCmd.lo
 g++ -Wall -pipe -DHAVE_CONFIG -DUSE_TCL_STUBS -I/usr/include -I../../.. 
-I../../../src/api/c++/include -I/usr/local/firstworks/include 
-I/usr/kerberos/include -c sqlrelayCmd.C  -fPIC -DPIC -o .libs/sqlrelayCmd.o
make[4]: Opuszczam katalog `/home/zope/sqlrelay/sqlrelay-0.37/src/api/tcl'
make[4]: Wchodzę katalog `/home/zope/sqlrelay/sqlrelay-0.37/src/api/tcl'
/bin/sh ../../../libtool --mode=compile g++ -Wall -pipe  -DHAVE_CONFIG  
-DUSE_TCL_STUBS -I/usr/include -I../../.. -I../../../src/api/c++/include 
-I/usr/local/firstworks/include -I/usr/kerberos/include -c sqlrelayCmd.C 
-o sqlrelayCmd.lo
 g++ -Wall -pipe -DHAVE_CONFIG -DUSE_TCL_STUBS -I/usr/include -I../../.. 
-I../../../src/api/c++/include -I/usr/local/firstworks/include 
-I/usr/kerberos/include -c sqlrelayCmd.C  -fPIC -DPIC -o .libs/sqlrelayCmd.o
make[4]: Opuszczam katalog `/home/zope/sqlrelay/sqlrelay-0.37/src/api/tcl'
first try errors :
In file included from sqlrelayCmd.C:13:
../../../config.h:278:1: warning: "INLINE" redefined
In file included from sqlrelayCmd.C:7:
/usr/include/tcl.h:268:1: warning: this is the location of the previous 
definition
sqlrelayCmd.C: In function `int sqlrcurObjCmd(void*, Tcl_Interp*, int, 
Tcl_Obj*
   const*)':
sqlrelayCmd.C:337: invalid conversion from `const char**' to `char**'
sqlrelayCmd.C:345: syntax error before `;' token
sqlrelayCmd.C:357: syntax error before `=' token
sqlrelayCmd.C:357: `Tcl_WideInt' undeclared (first use this function)
sqlrelayCmd.C:357: (Each undeclared identifier is reported only once for 
each
   function it appears in.)
sqlrelayCmd.C:357: syntax error before `->' token
sqlrelayCmd.C:360: syntax error before `,' token
sqlrelayCmd.C:360: uninitialized const `col'
sqlrelayCmd.C:361: syntax error before `,' token
sqlrelayCmd.C:361: conflicting types for `uint32_t col'
sqlrelayCmd.C:360: previous declaration as `const char col'
sqlrelayCmd.C:361: warning: unused variable `uint32_t col'
sqlrelayCmd.C:1134: syntax error before `;' token
sqlrelayCmd.C:1143: syntax error before `)' token
sqlrelayCmd.C:1143: `Tcl_GetWideIntFromObj' undeclared (first use this
   function)
sqlrelayCmd.C:1146: syntax error before `)' token
sqlrelayCmd.C:1147: syntax error before `)' token
sqlrelayCmd.C:1160: syntax error before `;' token
sqlrelayCmd.C:1168: syntax error before `)' token
sqlrelayCmd.C:1171: syntax error before `)' token
sqlrelayCmd.C: In function `int sqlrconObjCmd(void*, Tcl_Interp*, int, 
Tcl_Obj*
   const*)':
sqlrelayCmd.C:1709: invalid conversion from `const char**' to `char**'
sqlrelayCmd.C: In function `int sqlrconCmd(void*, Tcl_Interp*, int, Tcl_Obj*
   const*)':
sqlrelayCmd.C:1905: invalid conversion from `const char**' to `char**'
make[4]: *** [sqlrelayCmd.lo] Błąd 1
 
second try errors :
In file included from sqlrelayCmd.C:13:
../../../config.h:278:1: warning: "INLINE" redefined
In file included from sqlrelayCmd.C:7:
/usr/include/tcl.h:268:1: warning: this is the location of the previous 
definition
sqlrelayCmd.C: In function `int sqlrcurObjCmd(void*, Tcl_Interp*, int, 
Tcl_Obj*
   const*)':
sqlrelayCmd.C:345: syntax error before `;' token
sqlrelayCmd.C:357: syntax error before `=' token
sqlrelayCmd.C:357: `Tcl_WideInt' undeclared (first use this function)
sqlrelayCmd.C:357: (Each undeclared identifier is reported only once for 
each
   function it appears in.)
sqlrelayCmd.C:357: syntax error before `->' token
sqlrelayCmd.C:360: syntax error before `,' token
sqlrelayCmd.C:360: uninitialized const `col'
sqlrelayCmd.C:361: syntax error before `,' token
sqlrelayCmd.C:361: conflicting types for `uint32_t col'
sqlrelayCmd.C:360: previous declaration as `const char col'
sqlrelayCmd.C:361: warning: unused variable `uint32_t col'
sqlrelayCmd.C:1134: syntax error before `;' token
sqlrelayCmd.C:1143: syntax error before `)' token
sqlrelayCmd.C:1143: `Tcl_GetWideIntFromObj' undeclared (first use this
   function)
sqlrelayCmd.C:1146: syntax error before `)' token
sqlrelayCmd.C:1147: syntax error before `)' token
sqlrelayCmd.C:1160: syntax error before `;' token
sqlrelayCmd.C:1168: syntax error before `)' token
sqlrelayCmd.C:1171: syntax error before `)' token
make[4]: *** [sqlrelayCmd.lo] Błąd 1
 
make[3]: *** [all] Błąd 1
make[3]: Opuszczam katalog `/home/zope/sqlrelay/sqlrelay-0.37/src/api/tcl'
make[2]: *** [all] Błąd 2
make[2]: Opuszczam katalog `/home/zope/sqlrelay/sqlrelay-0.37/src/api'
make[1]: *** [all] Błąd 2
make[1]: Opuszczam katalog `/home/zope/sqlrelay/sqlrelay-0.37/src'
make: *** [all] Błąd 2

//-------------------------------------------------------------------------------------------------------
with --disable-tcl:

make -C cmdline all
make[2]: Wchodzę katalog `/home/zope/sqlrelay/sqlrelay-0.37/src/cmdline'
/bin/sh ../../libtool --mode=link g++  -Wall -pipe   -o query query.lo 
-L../../src/util -lsqlrutil -L../../src/api/c++/src -lsqlrclient 
-L/usr/local/firstworks/lib -lrudiments -lpthread -L/usr/kerberos/lib 
-lssl -lcrypto -lgssapi_krb5 -lkrb5 -lcom_err -lk5crypto -lresolv -ldl 
-lz -lrt -lrt -lcrypt -ldl
g++ -Wall -pipe -o .libs/query .libs/query.o  
-L/home/zope/sqlrelay/sqlrelay-0.37/src/util 
/home/zope/sqlrelay/sqlrelay-0.37/src/util/.libs/libsqlrutil.so 
-L/usr/local/firstworks/lib -L/usr/kerberos/lib 
-L/home/zope/sqlrelay/sqlrelay-0.37/src/api/c++/src 
/usr/local/firstworks/lib/libsqlrclient.so 
/usr/local/firstworks/lib/librudiments.so -lpthread -lssl -lcrypto 
-lgssapi_krb5 -lkrb5 -lcom_err -lk5crypto -lresolv -lz -lrt -lcrypt 
-ldl  -Wl,--rpath -Wl,/usr/local/firstworks/lib
.libs/query.o(.text+0x3a4): In function `main':
: undefined reference to `sqlrconnection::sqlrconnection[in-charge](char 
const*, unsigned short, char const*, char const*, char const*, int, int)'
.libs/query.o(.text+0x400): In function `main':
: undefined reference to `sqlrcursor::setResultSetBufferSize(unsigned 
long long)'
.libs/query.o(.text+0x4b1): In function `main':
: undefined reference to `sqlrcursor::getField(unsigned long long, 
unsigned)'
collect2: ld returned 1 exit status
make[2]: *** [query] Błąd 1
make[2]: Opuszczam katalog `/home/zope/sqlrelay/sqlrelay-0.37/src/cmdline'
make[1]: *** [all] Błąd 2
make[1]: Opuszczam katalog `/home/zope/sqlrelay/sqlrelay-0.37/src'
make: *** [all] Błąd 2

-- 
Maciej Wisniowski

>Hello all,
>
>I just uploaded 0.37 prerelease 2 to
>http://www.firstworks.com/sqlrelay-0.37pre2.tar.gz
>
>This pre-release includes fixes for several recently reported bugs and
>support for postgresql 8 bind variables.
>
>Since bind variables in postgresql 8 have different syntax than fake
>bind variables that had to be used with previous versions, there's a
>connectstring option fakebinds=yes for postgresql to force the
>connection to fake bind variables rather than use native ones so you
>don't have to rewrite your queries right away.
>
>David Muse
>dav...@fi...
>
>
>
>-------------------------------------------------------
>SF.Net email is Sponsored by the Better Software Conference & EXPO
>September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
>Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
>Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
>_______________________________________________
>Sqlrelay-discussion mailing list
>Sql...@li...
>https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
>
>  
>

Re: [Sqlrelay-discussion] Re: ZSQLRelayDA and freeing connections

[Maciej W. <mac...@co...>]

Hi!

We are using Zope 2.7.x and Zope 2.8.1 with SQLRelay 0.36.4
on Linux with Oracle and we didn't ecountered such problems.
We also have connections set to 3.

Maybe try to install DeadlockDebugger product for Zope and see
what it's threads are doing when it's "hanged".

-- 
Maciej Wisniowski

Jean Lagarde napisał(a):

> Oh yeah, I can make as many queries as I want using sqlrsh, so it 
> seems that the sqlrelay machinery is itself OK.
>
> -- Jean
>
> Jean Lagarde wrote:
>
>> Hi all,
>>
>> sqlrelay-0.35, Zope 2.7, Python 2.3, FreeBSD 5.0, FreeTDS
>>
>> I've got everything working up to this point: I can execute test 
>> queries, or run ZSQL methods, but only up to the "connections" 
>> parameter in the sqlrelay.conf file. For example, if connections="3", 
>> I can call the same ZSQL method three times, and the fourth time, 
>> Zope hangs until I run sqlr-stop and restart. Once I stop sqlrelay, 
>> the query that was hanging Zope returns with "failed to get cursor 
>> id". I went as far as checking that _finish() in db.py does seem to 
>> successfully call close() in PySQLRDB, but somehow I guess the 
>> previous connections/cursors aren't freed?
>>
>> Any ideas?
>>
>> Cheers,
>>
>> Jean
>>
>>
>>
>
>
>
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle 
> Practices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing 
> & QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> Sqlrelay-discussion mailing list
> Sql...@li...
> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
>

[Sqlrelay-discussion] Re: Connecting from Ruby to FreeTDS to Microsoft SQL Server 7.0

[Ben G. <me...@be...>]

Howdy,

Dave, thank you for your response. I just saw it today.

In the mean time, I've removed the code I complied as best I could and 
tried using only Debian packages obtained from apt-get.

Using a script similar to the one I previously posted (with or without a 
call to endSession()), I get:
"/usr/local/lib/site_ruby/1.8/i386-linux/sqlrelay.so: 
/usr/local/lib/site_ruby/1.8/i386-linux/sqlrelay.so: undefined symbol: 
__ti13genericsocket - 
/usr/local/lib/site_ruby/1.8/i386-linux/sqlrelay.so (LoadError)"

Would any of you know what's wrong? I tried Googling on this but didn't 
find help there.

Thank you
Ben

=========
From: Firstworks/4access <david@fi...>
Re: Connecting from Ruby to FreeTDS to Microsoft SQL Server 7.0   
2005-08-26 09:30  The only thing that stands out is the call to 
endSession(), but it
should be ok to call that as the result set should have been cached on
the client side.  Try removing that and see what you get.  Also, try
calling con.debugOn() after creating con and see if the output doesn't
reveal the problem.  If it doesn't, then send me the output and I'll
look at it too.

Dave
david.muse@fi...

On Thu, 2005-08-25 at 10:42 -0500, Ben Gribaudo wrote:
 > Hi,
 >
 > I am working on setting up a new Debian 3.1 server. So far, I can use
 > sqlrsh to query my database but I can't access the same sqlrelay connect
 > via Ruby.. Would any of you know what I'm doing wrong? I've even copied
 > my Ruby code to a Debian Woody server (which uses a Sybase driver
 > instead of FreeTDS) and it works.
 >
 > Thank you,
 > Ben Gribaudo
 >
 > Example: (usernane/password/socket/port replaced with dummy values)
 > ================================================
 > I can run this:
 > # sqlrsh localhost 8000 /tmp/name.socket user pass
 > SQLRShell - Version 0.22
 >        Connected to: localhost:8000 as user
 >
 >        type help; for a help.
 >
 > 0> select * from sysobjects;
 > name                           id        xtype uid info status      > 
base_schema_ver replinfo parent_obj crdate                   ftcatid
 > schema_ver stats_schema_ver type userstat sysstat indexdel
 > refdate                  version deltrig    instrig    updtrig    > 
seltrig category cache
 >
============================================================================================================================================================================================================================================================================================ 


 >
 > sysobjects                     1         S     1   25   -2147483645
 > 68              0        0          Fri Nov 13 03:00:19 1998 0       
 > 64         4                S    1        113  [...]
 >
 > but if I try to run
 >
 > require('sqlrelay')
 >
 > puts "HERE!"
 > 
con=SQLRConnection.new("localhost",8000,"/tmp/name.socket","user","pass",0,1) 

 >
 > cur=SQLRCursor.new(con)
 > cur.sendQuery("select * from sysobjects")
 > con.endSession()
 >
 > for row in 0..cur.rowCount()-1 do
 >        rowarray=cur.getRow(row)
 >            for col in 0..cur.colCount() do
 >                 puts rowarray[col]
 >              puts ","
 >           end
 >        puts "\n"
 > end
 >
 > all I get is "HERE!" and then I'm back at a shell prompt.
 >
 >
[snipped footer]

Re: [Sqlrelay-discussion] problem with sqlrelay 0.35-10 when I use INSERT query

[Firstworks/4access <dav...@fi...>]

That's very strange.

Are you sure you're not using a different user to log in when using psql
than is in the connection's parameter string in your sqlrelay.conf file,
or a different database maybe?

David Muse
dav...@fi...

On Tue, 2005-08-23 at 09:28 +0200, RAVELSON Tanjona wrote:
> Hi,
> 
> i'm trying to use sqlrelay with postgresql : here is
> the output when i launch the command as root
> ('sqlr-start -id pg_registry -config ./sqlrelay.conf')
> 
> Starting listener:
>   sqlr-listener -id pg_registry -config
> ./sqlrelay-config.conf
> 
> Starting 3 connections to pg_registry :
>   sqlr-connection-postgresql -id pg_registry
> -connectionid pg_registry -config
> ./sqlrelay-config.conf
>   sqlr-connection-postgresql -id pg_registry
> -connectionid pg_registry -config
> ./sqlrelay-config.conf
>   sqlr-connection-postgresql -id pg_registry
> -connectionid pg_registry -config
> ./sqlrelay-config.conf
> 
> Starting scaler:
>   sqlr-scaler  -id pg_registry -config
> ./sqlrelay-config.conf
> 
> Starting cache manager:
>   sqlr-cachemanager
> 
> 
> Thanks to MP3.com for sponsoring:
>         Clustered/Replicated database support.
>         Perl API.
> 
> and i think it run very well. I also use 'sqlrsh -id
> pg_registry' to connect to the instance (it's ok) and
> i test with some SELECT and UPDATE QUERIES (it's also
> ok). Whereas i'm running an INSERT QUERIE command,
> there is an issue (the  postgres's user, that sqlrelay
> use for his connection,  don't have permission to
> launch the sql command from sqlrelay) : 
> 
> sqlrsh -id pg_registry
> SQLRShell - Version 0.22
>         Connected to: localhost:9000 as pgtest
> 
>         type help; for a help.
> 0> insert into hosts(name) values('srv.nic.mg');
> ERROR:  permission denied for sequence
> objects_global_id_seq
> 
> Attempted Query:
> insert into hosts(name) values('srv.nic.mg')
>         Rows Returned   : 0
>         Fields Returned : 0
>         System time     : 50000
> 
> On the other hand, it's ok when i use psql with the
> same user as sqlrelay use for connection.
> 
> why the execution of UPDATE goes well whereas the
> INSERT is not?
> 
> thanks.
> 
> Note : i use the lastest version of debian (sarge) and
> postgresql 7.4.3-3 

[Sqlrelay-discussion] Re: Connecting from Ruby to FreeTDS to Microsoft SQL Server 7.0

[Ben G. <rub...@be...>]

Howdy,

Dave, thank you for your response. I just saw it today.

In the mean time, I've removed the code I complied as best I could and 
tried using only Debian packages obtained from apt-get.

Using a script similar to the one I previously posted (with or without a 
call to endSession()), I get:
"/usr/local/lib/site_ruby/1.8/i386-linux/sqlrelay.so: 
/usr/local/lib/site_ruby/1.8/i386-linux/sqlrelay.so: undefined symbol: 
__ti13genericsocket - 
/usr/local/lib/site_ruby/1.8/i386-linux/sqlrelay.so (LoadError)"

Would any of you know what's wrong? I tried Googling on this but didn't 
find help there.

Thank you
Ben

=========
From: Firstworks/4access <david@fi...>
 Re: Connecting from Ruby to FreeTDS to Microsoft SQL Server 7.0   
2005-08-26 09:30  
 The only thing that stands out is the call to endSession(), but it
 should be ok to call that as the result set should have been cached on
 the client side.  Try removing that and see what you get.  Also, try
 calling con.debugOn() after creating con and see if the output doesn't
 reveal the problem.  If it doesn't, then send me the output and I'll
 look at it too.
 
 Dave
 david.muse@fi...
 
 On Thu, 2005-08-25 at 10:42 -0500, Ben Gribaudo wrote:
 > Hi,
 >
 > I am working on setting up a new Debian 3.1 server. So far, I can use
 > sqlrsh to query my database but I can't access the same sqlrelay connect
 > via Ruby.. Would any of you know what I'm doing wrong? I've even copied
 > my Ruby code to a Debian Woody server (which uses a Sybase driver
 > instead of FreeTDS) and it works.
 >
 > Thank you,
 > Ben Gribaudo
 >
 > Example: (usernane/password/socket/port replaced with dummy values)
 > ================================================
 > I can run this:
 > # sqlrsh localhost 8000 /tmp/name.socket user pass
 > SQLRShell - Version 0.22
 >        Connected to: localhost:8000 as user
 >
 >        type help; for a help.
 >
 > 0> select * from sysobjects;
 > name                           id        xtype uid info status      
 > base_schema_ver replinfo parent_obj crdate                   ftcatid
 > schema_ver stats_schema_ver type userstat sysstat indexdel
 > refdate                  version deltrig    instrig    updtrig    
 > seltrig category cache
 >
============================================================================================================================================================================================================================================================================================

 >
 > sysobjects                     1         S     1   25   -2147483645
 > 68              0        0          Fri Nov 13 03:00:19 1998 0       
 > 64         4                S    1        113  [...]
 >
 > but if I try to run
 >
 > require('sqlrelay')
 >
 > puts "HERE!"
 > 
con=SQLRConnection.new("localhost",8000,"/tmp/name.socket","user","pass",0,1) 

 >
 > cur=SQLRCursor.new(con)
 > cur.sendQuery("select * from sysobjects")
 > con.endSession()
 >
 > for row in 0..cur.rowCount()-1 do
 >        rowarray=cur.getRow(row)
 >            for col in 0..cur.colCount() do
 >                 puts rowarray[col]
 >              puts ","
 >           end
 >        puts "\n"
 > end
 >
 > all I get is "HERE!" and then I'm back at a shell prompt.
 >
 >
[snipped footer]

Re: [Sqlrelay-discussion] New oracle7 driver build error with 0.36.4 and oracle 10.1.0 client?

[Firstworks/4access <dav...@fi...>]

I've seen something similar before.  On 64 bit machines, "unsigned long"
may be a 64 bit integer, which is the same as an "unsigned long long".
I've seen compilers make mistakes when generating function signatures.

Ie.  startSendingLong(unsigned long) gets turned into
startSendingLong(unsigned long long) and then programs fail to find
startSendingLong(unsigned long) at link time.

You might be running into a problem like that.  If you run:
nm --demangle src/connection/.libs/libsqlrconnection.so
and send me the result, I'll take a look and see if that's what's
going on.  If your nm doesn't support the --demangle option, you may
have to figure out what the c++ function name demangling option is for
your nm.

If that's what's going on, you might see if there's an update for your
compiler.

SQL Relay 0.37 will use (u)int(16|32|64)_t types rather than long/short
types and that should fix the problem.

There's a prerelease of sqlrelay/rudiments available at
http://www.firstworks.com/sqlrelay-0.37pre2.tar.gz

You might want to give it a try and see if you still have the problem.

Let me know what you find out.

David Muse
dav...@fi...

On Fri, 2005-08-26 at 07:15 -0700, Eric Rybski wrote:
> I've successfully built 64-bit SQL Relay using 0.36.3 with the following setup:
> 	Solaris 9 (Generic_117171-17)
> 	Oracle 10.1.0 Client
> 	Rudiments 0.28.2
> 
> I recently downloaded 0.36.4, and tried to perform an installation on the same
> system with the same configuration options, but ended up with the following
> error during make:
> 
> Undefined                       first referenced
>  symbol                             in file
> sqlrconnection::startSendingLong(unsigned long).libs/oracle7connection.o
> ld: fatal: Symbol referencing errors. No output written to
> sqlr-connection-oracle7
> 
> Has anyone else had this problem with this new version of SQL Relay?  Any ideas
> on how to resolve it?
> 
> I've included the entire event log below.
> 
> Regards,
> Eric
> 

[snip]

Re: [Sqlrelay-discussion] SQLRelay hangs with incorrect connection string

[Firstworks/4access <dav...@fi...>]

Looks like there's a bug in the sqlr-listener.

If you edit src/listener/sqlrlistener.C and look for the clientSession()
function, there's a bit of code like this in the function:

		// snooze before and after returning an
		// authentication error to discourage
		// brute-force password attacks
		snooze::macrosnooze(2);
		clientsock->write((unsigned short)ERROR);
		snooze::macrosnooze(2);

(or something similar)

insert a call to flushWriteBuffer(clientsock)
directly after the call to clientsock->write() and
it should fix the problem.

Dave
dav...@fi...

On Wed, 2005-08-31 at 10:53 +0200, Maciej Wisniowski wrote:
> Hi
> 
> I've just ecountered one problem with connection parameters to SQLRelay
> 
> . In sqlrelay.conf I have:
> 
> <instance id="xadm" port="9004"
>                socket="/tmp/xadm.socket"
>                dbase="oracle7" connections="3" maxconnections="5"
>                maxqueuelength="0"
>                growby="1" ttl="60" endofsession="commit"         
>                sessiontimeout="600" runasuser="nobody"
>                runasgroup="nobody" cursors="5"
>                authtier="listener" handoff="pass"
>                debug="None">
> <users>
>     <user user="test" password="test"/>
>     </users>
>     <connections>
>         <connection connectionid="xadm" 
> string="user=xadm;password=xadm;oracle_sid=devz" metric="1"/>
>     </connections>
> </instance>
> 
> 
> when using script like this:
> 
> // ------------------------------------------------------------
> from SQLRelay import PySQLRClient
> from SQLRelay import PySQLRDB
> 
> con=PySQLRDB.connect('localhost',9004,'','test','test',0,1)
> cur=PySQLRClient.sqlrcursor(con)
> res = cur.sendQuery('select * from xxx')
> // ------------------------------------------------------------
> 
> everything is OK (ORA exception is thrown (table xxx not known)
> and res == 0)
> but changing connection string to:
> 
> con=PySQLRDB.connect('localhost',9004,'','test','testx',0,1)
> 
> (here is incorrect password: testx) causes script to
> hang waiting for response...
> 
> This is PythonAPI and SQLRelay 0.36.4.
> Is this a bug or my config file or script is wrong?
> 

[Sqlrelay-discussion] SQLRelay hangs with incorrect connection string

[Maciej W. <mac...@co...>]

Hi

I've just ecountered one problem with connection parameters to SQLRelay

. In sqlrelay.conf I have:

<instance id="xadm" port="9004"
               socket="/tmp/xadm.socket"
               dbase="oracle7" connections="3" maxconnections="5"
               maxqueuelength="0"
               growby="1" ttl="60" endofsession="commit"         
               sessiontimeout="600" runasuser="nobody"
               runasgroup="nobody" cursors="5"
               authtier="listener" handoff="pass"
               debug="None">
<users>
    <user user="test" password="test"/>
    </users>
    <connections>
        <connection connectionid="xadm" 
string="user=xadm;password=xadm;oracle_sid=devz" metric="1"/>
    </connections>
</instance>


when using script like this:

// ------------------------------------------------------------
from SQLRelay import PySQLRClient
from SQLRelay import PySQLRDB

con=PySQLRDB.connect('localhost',9004,'','test','test',0,1)
cur=PySQLRClient.sqlrcursor(con)
res = cur.sendQuery('select * from xxx')
// ------------------------------------------------------------

everything is OK (ORA exception is thrown (table xxx not known)
and res == 0)
but changing connection string to:

con=PySQLRDB.connect('localhost',9004,'','test','testx',0,1)

(here is incorrect password: testx) causes script to
hang waiting for response...

This is PythonAPI and SQLRelay 0.36.4.
Is this a bug or my config file or script is wrong?

-- 
Maciej Wisniowski

Re: [Sqlrelay-discussion] [PATCH] Few tweaks to rudiments

[Dmitry S. <mi...@ca...>]

On Fri, Aug 26, 2005 at 12:37:16PM -0400, Firstworks/4access wrote:
> Ahhh, so on solaris, strchr/strstr/strrchr all return const char *
> rather than char *.  That's actually very cool.  It's always bugged


Probably not. From man strstr:
char *strstr(const char *s1, const char *s2);

I can provide you access to Solaris 10 box (SPARC) if you wish.



> me that glibc's versions of those functions return char * and I've been
> looking for an excuse to make my findFirst/findLast methods return a
> const char *.  Now I have one :)
> 
> I guess I need versions of them that take a char * first parameter
> though :).
> 
> I'll fix that in the next prerelease.
> 
> Dave
> 
> On Fri, 2005-08-26 at 19:23 +0400, Dmitry Sivachenko wrote:
> > On Fri, Aug 26, 2005 at 11:02:33AM -0400, David Muse wrote:
> > > Yes, the http://www.firstworks.com/sqlrelay-0.37pre2.tar.gz tarball
> > > contains both rudiments and sqlrelay and both contain
> > > sun-compiler-specific updates.
> > 
> > Well, I see.
> > 
> > Now I get:
> > 
> >  CC -pipe -D_REENTRANT -D_THREAD_SAFE -D__EXTENSIONS__ -I../ -I../include -c signalclasses.C -o signalclasses.o >/dev/null 2>&1
> > /bin/sh ../libtool --mode=compile CC -pipe  -D_REENTRANT -D_THREAD_SAFE -D__EXTENSIONS__     -I../ -I../include -c charstring.C -o charstring.lo
> >  CC -pipe -D_REENTRANT -D_THREAD_SAFE -D__EXTENSIONS__ -I../ -I../include -c charstring.C  -KPIC -DPIC -o .libs/charstring.o
> > CC: Warning: Option -pipe passed to ld, if ld is invoked, ignored otherwise
> > "charstring.C", line 740: Error: Cannot return const char* from a function that should return char*.
> > "charstring.C", line 744: Error: Cannot return const char* from a function that should return char*.
> > "charstring.C", line 748: Error: Cannot return const char* from a function that should return char*.
> > 3 Error(s) detected.
> > make[1]: *** [charstring.lo] Error 1
> > make[1]: Leaving directory `/tmp/rudiments-0.29/src'
> > make: *** [all] Error 2
> > 
> > 
> > 
> > I use Sun Studio 10.
> > 
> > 
> > > 
> > > Dave
> > > 
> > > On Fri, 2005-08-26 at 18:30 +0400, Dmitry Sivachenko wrote:
> > > > Please note that I submitted a patch for *rudiments*, not for sqlrelay.
> > > > 
> > > > 
> > > > 
> > > > On Fri, Aug 26, 2005 at 10:24:49AM -0400, Firstworks/4access wrote:
> > > > > Take a look at http://www.firstworks.com/sqlrelay-0.37pre2.tar.gz
> > > > > 
> > > > > I recently made a bunch of changes to support sun compilers including a
> > > > > check to see if the compiler supports -Wall at configure-time.  I
> > > > > believe it also includes these fixes.
> > > > > 
> > > > > I'm still working on some errors involving different compatibility
> > > > > modes, but hopefully they'll be fixed soon.
> > > > > 
> > > > > David Muse
> > > > > dav...@fi...
> > > > > 
> > > > > 
> > > > > On Fri, 2005-08-26 at 11:40 +0400, Dmitry Sivachenko wrote:
> > > > > > Hello!
> > > > > > 
> > > > > > I propose the following patch to rudiments (made against 0.28.2).
> > > > > > 
> > > > > > Patch to configure.in is to eliminate mandatory '-Wall' from CPPFLAGS
> > > > > > since there are compilers (say Sun Studio) which do not recongnize
> > > > > > this flag.  One can always add this switch to their CPPFLAGS if desired.
> > > > > > 
> > > > > > Patch to charstring.h fixes the inconsistency between declarations in
> > > > > > charstring.h and src/charstring.C. This is actually a bug which gcc hides,
> > > > > > I suppose.
> > > > > > 
> > > > > > Thanks!
> > > > > > 
> > > > > > 
> > > > > > --- include/rudiments/charstring.h.orig	Чтв Янв 27 06:35:07 2005
> > > > > > +++ include/rudiments/charstring.h	Птн Авг 26 11:34:38 2005
> > > > > > @@ -119,7 +119,7 @@
> > > > > >  							const char *needle);
> > > > > >  			// Returns true if "haystack" contains "needle" or
> > > > > >  			// false otherwise.
> > > > > > -		static	bool	contains(const char *haystack, char needle);
> > > > > > +		static	bool	contains(const char *haystack, const char needle);
> > > > > >  			// Returns true if "haystack" contains "needle" or
> > > > > >  			// false otherwise.
> > > > > >  		static	char	*findFirst(const char *haystack,
> > > > > > @@ -126,7 +126,7 @@
> > > > > >  							const char *needle);
> > > > > >  			// Returns a pointer to the first occurrance of "needle"
> > > > > >  			// in "haystack" or NULL if not found.
> > > > > > -		static	char	*findFirst(const char *haystack, char needle);
> > > > > > +		static	char	*findFirst(const char *haystack, const char needle);
> > > > > >  			// Returns a pointer to the first occurrance of "needle"
> > > > > >  			// in "haystack" or NULL if not found.
> > > > > >  		static	char	*findLast(const char *haystack,
> > > > > > @@ -133,7 +133,7 @@
> > > > > >  							const char *needle);
> > > > > >  			// Returns a pointer to the last occurrance of "needle"
> > > > > >  			// in "haystack" or NULL if not found.
> > > > > > -		static	char	*findLast(const char *haystack, char needle);
> > > > > > +		static	char	*findLast(const char *haystack, const char needle);
> > > > > >  			// Returns a pointer to the last occurrance of "needle"
> > > > > >  			// in "haystack" or NULL if not found.
> > > > > >  
> > > > > > --- configure.in.orig	Чтв Апр 21 18:42:05 2005
> > > > > > +++ configure.in	Птн Авг 26 11:31:13 2005
> > > > > > @@ -288,7 +288,7 @@
> > > > > >  	INCLUDE_ERROR="1")
> > > > > >  
> > > > > > 
> > > > > > -CPPFLAGS="-Wall -D_REENTRANT $CPPFLAGS"
> > > > > > +CPPFLAGS="-D_REENTRANT $CPPFLAGS"
> > > > > >  CXXFLAGS="$SMALLCODE $CXXFLAGS"
> > > > > >  
> > > > > >  AC_SUBST(CPPFLAGS)
> > > > > > 
> > > > > > 
> > > > > > -------------------------------------------------------
> > > > > > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > > > > > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> > > > > > Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> > > > > > Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> > > > > > _______________________________________________
> > > > > > Sqlrelay-discussion mailing list
> > > > > > Sql...@li...
> > > > > > https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
> > > > > > 
> > > > > 
> > > > > 
> > > > > 
> > > > > -------------------------------------------------------
> > > > > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > > > > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> > > > > Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> > > > > Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> > > > > _______________________________________________
> > > > > Sqlrelay-discussion mailing list
> > > > > Sql...@li...
> > > > > https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
> > > > 
> > > > 
> > > > -------------------------------------------------------
> > > > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > > > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> > > > Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> > > > Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> > > > _______________________________________________
> > > > Sqlrelay-discussion mailing list
> > > > Sql...@li...
> > > > https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
> > > > 
> > > 
> > > 
> > > 
> > > -------------------------------------------------------
> > > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> > > Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> > > Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> > > _______________________________________________
> > > Sqlrelay-discussion mailing list
> > > Sql...@li...
> > > https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
> > 
> > 
> > -------------------------------------------------------
> > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> > Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> > Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> > _______________________________________________
> > Sqlrelay-discussion mailing list
> > Sql...@li...
> > https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
> > 
> 
> 
> 
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> Sqlrelay-discussion mailing list
> Sql...@li...
> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion