OutOfMemoryException with SQL Server Mobile

Hi,

I'm getting always an OutOfMemoryException after my program runs for a while. It's a smartphone application running on .NET 2.0 CF and using the SQL Server Mobile 2005. Sometimes it's even not a OutOfMemoryException but a SqlCeException. But it's always when the phone is having little free memory left.

I have also created a few custom controls (I even implemented the Disposable interface) to dispose all resources (Fonts, Bitmaps). I'm using all Pens and Brushes always in "using" blocks, to have them disposed immediately after having used them. The program does not have any static objects or lists of objects that are hold in memory for a long time.

I really don't understand why I get this exception. I see it always happen after doing a lot queries on the SQL Server Mobile. Is this tool using a lot memory? Can I flush somehow the memory of the SQL Server? Should I run also manually sometimes the GarbageCollector? Does this thing not run automatically?

Can anybody, perhaps of MSFT help me? I'm also willing to share the full source code with MSFT, if required.

You will probably get a faster answer on the Sql Server Mobile forum.

Moving this thread to that forum.

|||

Here are a few suggestions (from an earlier thread)

1) Try to use SqlCeResultSet than DataSet

2) Please dispose the managed objects explicitly instead of waiting for garbage collector to pick them up. Please also note that SqlCeCommand object can be reused and they need to be disposed explicitly

In addtion you can email me your project and I'll be glad to look into this issue for you.

Mark dot Ihimoyan at microsoft dot com

Outer join runs differently on SQL Server and Oracle

Oracle 9i:
create table t(i integer)
Table created
insert into t values(1)
1 row inserted
select t1.i i1, t2.i i2
from t t1 left join t t2 on 1=0
0 rows selected
-- I beleive this is wrong
drop table t
Table dropped
the same query against MS SQL Server 2000:
create table t(i integer)
insert into t values(1)
select t1.i i1, t2.i i2
from t t1 left join t t2 on 1=0
i1 i2
-- --
1 NULL
-- I think this is correct
(1 row(s) affected)
drop table t
What do you thinkAK wrote:
> Oracle 9i:
> create table t(i integer)
> Table created
> insert into t values(1)
> 1 row inserted
>
> select t1.i i1, t2.i i2
> from t t1 left join t t2 on 1=0
> 0 rows selected
> -- I beleive this is wrong
> drop table t
> Table dropped
> the same query against MS SQL Server 2000:
> create table t(i integer)
> insert into t values(1)
> select t1.i i1, t2.i i2
> from t t1 left join t t2 on 1=0
> i1 i2
> -- --
> 1 NULL
> -- I think this is correct
> (1 row(s) affected)
> drop table t
> What do you think
I think you're using an unpatched release of Oracle 9i, whichever
release that may be (9i says NOTHING about which release or patch level
you're using):
Connected to:
Oracle9i Enterprise Edition Release 9.2.0.6.0 - 64bit Production
With the Partitioning, OLAP and Oracle Data Mining options
JServer Release 9.2.0.6.0 - Production
SQL> create table t(i integer);
Table created.
SQL>
SQL> insert into t values(1);
1 row created.
SQL>
SQL>
SQL> select t1.i i1, t2.i i2
2 from t t1 left join t t2 on 1=0;
I1 I2
-- --
1
SQL>
The results look exactly the same to me between the two. Care to try
again, this time with a patched release of Oracle?
David Fitzjarrell|||AK schrieb:
> Oracle 9i:
> create table t(i integer)
> Table created
> insert into t values(1)
> 1 row inserted
>
> select t1.i i1, t2.i i2
> from t t1 left join t t2 on 1=0
> 0 rows selected
> -- I beleive this is wrong
> drop table t
> Table dropped
> the same query against MS SQL Server 2000:
> create table t(i integer)
> insert into t values(1)
> select t1.i i1, t2.i i2
> from t t1 left join t t2 on 1=0
> i1 i2
> -- --
> 1 NULL
> -- I think this is correct
> (1 row(s) affected)
> drop table t
> What do you think
>
Oracle ansi join implementation had some bugs ( much of them fixed in
recent versions/patches ), but in your case i definitely can't reproduce
your behaviour ( both on 9iR2 and 10gR1/R2)
oracle@.col-fc1-02:~/sql >sqlplus scott/tiger
SQL*Plus: Release 9.2.0.6.0 - Production on Thu Aug 11 19:19:29 2005
Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.
Connected to:
Oracle9i Enterprise Edition Release 9.2.0.6.0 - Production
With the Partitioning, OLAP and Oracle Data Mining options
JServer Release 9.2.0.6.0 - Production
SQL> create table t(i integer);
Table created.
SQL> insert into t values(1);
1 row created.
SQL> select t1.i i1,t2.i i2
2 from t t1 left join t t2 on 1=0;
I1 I2
-- --
1
SQL>
Here is output for 10gR1
oracle@.wks01:~> sqlplus scott/tiger
SQL*Plus: Release 10.1.0.3.0 - Production on Do Aug 11 18:05:22 2005
Copyright (c) 1982, 2004, Oracle. All rights reserved.
Connected to:
Oracle Database 10g Enterprise Edition Release 10.1.0.3.0 - Production
With the Partitioning, OLAP and Data Mining options
SQL> create table t(i integer);
Table created.
SQL> insert into t values(1);
1 row created.
SQL> select t1.i i1,t2.i i2
2 from t t1 left join t t2 on 1=0
3 /
I1 I2
-- --
1
SQL>
Best regards
Maxim|||yep, one more reason to patch up:
Connected to:
Oracle9i Release 9.2.0.1.0 - Production
JServer Release 9.2.0.1.0 - Production
Thanks for the feedback

OSQL returns error code of 1 - intermittently

Hello,
I've written an program that runs OSQL to set up a database. I got a weird
situation yesterday where OSQL returns an error level of 1 and didn't
generate any output file at all. The command I use is:
OSQL.EXE -E -b -i SetupRTEDatabase.sql -o SetupRTEDatabase.log
Normally, this command runs OK and produces a .log file. Why would it
return 1 and no log file?
I call OSQL from a Borland Builder C++ program.
Any ideas?
Regards,
Robert
PS: I ran the OSQL command straight after I installed MSDE and started up
SQL Server.Hi,
you are assuming some things:
OSQL.EXE -E -b -i SetupRTEDatabase.sql -o SetupRTEDatabase.log
-E are you really able to use integrated authentication, if not you
have to provide a user name and a password. Perhaps you are running
your program under a special account which isn=B4t priviledged in the
database ?!
-i Are you in the right folder that SOQL can find the query file while
executing the file from the program ?
(missing -S) If you don=B4t specify a server, the default behaviour is
to go to the local server, the default instance, to the default port
1433. If this is not valid, you have to define the server/indtance,port
e=2Eg. Servername\instancename,Port.
HTH, Jens Suessmeyer.|||Hi,
Thanks for your reply.
OSQL.EXE -E -b -i SetupRTEDatabase.sql -o SetupRTEDatabase.log
-E are you really able to use integrated authentication, if not you
have to provide a user name and a password. Perhaps you are running
your program under a special account which isnt priviledged in the
database ?!
I think so. I am running as (machine) administrator when I execute OSQL
-i Are you in the right folder that SOQL can find the query file while
executing the file from the program ?
The SetupRTEDatabase.sql script is in the same directory as the calling
program.
(missing -S) If you dont specify a server, the default behaviour is
to go to the local server, the default instance, to the default port
1433. If this is not valid, you have to define the server/indtance,port
e.g. Servername\instancename,Port.
This is OK. I run OSQL pretty much straight after I do an install of SQL
Server. I don't change any of the default settings.
Regards,
Robert|||I've fixed the problem.
As it turns out, I was running the program in the wrong folder after all.
OSQL was being run (indirectory) from a Setup program which ran on a network
drive.
I did a ChangeDirectory to the path with the SetupRTEDatabase.sql script and
everything worked fine.
"Robert Wheadon" <robert.wheadon@.monitorbm.co.nz> wrote in message
news:uC$01qrTGHA.5108@.TK2MSFTNGP11.phx.gbl...
> Hi,
> Thanks for your reply.
> OSQL.EXE -E -b -i SetupRTEDatabase.sql -o SetupRTEDatabase.log
> -E are you really able to use integrated authentication, if not you
> have to provide a user name and a password. Perhaps you are running
> your program under a special account which isnt priviledged in the
> database ?!
> I think so. I am running as (machine) administrator when I execute OSQL
> -i Are you in the right folder that SOQL can find the query file while
> executing the file from the program ?
> The SetupRTEDatabase.sql script is in the same directory as the calling
> program.
> (missing -S) If you dont specify a server, the default behaviour is
> to go to the local server, the default instance, to the default port
> 1433. If this is not valid, you have to define the server/indtance,port
> e.g. Servername\instancename,Port.
> This is OK. I run OSQL pretty much straight after I do an install of SQL
> Server. I don't change any of the default settings.
> Regards,
> Robert
>

osql return to check backup success/failure

I am running an osql using input script which runs a database backup. Is
there a way I could get a return of 1 or 0, 0 being success and 1 being
error. I tried using -b option but it returns the entire details of failure
followed by the return number and if i use -o paramater I dont see any
result at all. Can anyone please suggest some solution to this. any help
will be greatly appreciated. Thanks
Message posted via http://www.webservertalk.comTry catching the error in the batch being executed with OSQL and use
RAISERROR to set the DOS variable ERRORLEVEL.
Example:
Create a batch file with the following instructions:
osql -S my_server -E -d master -Q "backup database northwind to blahblah if
@.@.error != 0 raiserror ('Error during backup.', 16, 127)"
echo %ERRORLEVEL%
Execute the batch file and you will see 50000 as the echo of the variable
ERRORLEVEL, because I am specifying a msg_str in RAISERROR. You can create
your own error msg using sp_addmessage.
AMB
"ishaan99 via webservertalk.com" wrote:

> I am running an osql using input script which runs a database backup. Is
> there a way I could get a return of 1 or 0, 0 being success and 1 being
> error. I tried using -b option but it returns the entire details of failur
e
> followed by the return number and if i use -o paramater I dont see any
> result at all. Can anyone please suggest some solution to this. any help
> will be greatly appreciated. Thanks
> --
> Message posted via http://www.webservertalk.com
>|||I forgot to mention that in order to set DOS variable ERRORLEVEL, using
RAISERROR, you have to specify state 127.
Error Handling in SQL Server – a Background
http://www.sommarskog.se/error-handling-I.html
Implementing Error Handling with Stored Procedures
http://www.sommarskog.se/error-handling-II.html
AMB
"Alejandro Mesa" wrote:
> Try catching the error in the batch being executed with OSQL and use
> RAISERROR to set the DOS variable ERRORLEVEL.
> Example:
> Create a batch file with the following instructions:
> osql -S my_server -E -d master -Q "backup database northwind to blahblah i
f
> @.@.error != 0 raiserror ('Error during backup.', 16, 127)"
> echo %ERRORLEVEL%
> Execute the batch file and you will see 50000 as the echo of the variable
> ERRORLEVEL, because I am specifying a msg_str in RAISERROR. You can create
> your own error msg using sp_addmessage.
>
> AMB
>
> "ishaan99 via webservertalk.com" wrote:
>|||Thanks so much. That helped and worked fine.
Message posted via http://www.webservertalk.com

OSQL Batch File Problem

I have a scheduled task that runs a batch file in windows 2003 server which
is running SQL Server 2000 Enterprise.
The batch file contains the line:
OSQL -i myscript.sql -Umyuser -Pmypassword -Slocalhost -o mytrans.log
the 'myscript.sql' file updates a table with another table of the exact
number of rows.
When the scheduled task runs, the 'mytrans.log' file shows 0 rows were
affected.
If I run this myself from the command line, 22,000 rows are affected which
is correct.
I am guessing there is some sort of permission/authentication issue here. I
am sending the right username/password for SQL and for windows to run the
task.
Any idea what I could do to fix this?
Thanks,
J. BaezGo to schedule tasks -- Tasks tab--See the user in RUN AS: . The user
displayed there should be having write access to the folder you are writing
the log file.
After that just right click and run the task and see.
Thanks
hari
SQL Server MVP
"J. Baez" wrote:
> I have a scheduled task that runs a batch file in windows 2003 server which
> is running SQL Server 2000 Enterprise.
> The batch file contains the line:
> OSQL -i myscript.sql -Umyuser -Pmypassword -Slocalhost -o mytrans.log
> the 'myscript.sql' file updates a table with another table of the exact
> number of rows.
> When the scheduled task runs, the 'mytrans.log' file shows 0 rows were
> affected.
> If I run this myself from the command line, 22,000 rows are affected which
> is correct.
> I am guessing there is some sort of permission/authentication issue here. I
> am sending the right username/password for SQL and for windows to run the
> task.
> Any idea what I could do to fix this?
> Thanks,
> J. Baez|||That user already has full control of that directory. So it still not
working. Any other ideas?
"Hari Prasad" wrote:
> Go to schedule tasks -- Tasks tab--See the user in RUN AS: . The user
> displayed there should be having write access to the folder you are writing
> the log file.
> After that just right click and run the task and see.
> Thanks
> hari
> SQL Server MVP
> "J. Baez" wrote:
> > I have a scheduled task that runs a batch file in windows 2003 server which
> > is running SQL Server 2000 Enterprise.
> >
> > The batch file contains the line:
> >
> > OSQL -i myscript.sql -Umyuser -Pmypassword -Slocalhost -o mytrans.log
> >
> > the 'myscript.sql' file updates a table with another table of the exact
> > number of rows.
> >
> > When the scheduled task runs, the 'mytrans.log' file shows 0 rows were
> > affected.
> >
> > If I run this myself from the command line, 22,000 rows are affected which
> > is correct.
> >
> > I am guessing there is some sort of permission/authentication issue here. I
> > am sending the right username/password for SQL and for windows to run the
> > task.
> >
> > Any idea what I could do to fix this?
> >
> > Thanks,
> >
> > J. Baez

OSQL Batch File Problem

I have a scheduled task that runs a batch file in windows 2003 server which
is running SQL Server 2000 Enterprise.
The batch file contains the line:
OSQL -i myscript.sql -Umyuser -Pmypassword -Slocalhost -o mytrans.log
the 'myscript.sql' file updates a table with another table of the exact
number of rows.
When the scheduled task runs, the 'mytrans.log' file shows 0 rows were
affected.
If I run this myself from the command line, 22,000 rows are affected which
is correct.
I am guessing there is some sort of permission/authentication issue here. I
am sending the right username/password for SQL and for windows to run the
task.
Any idea what I could do to fix this?
Thanks,
J. BaezTry fully using fully qualified path for the input file -
C:\SomeFolder\mysqlscript.sql
or whatever the path is.
-Sue
.
On Tue, 22 Aug 2006 05:26:01 -0700, J. Baez
<JBaez@.discussions.microsoft.com> wrote:

>I have a scheduled task that runs a batch file in windows 2003 server which
>is running SQL Server 2000 Enterprise.
>The batch file contains the line:
>OSQL -i myscript.sql -Umyuser -Pmypassword -Slocalhost -o mytrans.log
>the 'myscript.sql' file updates a table with another table of the exact
>number of rows.
>When the scheduled task runs, the 'mytrans.log' file shows 0 rows were
>affected.
>If I run this myself from the command line, 22,000 rows are affected which
>is correct.
>I am guessing there is some sort of permission/authentication issue here.
I
>am sending the right username/password for SQL and for windows to run the
>task.
>Any idea what I could do to fix this?
>Thanks,
>J. Baez|||It is really, really helpful if you post ONLY one copy of your request.
Posting in multiple locations WITHOUT crossposting loses you the benefit of
our 'bouncing' ideas off of each other.
Arnie Rowland, Ph.D.
Westwood Consulting, Inc
Most good judgment comes from experience.
Most experience comes from bad judgment.
- Anonymous
"J. Baez" <JBaez@.discussions.microsoft.com> wrote in message
news:90E8F840-025C-4FFE-8926-55FC25F94567@.microsoft.com...
>I have a scheduled task that runs a batch file in windows 2003 server which
> is running SQL Server 2000 Enterprise.
> The batch file contains the line:
> OSQL -i myscript.sql -Umyuser -Pmypassword -Slocalhost -o mytrans.log
> the 'myscript.sql' file updates a table with another table of the exact
> number of rows.
> When the scheduled task runs, the 'mytrans.log' file shows 0 rows were
> affected.
> If I run this myself from the command line, 22,000 rows are affected which
> is correct.
> I am guessing there is some sort of permission/authentication issue here.
> I
> am sending the right username/password for SQL and for windows to run the
> task.
> Any idea what I could do to fix this?
> Thanks,
> J. Baez
>|||Yeah...I noticed that after I replied. I think there are all
different answers, directions trying to help out with this
one on at least a few different groups. And no one knows who
is suggesting what.
It also makes it hard for others to find the answer when
they search things in the future. One post has the answer,
others don't, you only find the one without the solution,
etc.
-Sue
On Tue, 22 Aug 2006 13:19:49 -0700, "Arnie Rowland"
<arnie@.1568.com> wrote:

>It is really, really helpful if you post ONLY one copy of your request.
>Posting in multiple locations WITHOUT crossposting loses you the benefit of
>our 'bouncing' ideas off of each other.

OSQL Batch File Problem

I have a scheduled task that runs a batch file in windows 2003 server which
is running SQL Server 2000 Enterprise.
The batch file contains the line:
OSQL -i myscript.sql -Umyuser -Pmypassword -Slocalhost -o mytrans.log
the 'myscript.sql' file updates a table with another table of the exact
number of rows.
When the scheduled task runs, the 'mytrans.log' file shows 0 rows were
affected.
If I run this myself from the command line, 22,000 rows are affected which
is correct.
I am guessing there is some sort of permission/authentication issue here. I
am sending the right username/password for SQL and for windows to run the
task.
Any idea what I could do to fix this?
Thanks,
J. BaezGo to schedule tasks -- Tasks tab--See the user in RUN AS: . The user
displayed there should be having write access to the folder you are writing
the log file.
After that just right click and run the task and see.
Thanks
hari
SQL Server MVP
"J. Baez" wrote:

> I have a scheduled task that runs a batch file in windows 2003 server whic
h
> is running SQL Server 2000 Enterprise.
> The batch file contains the line:
> OSQL -i myscript.sql -Umyuser -Pmypassword -Slocalhost -o mytrans.log
> the 'myscript.sql' file updates a table with another table of the exact
> number of rows.
> When the scheduled task runs, the 'mytrans.log' file shows 0 rows were
> affected.
> If I run this myself from the command line, 22,000 rows are affected which
> is correct.
> I am guessing there is some sort of permission/authentication issue here.
I
> am sending the right username/password for SQL and for windows to run the
> task.
> Any idea what I could do to fix this?
> Thanks,
> J. Baez|||That user already has full control of that directory. So it still not
working. Any other ideas?
"Hari Prasad" wrote:
[vbcol=seagreen]
> Go to schedule tasks -- Tasks tab--See the user in RUN AS: . The user
> displayed there should be having write access to the folder you are writin
g
> the log file.
> After that just right click and run the task and see.
> Thanks
> hari
> SQL Server MVP
> "J. Baez" wrote:
>

osgl syntax problem

I have a script that runs fine in SQL Query Analyzer but has a syntax error
using osql. It restores several databases and log files. One of the database
s
has a space in the database name. Similar to RESTORE DATABASE "XXXX XXXX"
FROM DISK (BLAH). When run using osql all the databases restore correctly
except for the database with the space in the name. SQL Query Analyzer has n
o
problems with the statement. Is there any way to make this work or do I have
to rename the database with no spaces?
--
Mark Koenig
Tucson, AZIdeally, you should have no spaces in object names, or you get the problem
you encountered. Try:
SET QUOTED IDENTIFIER ON
GO
Tom
----
Thomas A. Moreau, BSc, PhD, MCSE, MCDBA
SQL Server MVP
Columnist, SQL Server Professional
Toronto, ON Canada tom@.cips.ca
www.pinpub.com
"KMK" <kkoenig@.email(DOT)arizona(DOT)edu> wrote in message
news:BFFEDC2D-C3BE-4365-8369-F6F1C7D6D875@.microsoft.com...
>I have a script that runs fine in SQL Query Analyzer but has a syntax error
> using osql. It restores several databases and log files. One of the
> databases
> has a space in the database name. Similar to RESTORE DATABASE "XXXX XXXX"
> FROM DISK (BLAH). When run using osql all the databases restore correctly
> except for the database with the space in the name. SQL Query Analyzer has
> no
> problems with the statement. Is there any way to make this work or do I
> have
> to rename the database with no spaces?
> --
> Mark Koenig
> Tucson, AZ|||It's been awhile since I've used osql but with embedded spaces, you might
need to use brackets. i.e. [My Database] Perhaps, enclose in double quote
marks. Or, single and double quote marks.
If this was a stored procedure, a fully qualified name would be:
[My Server].[My Databse].[My Owner].[sp Select Data]
I did some batch osql about a year ago and I had to send filenames into SQL
Server to generate and execute some bulk insert statements.
In DOS, I passed the filename to a common script that called SQL Server. In
the calling script, I passed the filename as "'C:\Program Files...TXT'".
Note, the single quotes inside the double quotes. Then in the common script
that called SQL Server, I had the following line:
@.osql -E -S %SQLServerName% -d %Database% -Q "exec %1 %~2"
Note: Parameter %1 contained the name of the stored procedure to execute.
Parameter %~2 is the fully qualified name of the input file.
In the stored procedure, I received the filename as varchar(128). When I
concatenated it into the bulk insert statement, I had to enclose it in singl
e
quote marks in order to accomodate embedded spaces. For example, the
statement ended up as BULK INSERT...FROM 'C:\PROGRAM FILES...TXT'. To embed
single quote marks just repeat it. So it's EXEC ('bulk insert ' + @.TableNam
e
+ ' from ''' + @.DataFilename + ''' with (tablock, firstrow = ' +
@.StartRowNumber + ', formatfile = ''' + @.FormatFilename + ''', maxerrors = '
+ @.MaxNumberOfErrors + ')')
Bottom line, you need double quote marks to pass a filename with embedded
spaces from one DOS/batch command script to another. (I don't know if you'r
e
doing this part or not.) Then, it has to be enclosed in single quote marks
for the osql statement. (At least it did for me using the -Q (query)
parameter.) Then, if your concatenating it in a stored procedure, you need
to embed the incoming value in single quote marks.
Hope that helps,
Joe
"KMK" wrote:

> I have a script that runs fine in SQL Query Analyzer but has a syntax erro
r
> using osql. It restores several databases and log files. One of the databa
ses
> has a space in the database name. Similar to RESTORE DATABASE "XXXX XXXX"
> FROM DISK (BLAH). When run using osql all the databases restore correctly
> except for the database with the space in the name. SQL Query Analyzer has
no
> problems with the statement. Is there any way to make this work or do I ha
ve
> to rename the database with no spaces?
> --
> Mark Koenig
> Tucson, AZ|||This did work after using SET QUOTED_IDENTIFIER ON. Your response left out
the underline. Thanks for the info!
--
Mark Koenig
Tucson, AZ
"Tom Moreau" wrote:

> Ideally, you should have no spaces in object names, or you get the problem
> you encountered. Try:
> SET QUOTED IDENTIFIER ON
> GO
>
> --
> Tom
> ----
> Thomas A. Moreau, BSc, PhD, MCSE, MCDBA
> SQL Server MVP
> Columnist, SQL Server Professional
> Toronto, ON Canada tom@.cips.ca
> www.pinpub.com
> "KMK" <kkoenig@.email(DOT)arizona(DOT)edu> wrote in message
> news:BFFEDC2D-C3BE-4365-8369-F6F1C7D6D875@.microsoft.com...
>
>