I have an old intranet website written with vb.net and using Oracle.ManagedDataAccess mostly doing read operations from Oracle database 11g.
My db connection code is as follows.
Public Shared Function MyDBConnection(ByVal command_text As String, ByVal connstring As String, ByVal ParamArray parameters As OracleParameter()) As DataTable
Dim OraCommand As New OracleCommand
Dim tmp As New DataTable
Using ORAconnection = New OracleConnection(ConfigurationManager.ConnectionStrings(connstring).ConnectionString)
OraCommand.CommandText = command_text
If parameters IsNot Nothing AndAlso parameters.Length > 0 Then
For Each p In parameters
OraCommand.Parameters.Add(p.ParameterName, p.OracleDbType, p.Size).Value = p.Value
Next p
End If
OraCommand.Connection = ORAconnection
Try
ORAconnection.Open()
tmp.Load(OraCommand.ExecuteReader())
Catch ex As OracleException
End Try
End Using
Return tmp
End Function
And my oracle connection string is like this
<add name="ConnectionString" connectionString="User Id=userid;Password=userpadd;Data Source=servername:port/port_dp"/>
I was testing if my connections to the database were closing properly but it looks like connections on the database stayed open after being closed on my code. Eventually they close way after my query completed about 2 minutes, 10 minutes or hours later.
Is this connection pooling at work? or if there is something wrong with my code?
After reading about oracle pooling, it looks like the application should be reusing the same opened connection on the DB but in my case it looks like is opening new connections anyway.
So my question is, should I disable pooling on my connection string to make sure all connections open/close and not have connections lingering on the DB?
No, you should not disable connection pooling. In .NET connection pooling is managed by a mechanism outside of your reach, and you should proceed as if it is not there: open and close your connections as you normally would (i.e. at the beginning and end of every set of operations that you wish to enroll in a transaction/every set of ops that defines a good "unit of work" such as running a report, updating a table etc
The action of opening and closing connections in your code simply leases the from/returns them to the pool. .NET will manage the rest regarding maintaining a cache of some open connections to the db
Yes, there's pooling. The number of connection might have to do with the default value of min and max pool size. Also, OracleCommand does have a Dispose method that you are not calling.
Added: I see your empty catch statement, you don't need it when using "Using" (or ever). The Dispose will still be called if there's an exception.
Related
I am trying to run a non query using a Oracle connection in ASP C# with CLR 4.5. Here is my code:
string connectionString = ConfigurationManager.ConnectionStrings["OracleConnectionString1"].ConnectionString;
OracleConnection conn = new OracleConnection(connectionString);
conn.Open();
OracleCommand cmd = new OracleCommand();
cmd.Connection = conn;
cmd.CommandText = "update SALES_ADVENTUREWORKS2012.SALESORDERDETAIL set UNITPRICEDISCOUNT=0 where ROWGUID='4A399178-C0A0-447E-9973-6AB903B4AECD'";
cmd.CommandType = CommandType.Text;
cmd.CommandTimeout = QUERY_TIMEOUT;
int row_affected = cmd.ExecuteNonQuery();
HttpContext.Current.Response.Write("Rows affected:" + row_affected + "<br/>");
conn.Close();
when I run the query in oracle development tool, it works fine.
when I use the asp code above, it freezes when performing the query. It freezes forever even though I used a 5 second timeout.
I've tried using the managed and unmanaged oracle libraries; both behave the same.
Note that using the fill or scalar query work perfectly fine so there is nothing wrong with my connection string. Also the fact that oracle development can perform this update query proves that this is not a permission problem.
Any ideas?
Most likely your query is waiting to get access to the record. You probably have modified that row in "oracle development tool" and have not committed or rolled back that transaction.
Just commit/rollback in your tool or close open session.
You can check for open transactions in v$transaction view.
More on automatic locks in Oracle:
http://docs.oracle.com/cd/E11882_01/server.112/e41084/ap_locks001.htm
Are you certain you are using the 4.5 library? The 3.5 documentation states that the CommandTimeout property has no effect.
The 4.5 documentation suggests it should work, but the Remarks section doesn't mention the change, which warrants suspicion.
Otherwise, the code you posted doesn't seem to show where you actually set the value of QUERY_TIMEOUT to 5 seconds. If QUERY_TIMEOUT has a value of zero, then any other provider (SQLCommand, for example) would wait indefinitely. As vav suggested, locks from other sources could cause an indefinite wait.
I got an asp.net gridview connected to my sql database. When Inserting a new record or updating a record im doing some serverside checks and then either update/insert a record or do nothing. right now i got 2 methods CheckArtistExists and CheckSongExists which are both using a SqlConnection Object e.g.
public bool CheckSongExists(string _title, int _artistId)
{
int cnt = -1;
using (SqlConnection con = new SqlConnection(CS))
{
//check if song already is exists in DB
SqlCommand cmd = new SqlCommand("Select Count(ID) from tblSong WHERE Title = #newTitle AND ArtistId = #newArtistId;", con);
cmd.Parameters.AddWithValue(#"newTitle", _title);
cmd.Parameters.AddWithValue(#"newArtistId", _artistId);
con.Open();
cnt = (int)cmd.ExecuteScalar();
// if cnt ==1 song exists in DB, of cnt == 0 song doesnt exist
if(cnt == 1)
{ return true; }
else
{ return false; }
}
}
So for the Update function in the gridview i need to establish 3 SqlConnections (at max) one to check for the artist(if artist doesnt exist i have to insert a record to tblArtist first)
then a check if the song exists(only if artist exists) and finally if song doesnt exist I have to insert a new record.
I know database connections are valuable resources thats why i put them in a using block. So im not quite sure if its good style to use 3 SqlConnection objects to update/insert. Can you please tell me if my code is ok or if i should rather use another approach for this problem.
thank you
ADO.NET internally manages the underlying Connections to the DBMS in the ADO-NET Connection-Pool:
In practice, most applications use only one or a few different
configurations for connections. This means that during application
execution, many identical connections will be repeatedly opened and
closed. To minimize the cost of opening connections, ADO.NET uses an
optimization technique called connection pooling.
Connection pooling reduces the number of times that new connections
must be opened. The pooler maintains ownership of the physical
connection. It manages connections by keeping alive a set of active
connections for each given connection configuration. Whenever a user
calls Open on a connection, the pooler looks for an available
connection in the pool. If a pooled connection is available, it
returns it to the caller instead of opening a new connection. When the
application calls Close on the connection, the pooler returns it to
the pooled set of active connections instead of closing it. Once the
connection is returned to the pool, it is ready to be reused on the
next Open call.
So obviously there's no reason to avoid creating,opening or closing connections since actually they aren't created, opened and closed at all. This is "only" a flag for the connection pool to know when a connection can be reused or not. But it's a very important flag, because if a connection is "in use"(the connection pool assumes), a new physical connection must be openend to the DBMS what is very expensive.
So you're gaining no performance improvement if you "reuse" connections but the opposite.
Create, open(in case of Connections), use, close and dispose them where you need them(f.e. in a method)
use the using-statement to dispose and close(in case of Connections) implicitely
So yes, it's absolutely fine to use one connection per method since you are not using a physical connection at all if connection-pooling is enabled (default).
Another question is if you could improve your approach. You could create a stored-procedure which checks existence and updates or inserts accordingly.
Solutions for INSERT OR UPDATE on SQL Server
I see a couple of DataContext connection string questions. I'm going to try to differentiate this one a bit:
How does one construct a generic connection string to a database, localhost | User-PC\User | Some database... (it is hosted/managed by Microsoft SQL 2008)
I notice that it is IDisposable. So if I have multiple users hitting my site, my code can only access the database one instance at a time, and has to wait until each instance is disposed, in order for the data to be consistent for each user?
Is it possible, by any chance, to somehow enable LINQ in F#-Interactive, and connect to the database from there? I cannot figure out how to enable/load the System.Data dll into fsi. Maybe that is unique to my installation, or it is a common thread? (ie, my installation also does not recognize windows.base.dll--I have to manually get it from programs\reference assemblies).
Anyhow, I've pretty much conclusively discovered that
let x = new System.Data.Linq.DataContext("localhost")
...does not work.
1) How does one construct a generic connection string to a database?
There is no generic way to construct a connection string. The best thing to do is to keep the connection string in some configuration file where you can change it depending on your configuration (the name of SQL Server machine, authentication options, whether it is a file-based database or normal). There is a web site with examples for most of the options.
2) I notice that it is IDisposable. So if I have multiple users hitting my site, my code can only access the database one instance at a time [...]?
No, this is not how DataContext works. The DataContext does not keep a live connection to the server that would block anybody else from using the SQL server. It keeps some state (i.e. cached entities that were already obtained) and it uses optimistic concurrency to make sure that the state is consistent (you can use transactions to prevent other connections, if that's what you want).
3) Is it possible, by any chance, to somehow enable LINQ in F#-Interactive [...]?
That shouldn't be a problem. You can reference assemblies using #r "foo.dll" in F# interactive. The typical approach for F# 2.0 is to generate the data context using C# tools and then just reference it (for F# 3.0, things are easier because you can just use type provider).
If you generate LINQ to SQL data context for Northwind in C#, the F# Interactive use would look like this:
#r #"<whatever_path>\Northwind.dll"
#r "System.Data.Linq.dll"
open Northwind
open Microsoft.FSharp.Linq
let connStr = #"Data Source=.\SQLEXPRESS;AttachDbFilename=<path>\NORTHWND.MDF;" +
#"Integrated Security=True;User Instance=True"
let operation () =
// Using 'use' to make sure it gets disposed at the end
use db = new NorthwindDataContext(connStr)
// do something with the database
There actually is a somewhat generic way to construct a connection string:
open System.Data.Common
open System.Data.SqlClient
let providerName = "System.Data.SqlClient"
let factory = DbProviderFactories.GetFactory(providerName)
let cnBuilder = factory.CreateConnectionStringBuilder() :?> SqlConnectionStringBuilder
cnBuilder.DataSource <- "localhost"
cnBuilder.InitialCatalog <- "MyDatabase"
cnBuilder.IntegratedSecurity <- true
let connStr = cnBuilder.ConnectionString
My approach was to have 1 connection string and then use that for all of my DataContext connections. So this code builds the EntityConnectionString based on MyConnString:
protected override MyEntities CreateObjectContext()
{
string ConnString =ConfigurationManager.ConnectionStrings["MyConnString"];
string seConn = ConfigurationManager.ConnectionStrings["MyEntities"].ToString();
EntityConnectionStringBuilder ecsb = new EntityConnectionStringBuilder(seConn);
ecsb.ProviderConnectionString = ConnString;
EntityConnection ec = new EntityConnection(ecsb.ToString());
ScheduleEntities ctx = new ScheduleEntities(ec);
return ctx;
}
I'm having difficulty with an SQL query against Server 2008 from IIS7. I have a VB.NET class library which runs an update statement. The underlying code used to create the connection hasn't changed, but suddenly the query is failing in our testing and development environments. It does, however, still work against the same server/database using the slightly older code in our production environment.
I've tried setting the connection timeout in the web.config and I'm at a loss to explain the cause.
The basic structure of the query is:
Dim conn = New SqlConnection()
conn.ConnectionString = "Data Source=someserver\sqlexpress2008;Initial Catalog=DatabaseName;User ID=sa;Password=pass"
conn.Open()
Using cmd As SqlCommand = conn.CreateCommand()
cmd.CommandText = "UPDATE ..."
cmd.Parameters.AddWithValue("#UName", user.name)
cmd.ExecuteNonQuery() 'fails with error
End Using
The error is:
A transport-level error has occurred when sending the request to the
server. (provider: TCP Provider, error: 0 - An existing connection was
forcibly closed by the remote host.)
I've tried restarting IIS and the SQL server and I'm totally out of ideas. I just need a fix
You need to open the connection before calling SqlCommand.ExecuteNonQuery(). You do this by calling SqlConnection.Open().
Dim conn = New SqlConnection()
conn.ConnectionString = "Data Source=someserver\sqlexpress2008;Initial Catalog=DatabaseName;User ID=sa;Password=pass"
Using cmd As SqlCommand = conn.CreateCommand()
cmd.CommandText = "UPDATE ..."
cmd.Parameters.AddWithValue("#UName", user.name)
conn.Open()
cmd.ExecuteNonQuery() 'fails with error
conn.Close()
End Using
Also, ensure you database isn't in single user mode.
This helped another person who was stuck recently. You could examine the problem from the database server by setting up a SQL Server Profiler.
You can find lots of info about SQL Profiler by just googling around. Here's a site with a video that might help you get started. For starters, you would be able to see if the request is even reaching the database server.
This was a nightmare to track down. It turned out to be cause by a horrible quirk in VB.NET. Nullable datetimes seem to be coerced to DateTime.MinValue, which resulted in a DateTime.MinValue being inserted into an sql datetime. The fix was to check for either !property.HasValue && property.Value != DateTime.MinValue when setting the parameters for the command.
This is a network-level error. The database server is killing the connection for some reason. In order to troubleshoot this, I would open a connection using SSMS to the DEV and TEST servers and make sure that I can run simple queries w/o problems. It's unlikely that the issue is your library since you would be getting timeout or some other kind of errors.
as Lcarus, said, database server is killing the connection for unknown reason.
you can check the logs, to verfiy. Log path will be C:\Program Files\Microsoft SQL Server\<your instance>\MSSQL\LOG
from MSDN Blog MSDN Blog
this will occur when A connection is taken from the connection pool,
the application does not know that the physical connection is gone, an
attempt to use it is done under the assumption that the physical
connection is still there.
I have a dropdown box and a literal tag inside an Update Panel. On the selection change event of the dropdown up requery the database and repopulate the literal tag and then call UPdatePanel.Update().
below, is there are way i can avoid having to create a new Oledbconnection each time as this seems slow. Can i reuse and store:
The Datasource
The connection in the page.
if so, how do i keep this state between calls from the GUI to the server? Here is my selection change code below
protected void cboPeople_SelectedIndexChanged(object sender, EventArgs e)
{
string dataSource = ConfigurationSettings.AppSettings["contactsDB"];
var objConn = new OleDbConnection(dataSource);
string id = People[cboPeople.Text];
UpdateLiteral(objConn, id);
}
With .NET is not a good idea to keep your connection alive longer than needs. Good practice would be to put a using statement around it (so it always gets cleaned up):
string dataSource = ConfigurationSettings.AppSettings["contactsDB"];
using(var objConn = new OleDbConnection(dataSource))
{
string id = People[cboPeople.Text];
UpdateLiteral(objConn, id);
}
.NET uses connection pooling, which means that when you close/dispose of the connection it doesn't actually close the connection, rather resets it and add it back to the pool. The next time a connection is needed it is used from the pool. So the overhead is not as much as you think and it is not slow. In actual fact you will find that it will use the same connection as long as only one at a time is needed.
The danger with keeping connections open is that they never get closed, and in high demand situations you run out of connections.
You need to recreate this for each request. You have a a state less server. you never know when or if your client will call back. You do not want to keep an open connection to the database nor could you simply service multiply clients while maintaining one database connection.
To deploy high-performance
applications, you must use connection
pooling. When you use the .NET
Framework Data Provider for OLE DB,
you do not have to enable connection
pooling because the provider manages
this automatically. For more
information about how to use
connection pooling with the .NET
Framework Data Provider for OLE DB,
see OLE DB, ODBC, and Oracle
Connection Pooling (ADO.NET).
From OleDbConnection Class