|
PDTUM wrote: the app actually crashes Windows 7 if I do not shut it down immediately following failure
The OS actually crashes versus you receiving an OS exception?
|
|
|
|
|
Hello js,
Thank you for your question....Yes. Since I made the earlier post, I have installed 10g on the local server and everything works perfectly. Although I have gotten great help from the people here, I still have the original issue in this case, so I just uploaded a post to another forum as I am still baffled by this. I will post the new text below. If you have any ideas...I am ALL ears... ...Thanks, Pat
I have a Windows 7 dev machine used to create .Net applications with VS 2010. I want to connect to a remote network server (using Windows 2003) that is running an Oracle 10g database. I also need to mention (just in case) that the remote server is also hosting an MS SQL 2005 application as well. I am using the code below (with the first connection string) to connect directly to a version of 10g that is running on the same machine with no problems, however when I try to connect to the network machine, it actually crashes Windows 7. I have tried several variations of connection strings as I feel that I must be making a sytax error somewhere. What concerns me is that I have dual try/catch statements in the application and I do not understand why it simply does not refuse the connection and report the error. I suppose the real question is 'what is the correct syntax for the connection string'....or WHATEVER the hell I am doing wrong. This is my first go around with Oracle (and I am not liking it at all), but it is not a matter of choice for me. SO.........any help or suggestions are greatly appreciated. Thank you in advance...Best Regards, Pat
string CONNSTR = "Server=192.168.2.5:1521;User ID=Patt;Password=Mine;";
public Oracle()
{
InitializeComponent();
}
private void TestMyOracleConnection()
{
OracleConnection Conn = new OracleConnection(CONNSTR);
try
{
Conn.Open();
MessageBox.Show("Oracle Connection Established", "Success");
}
catch (OracleException ex)
{
MessageBox.Show(ex.Message, "Oracle Connection Failed!");
}
catch (Exception ex)
{
MessageBox.Show(ex.Message, "Oracle Connection Failed!");
}
finally
{
Conn.Close();
MessageBox.Show("Oracle Connection Closed", "Success");
}
}
private void buttonTestConnection_Click(object sender, EventArgs e)
{
TestMyOracleConnection();
}
|
|
|
|
|
An OS crash is a pretty severe problem and one which has nothing to do with your code.
It just simply shouldn't be possible.
Some possible avenues.
1. Try it on another machine. Perhaps you have some significant failure (like hardware) on your current box.
2. Try a different Oracle driver, if that is possible.
3. Insure the OS is up to date.
Solutions:
1. Is the target run machine going to be windows 7 (or server 2008)? If not then use a different OS for development.
2. Use a proxy server. I would suggest java. That is going to require a lot of work. But your app connects to the proxy app, it does the database work, and then returns the result.
The last is a rather severe solution but so is an OS crash.
|
|
|
|
|
Guys,
In a few days I have a meeting internally about redesigning the Oracle database from scratch.
The downside is that none of the meeting participants is a real DBA (IMHO we should hire a DBA consultant for advice), but we do have some knowledge about Oracle.
one of my major concerns is that someone opted for multiple smaller databases (that should communicate if necessary) I can think of multiple reasons why this is a bad idea, but I couldn't find any satisfactory links to proove it. (Maybe it isn't that bad after all?)
I did write some stuff down already (never delete, but rather 'inactivate', use history mechanism etc..)
In short: What do I need to watch out for when designing a new database ?
If it can help, we're an insurance company, but we also do call taking (could result in some tables having rapid changes when a lot of calls come in) Also we will start with adding one project to it, then a second, a third etc... so no one time porting to another database.
Many thanks in advance.
V.
|
|
|
|
|
V. wrote: IMHO we should hire a DBA consultant for advice
This is what you should do, getting a good one is always difficult and expensive. Having recently been dipped in Oracle for the first time in over a decade I can tell you the design and setup is VASTLY more difficult than sql server. The cost of getting a database design wrong enormously outweighs the initial expense of getting in a DBA.
For one, I'm told Oracle does not like multiple databases on a server, the design is to use schemas, I thought this was pure bullshit but it came from a DBA so it may be right.
For another Oracle is case sensitive when dealing with data so where name = 'Johny' will miss 'johny' IMHO this and the ongoing support requirement enough reason to use another database (sql server). The ONLY reason to use Oracle is if your data is so huge that SQL Server chokes. If your design spec says < 1tb per year I would use another database.
Never underestimate the power of human stupidity
RAH
|
|
|
|
|
Mycroft Holmes wrote: For one, I'm told Oracle does not like multiple databases on a server, the design is to use schemas, I thought this was pure bullsh*t but it came from a DBA so it may be right.
It's BS. It works quite fine. It's just pointless. Multiple schemas is the way to go.
Mycroft Holmes wrote: For another Oracle is case sensitive when dealing with data so where name = 'Johny' will miss 'johny' IMHO this and the ongoing support requirement enough reason to use another database (sql server).
And the problem is what? Use a function based index and you got the best of both worlds.
Example:
CREATE INDEX foo_bar_ix
ON foo (
LOWER("bar")
)
/
And then you can search the table with Select * from Foo where lower(bar) = lower('Johnny');
"When did ignorance become a point of view" - Dilbert
|
|
|
|
|
Jörgen Andersson wrote: Use a function based index and you got the best of both worlds
Nice. I didn't know you could do that. A 5 just for that trick.
Jörgen Andersson wrote: It's BS. It works quite fine. It's just pointless
I suspect he's referring to the load on having multiple instances running as opposed to a single instance running multiple schema.
|
|
|
|
|
Pete O'Hanlon wrote: Nice. I didn't know you could do that. A 5 just for that trick.
That single trick alone is worth the extra trouble of using Oracle
"When did ignorance become a point of view" - Dilbert
|
|
|
|
|
Materialised Views. They are worth it as well. Mind you, we have a top notch Oracle DBA to make sure that I don't have to know how to do this.
|
|
|
|
|
Pete O'Hanlon wrote: Materialised Views. They are worth it as well
Agreed.
But sqlserver has indexed views.
Pete O'Hanlon wrote: Mind you, we have a top notch Oracle DBA to make sure that I don't have to know how to do this.
I guess that's the major problem with Oracle, that you need one.
"When did ignorance become a point of view" - Dilbert
|
|
|
|
|
My biggest problem with materialised views is that we can't use any SDO geometry types in them (something that would be really useful for us).
|
|
|
|
|
Are you sure about that?
I understand that it's severely limited[^]. But not impossible.
"When did ignorance become a point of view" - Dilbert
|
|
|
|
|
It's point 9 that's the killer. We have some data that comes in as x,y coords and we want our materialised views to be spatial, so you get screwed up when you try to construct a location sdo geometry.
|
|
|
|
|
Spatial is totally out of my line of work, but I would have thought point one or ten would have been the real killer.
I'm curious, what would have been the purpose of the mv if it had been possible?
"When did ignorance become a point of view" - Dilbert
|
|
|
|
|
We are reliant on a separate supplier for a standardised gazetteer implementation in one particular project (the data's provided by the government), but we need it in a different format (a spatial one for spatial searches). We need the view to be updated on import of data into the master gazetteer. We've worked around the issue, but it would be nice not to have to work around the issue.
|
|
|
|
|
I've been thinking about this one for a while.
If I understand you correctly (and it's quite possible I don't), you're getting XY coords from an external source that you load into a table, and you want them transformed into a sdo_geometry datatype which you do via a materialized view. And for some reason you can't do this at the import.
One possible solution is to add a virtual column into the master gazetteer that calculates the sdo_geometry.
From 11g you can index, gather statistics and partition by a virtual column, which makes them a lot faster and quite useful nowadays.
There are probably other limitations that makes it impossible to drop the view, but nothing I could find on a quick google.
"When did ignorance become a point of view" - Dilbert
|
|
|
|
|
Pete O'Hanlon wrote: we have a top notch Oracle DBA to make sure that I don't have to know how to do this.
I wish. Difficult to find and expensive to retain. I'm in the throes of getting to know Oracle again after more than a decade in SS, I think SS must be like VB, lots of work to make it easy for the developer.
Then I aggregated 150m records in minutes where SS was taking hours and remembered why we are using Oracle.
Never underestimate the power of human stupidity
RAH
|
|
|
|
|
I've balanced out the univote.
|
|
|
|
|
Oh, hadn't noticed. Thanks!
"When did ignorance become a point of view" - Dilbert
|
|
|
|
|
Jörgen Andersson wrote: It's BS. It works quite fine. It's just pointless. Multiple schemas is the way to go.
I suspected as much, it was probably a DBA enforcing the schema requirement.
Jörgen Andersson wrote: And then you can search the table with Select * from Foo where lower(bar) = lower('Johnny');
I will need to do some work to understand the relevance on the function based index. However I'm used to SS where case is irrelevant when comparing text, there for the lower() is not required. This will continue to bite me until I get used to it.
My real peeve with Oracle is the all upper case objects, I hate underscores so my names look like FILENAMEDSOMETHING instead of FileNamedSomeThing.
Never underestimate the power of human stupidity
RAH
|
|
|
|
|
Mycroft Holmes wrote: I will need to do some work to understand the relevance on the function based index
Say for example that you have a log table where one of the columns has the date datatype which carries both date and time, and you want to get all occurences from one specific day.
A query for trunc(logdate) would make a full table scan, unless you have an index on trunc(logdate).
Another example, You have a really large table where you have a status column, and the only value you ever make a search on is 'PENDING'. Then an index on "Case When status = 'PENDING' Then 1 Else Null" will be very small and fast, as only those entries where the status is PENDING will be stored in the index as null values are not indexed.
The backside is that function based indexes is costing more to maintain.
"When did ignorance become a point of view" - Dilbert
|
|
|
|
|
Thanks for that Jorgen
While the ability to do these tweaks is very good, the requirement to do them is painful. I beging to understand why a full time DBA is a requirement, I consider this type of tuning beyond the requirements of a developer.
Never underestimate the power of human stupidity
RAH
|
|
|
|
|
Mycroft Holmes wrote: , the requirement to do them is painful
Having to use Oracle?
Mycroft Holmes wrote: beyond the requirements of a developer
The learning threshold is indeed high, especially if you come from the world of SqlServer.
BTW, I liked your comparison with VB, had it been more catchy I would have stolen it for a sig 
"When did ignorance become a point of view" - Dilbert
|
|
|
|
|
V. wrote: What do I need to watch out for when designing a new database ?
I'm with Mycroft, try to save a dime here and it'll cost a fortune down the road.
V. wrote: but we do have some knowledge about Oracle
I have some knowledge about cars, but that doesn't qualify me to design a new one. People that are taking design-decisions should bring forth rational arguments for their proposals.
I are Troll 
|
|
|
|
|
What's the reasoning behind getting several databases?
If it's getting slow because of large dataamounts you should have a look at partitioning instead.
Get a good DBA for the project to give you proper advice. It'll pay back later.
"When did ignorance become a point of view" - Dilbert
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
