My program enables user's to play online against other players. the way I achieve this is by having a lobby form with a datagridview displaying all the matches in a datatable entitled "matches". from their you can select a match (send a request) or host one which will direct you to a "waiting" form and insert your match into the database. The problem is that this database fills up fast with old matches because I have coded it to delete records on form close however this does not run if the process (whole program) is ended from say task manager and I am left with an old game in the database. my question is how can I ensure that when a user stops hosting, their corresponding match is deleted. The only solution I can think of involves having a server constantly pinging all the matches and deleting them if they do not respond.
Your idea of the server pinging the client is a good one. But it can be reversed. For example, by using a TCP connection, you can send the username or ID of the player to the server every two minutes through the client. If the player's client has not pinged the server in two minutes, the server can assume that the player is offline and delete the old 'match' instead of the client. You can run the server dedicated and it will always work well.
For more information on TCP data communication, just google it!
Hope this helped,
Rodit
Related
I have an application that connects with a SQL Server database and cycles through batches of records to perform various tasks and then updates the database accordingly (i.e. "success", "error", etc...).
The potential problem I'm running into is that since it takes roughly a minute or so to get through each record (long story), if I have more than one user running the application there's a high chance of "data collisions" or users trying to process the same records at the same time. Which cannot happen if it is to execute properly.
Initially, I thought about adding a LOCKED column to help the application determine if the record was already opened by another user, however if the app were to crash or to be exited without completing the record it was currently on, then it would show that record as opened by another user indefinitely... right? Or am I missing an easy solution here?
Anyway, what would be ideal is if it were possible to have the application SELECT 100 records at a time, and "lock them out" on the database while the application processes them AND so that other users can run the application and SELECT a different set of 100 so as not to overlap. Is that possible? I've tried to do some research on the matter, but to be honest my experience in SQL Server is very limited. Thanks for any and all help!
There are Many SQL Servers hosted on different different Servers.
All Servers are working based on "SQL Server Authentication". So the Same Login is used by many people in the Organization.
How to trace who deleted some of the records in particular table?
Do we need any additional coding like Triggers are required or its a in-build feature of SQL server to provide those details?
Please help me.
Thank You.
If the deletion has already occurred and you had nothing in place to track / log this, then the chances are going to be very low - they are not zero, but not far above it.
If you use the transaction log to identify the exact deletion and the session id of the deletion, which we already know is the shared user login - and you have got successful login security auditing enabled you would in theory be able to trace it back to the IP address that made the deletion.
However - that is a pretty slim chance - I would suspect that the login is from the actual application software and you would of needed that to be running directly on the users machine, e.g not a 3-tier / web based server of any flavor, but a good old thick client app making direct connections.
That gets you an IP and a time, but not a who was logged in on that machine at that time, if its shared in any form, then you are having to get login records on the machine etc.
I've got a virtual world in Adobe Flash AS2 using SmartFoxServerPro. One problem, everytime a user logs back in, his data resets, well the server never saves it. My question is: how to save permanent data to a server, so even when other people join, they'd see it. For example, if Player1 places a block from his inventory, then the following week he re-joins the game, he'd still see the block, cause the server would have saved hat block to it's map. Another example, how to make the server automatically have mobs on it, so that when user log in, they'd have monsters waiting for them, and even when the mobs, life is 50%, all players would see his life as 50%.
That is my question.
You need to persist your game state in a database and load the information when your user logs in to the game.
More info on databases and smartfox here:http://docs2x.smartfoxserver.com/DevelopmentBasics/database-recipes
Just after some advice and recommendations here if I may.
I'm creating an iPad app (IOS6) that will write data to the local database on the device and then either straight away or later on replicate that record to a web service (so a Cloud service basically).
What is the best way to go about this you think?
I was thinking of just having a column in the local DB called "synced" and set the flag to '0' right away when the record is created, then sync records with a '0' either right away or during regular intervals. Then obviously set the flag to '1' when each record is replicated.
I want the app to work offline and then sync when the device has an available connection to my web service.
Ideally every record should be replicated right away or seconds later, but in the event of no network connectivity I want to be able to queue the replication to occur.
So what's the best way or achieving this you think? Thanks in advance :)
The solution is going to depend a lot on how complex your total solution is.
For example, if the records are only being created on the local device and then uploaded, without ever being modified, then your solution will be more than adequate.
However, if you allow update of the records on the local device or the records can be updated once they get into your web service, then you need to start managing conflict resolution. The way that we address this situation is to record a timestamp in the "master" database (the one updated by the web service) and synchronize that timestamp when a record is uploaded either as a new record or as an update. When the user updates a record, we send the timestamp and if the value in the database is different than the sent database, the update request is rejected. Of course there are different approaches to this conflict resolution, this is just one that works for our application and users.
I have the following situation:
-I have a server A hooked up to a piece of hardware that sends values and information out of every second. Programs on the server machine can read these values. This server A is in a very remote location so Internet connection is very slow and not reliable but the connection does exist. Let's say it's a weather station in the Arctic.
-Users from the home location want to monitorize the weather values somehow. Well, the users can use a remote desktop connection the server A but that would be too too slow.
My idea is somehow to have a website on a web server (let's call the webserver - B and B is in a home location ) and make the server A connect to the server B and somehow send values and the web application reads the values and displays them....... but how to do such a system ?? I know I can use MySQL and have the server A connect to a SQL server on server B and send INSERT queries and have the web application running on server B constantly read from the SQL server but I think that would be way way too slow and I think there has to be a better solution.
Any ideas ?
BTW. The users should be able to send information to the weather station from the website as well ( so an ADMIN user should be allowed to shut down the weather station from the website or whatever)
Best regards,
MadSeb
Ganglia (http://ganglia.sourceforge.net/) is a popular monitoring tool that supports collecting arbitrary statistics using the gmetric tool. You might be able to build something around that.
If you do need to roll your own solution then you could have a persistent message queue at A (I'm a fan of RabbitMQ) to which you could log your metrics. You could then have something at B which listens for messages on the queue and saves the state at B.
This approach means you don't lose data when the connection drops.
The message might be a simple compressed data value, say csv or json so should be fine on low bandwidth connections.
All the work (parsing the csv or json, and saving the data to a database for example) is done at B where you don't have limitations.