PostgreSQL DB for multiple clients

[MoNDoGuY]

Hi guys, I gave the forum a good search I couldn't find anything that reflected our setup. So here is our situation.

Myself and 3 or 4 friends have been playing poker online for little while and we've recently installed PT3 to analyze our games. One of my friends is a statistical genius, and a tool like this goes a long way in his hands. We want a setup that makes it easy for our math genius to access all our data and analyze it.

So our plan is setup a remote postgres DB and have all our data dump to that server. From my testing and research it seems that we can' tall simultaneously dump our data to the same DB, but make separate ones for each of us. So if I'm understanding this correctly, the math genius would need to bounce from DB to DB to run and export the reports.

Is there a more efficient way of running this type of setup? Are we over complicating things for nothing?

Thanks.

[kraada]

If you give him all of your original hand history files as you go, he can import them into his one database for analysis, then look at all stats combined should that be easiest. Otherwise, your proposed setup will work but you are correct he'll need to go from database to database and he won't be able to run any global data.

[MoNDoGuY]

Firstly, you guys have awesome support. Thanks for the quick response

So another option would be to FTP our hand histories over to his server. Or is it the processed files? Then he would have to run a manual import. This would create one big database with all our data in one place. Is this correct?

Thanks again.

[WhiteRider]

That's correct. The processed files ARE the hand histories - they're just moved after they've been imported so that they don't have to be read again next time you start auto import.

[MoNDoGuY]

Thanks again guys. I'd like to play out this process from the math geniuses (Simon) perspective just to make sure I have this right.

- Individually we would all play online with PT3 running like normal. Importing the data in our own local DB's.
- Once done playing we FTP our processed files to the main server.
- Simon would have to go into PT3 and change the active DB from his local DB on his local Postgres server, to the remote server's GroupDB.
- Run the Manual Import Process on the files that have been FTP'd.
- Upon completion, individually, if we want to see all the data in the GroupDB we would have to go into PT3 and change our active database from our local one to the remote server.

Thats my understanding of the setup required. Do I have that correct?

Fantastic support guys. Keep up the great work.

[kraada]

That's correct. Expect things to be a little slower when using the remote database though just due to network latency.

[MoNDoGuY]

Awesome thank you! I'll be setting up some VM's to test this setup. I'll be back if I run into issues

[MoNDoGuY]

High guys. I have a couple of more questions about this setup.

Firstly, the remote server we're going to setup has a 2.8g Pentium Celeron CPU. Do you think this will be sufficient? I'm aware your minimum requirements require a P4. Considering this remote server, running XP, is strictly for historical analysis of our stats I'm thinking this is ok. It might be a little slow, but not horrible. Or are we setting ourselves up for disappointment.What are you thoughts?

Second, is there a way to setup a batch file that can run the manual import process? I know postgres has some command line functionality, but I'm no DBA. So for example,

1. Our files are stored on C:\History.
2. Our global team DB is called PokerTeam.
3. We run a batch file that imports all the hand histories from C:\history into the PokerTeam DB
4. The batch file deletes the content of C:\history upon completion.

I'm basically trying to remove any chance of human error and effort required from our math genius.

Thanks guys.

[tarix]

Firstly, the remote server we're going to setup has a 2.8g Pentium Celeron CPU. Do you think this will be sufficient?

You are going to have an exceptional amount of latency with this solution. PT3 expects requests to and from the server to be at zero latency. Having your server remote infinitely increases this. In addition the database server is the most responsible for doing all of the analysis.

If you are going to have one person analyze the data then he should import all of the data into a local database.

[MoNDoGuY]

Thanks for the feedback.

To clarify, our remote server will be located at the math geniuses house, so it will be connected to his local LAN. His network latency should be minimal. That being said, I assume the celeron CPU is going to be a big bottleneck?