cluster

[Suerte]

I have just started two fresh dbs. I have manually imported about 3million hands into them. I set it so observed go to 1 and played go to db 2.

It imported overnight, and has been in the clustering stage for 5 hours now. Is this normal for it to take so long to cluster? How long will it take? I have a good cpu with fast stuff on it. TY.

[kraada]

The first one can take a while.

Press ctrl+alt+del and click on Task manager. Is the postgres.exe still using the CPU? Is your computer still sounding like it's doing things? If so, let it keep working.

If it isn't, you can kill the PokerTracker process via the Processes tab and try again; it has been known to get stuck on occasion for non-obvious reasons (but then work the second time).

[Suerte]

it has been like 8 hours now. Postgres is running in task manager..... it goes from 0% to like 1% every couple seconds.

Could I cancel the process and resume later? Or if not, possibly delete the dbs and import hands into two new fresh ones in like 3 one million hand batches? Cluster after each one ect.... Would that make it go faster? This is getting ridiculous

[kraada]

Yes, you can cancel the cluster, but you'll need to actually kill PokerTracker from the Task Manager's processes tab as the "cancel" button won't interrupt the cluster.

(If it's updating the cache you should never cancel that will cause database instability.)

After you cancel the cluster it's up to you, you can either try again later (you can't resume, just start over) or you can create a new database, import into new databases and re-cluster at that point. If you're going to create new DBs I'd recommend just importing all of the hands you're going to bring into that database in one go and let it run housekeeping overnight, as it's easier to wake up in the morning with the process done.

The smaller the database, the faster housekeeping will run.

[Suerte]

I did in fact cancel the cluster process. I did do some housekeeping after I canceled it though. I vac/analyzed the db. Before I went to be last night, I started the cluster/update cache process.

I did this by going to "housekeeping" and having only the "cluster" and "update cache" boxes checked. It has been running for 16 hours now and is still not finished.

I am wondering.... which process went first? Was is cluster or cache? It is on updating cache at the moment. I think cluster is done, but I do not know. Please tell me which one goes first, and if the cache should take longer or shorter than the cluster.

This process seems silly if it takes this long every time. Thanks.

[kraada]

Cluster goes first, then Update Cache.

After the cache finishes updating you should vacuum/analyze again. That part is thankfully very quick and it's useful to do that at the end after updating the cache. If you have logging enabled you can monitor the end of the C:\Program Files\PokerTracker 3\PokerTracker.log file to see where the cluster currently is. (I use tail for the monitoring; it's useful

If you don't, after this time finishes, you can enable logging and next time you can see where the cache is when it's working at least.

You don't need to recluster or update the cache until you've imported at least another hundred thousand hands (and you can probably get away with more depending on your computer).

[Suerte]

thank you. Do you know how long cache takes in relation to cluster? Is cache longer or shorter than cluster?

[kraada]

It really depends on your system and how large the database is. As a default I'd assume they take roughly the same amount of time, but after you've done it a few times you'll get a feel for whether your computer is faster at one than the other.