63.2 GB Postgresql DB clustering / housekeeping for 40 hours

[Steve Ford]

I finally finished loading all of my hand histories into my db and my post gres folder is 63.2 GB (with just the one poker tracker 3 db in it). I started it doing all of the housekeeping and it clustered for about 30 hours and it is currently updating cache about about to hit the 40 hour mark for total time. My question is I will be loading about 10k hands a week into this thing and I can't be down for 40 hours a week for maint. Is it only taking this long because it is the first time since i loaded 8 mil hands into it? How long will this take approx if it is 70 gig and is updating about 10k hands since the last time I did it? I am just trying to get an idea because this may be a waste of time for something I will end up not using after all and I can go ahead and stop now. Also it took forever to give me a view of all my players etc. Is there a way to speed that up?

[kraada]

I would not recommend clustering and/or updating the cache every week if you're only importing 10k hands.

At that pace you can run housekeeping about once a month without any real problems.

You should vacuum and analyze after the cache update finishes (JUST vacuum and analyze; those two should only take a minute or two each).

Unfortunately the clustering and cache updating take about the same amount of time each time you do it.

If you have more free space on your hard drive the cluster might go a bit faster.

[dsavereide]

I was wondering about this...

Would it make sense to start a new db everytime you reached a certain size in your current db? If I wanted to keep clustering and cache updating down to an over night task ... anyone have an idea how many hands that would be?

[kraada]

How many hands works well for your databases is going to depend on your personal usage and your computer's speed and power.

Personally, I create a new database for observed hands every 2 months.