I said I’d never be one of those neglecting bloggers…

It’s been a busy last month…finally (after six months of work) got everything up and running. When I started this blog I told myself, "I’m not going to be one of those people who starts a blog and then neglects it. Not me." Ugh, I’ll just blame it on the holiday breaks.
 
Total damage
Size: 350GB
Growth: 1.25GB/day
Databases: Six
Measure Groups: 38
Dimensions over one million members: Ten
Largest dimension: ten million members
Partitions: appx 1500 and growing daily
 
Based on growth we’re looking at being 750GB by the end of the year. That is, if we just let everything be. However, we have a migration project coming that has the possibility of doubling the warehouse. My guess is that we could be anywhere between one and two TB by the end of the year. Fun!
 
 
 
 
Advertisements
This entry was posted in Uncategorized. Bookmark the permalink.

4 Responses to I said I’d never be one of those neglecting bloggers…

  1. Rahil says:

    We\’ve got a cube that is going on to a size of yours. Did you adjust any advanced AS2005 properties? If so then which ones? We have an 8 processor AMD Opteron with 32GB RAM.
     
    Thanks.

  2. Jesse says:

    We\’re on a four proc, 16GB Opteron as well. It\’s holding together quite well. The most RAM I\’ve seen used is 11GB. CPU will get pegged, but that\’s nothing a few procs won\’t help.
     
    As far as adjusting advanced properties, nothing as of yet. Haven\’t seen any reason to change anything as the performance is great. I suppose once peformance degrades I\’ll go looking for things to change.
     
    The biggest gain we\’ve got is by upgrading disk. Get the fastest disk you can afford. Just a few days ago we were able to upgrade to SAN disk and it cut down processing time by up to 50%. Whereas RAM was the biggest hardware limiting factor for most in AS2000, my guess is that it\’s going to be disk for SSAS2005.

  3. Rahil says:

    Was that while querying? We seem to get poor querying performance because AS only uses around 1Gb of memory and starts paging even though it has about 30 Gb of memory available. I\’m tweaking the no of threads AS uses etc (as recommended by most pundits), however since your cube is of similar size, just wanted to know what you might have done.
     
    Also when querying, do you have the issue where AS starts reading partitions it shouldn\’t have to read from (i.e. if I have  monthly partitions, and I run a query for a particular date, sql profiler shows random partitions being read for that measure group).
    Each partitions has read 300,000,000 rows, so it\’s above any threshold that AS uses.

  4. Jesse says:

    A couple of questions:
     
    How are you partitioning? How do you decide what data goes into what partition?
    How do the users query the data? By dates?
     
    What aggregations have you designed? Have you done any custom or just let the Agg Design Wizard go at them?
     
    Paging is normal. We get paging all the time. We did get some weirdness where queries were going against all partitions, however processing the indexes seemed to get rid of that.
     

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s