How have you improved the SQL Servers at your work

  • Interesting. How are you handling a trend analysis that would go across the year end? Say looking at 6 months from Oct 2001 - Mar 2002?

    K. Brian Kelley

    K. Brian Kelley

  • Currently we keep the most current year in the table. The data is rolling so today the data from Feb 12 of 2001 was over written. We have only had this up for about 8 months, prior to the year end we are going to create another set of tables that holds month end snapshots.

    Currently the interface dispays up to 7 days\ 7 weeks \ 7 months worth of data, so you can look at the scenario that you described above. I would be happy to email you a screen shot if you want to see what it looks like.

  • I have done something similar but on a smaller scale (much smaller) and for a different purpose than trending table sizes. Daily I insert into a table how many rows a table has along with how many rows have a specified column with a specified value. I get an e-mail that shows me the results. I use this to help me see if there is a problem with some of the columns we populate based on values from other columns, such as the current value of stocks an institution holds. I have considered developing something to see the trend on these few tables and columns since I have a year or more of this data.

    Robert Marda

    Robert W. Marda
    Billing and OSS Specialist - SQL Programmer
    MCL Systems

  • I'm working at making small changes, but the biggest so far is reclaiming ~25GB of wasted log space. This saved us from going out and buying more disks we didn't need, which they were about to do.

    That particular server is used for data warehousing and the logs were huge. Basically, they were set to autogrow. So with each monthly load they grew and grew and grew. (For one the database was ~11GB, the log was over 15GB! and this for a data warehouse that is only refreshed once a month) The previous DBA (my boss) had set up the wizard generated job to weekly reclaim unused space. But logs are tricky and the job was ineffective.

    The cool part is that in the process I learned alot about the structure of log files (thank goodness for Inside SQL Server - don't leave home without it).



  • The biggest improvement I made was to get rid of replication!

    When I inherited the servers the set up was 1 server here which had all the data on it, and then servers at each of the manufacturing sites which had their own data on them. There was two way merge replication set up between the servers. Everytime it failed, which it did on a regular basis, I got bleeped by the monitoring system! Not only was that a pain, but when we tried to do a disaster recovery test ( the first one ever, and the servers had been in place for almost 2 years!) we could not get the database to come out of suspect!

    I ended up getting rid of replication and setting it up as a central database server with remote application servers. The next DR we did was plain sailing. As for performance - it got better.

Viewing 5 posts - 16 through 20 (of 20 total)

You must be logged in to reply to this topic. Login to reply