SQL Server To Do List

This is my SQL Server to do list, I suspect I'll never get to the end, in fact let's start with that thought in mind :

  1. Keep learning!
  2. Tinker with the latest version.
  3. Read the blogs of the DBA greats, current favourites (alphabetical order, no favouritism) are Brent Ozar & associatesDenny CherryGlenn Berry (feel free to suggest more!).
  4. Read the enormous pile of manuals that is ever growing, remember to be thankful for PDFs.
  5. Alter the fill factor review process to also output suggested scripts for new fill factor levels and change the index defrag routine to automatically pick up changed fill factors (or find out a way of updating the fill factor setting without rebuilding the index so it will pick up automatically).
  6. Alter the fill factor review to integrate into one process with the index structural review.
  7. Implement a daily scheduled task to check the ProcessLog for new errors and email them over (so, for example, if the index routine has an issue we can pick that up in the process log rather than ignoring errors as at present).Will be handled by job checking routine.
  8. Alter the Check State routine to log frequency of checking against each instance record to accommodate different server priorities (so live boxes might be every 15 minutes but dev hourly).
  9. Modify the Check State routine to specify which periods checking should be checked (most environments I have worked in operate on an extended business day Monday to Friday so little point flagging server state on a Sunday night when no one is going to respond).
  10. (above item) suggests that I need to change that thought so that I have both a checking period and a reporting period so that servers can be checked 24 hours and create the damning logs to wave a finger at but not fill my inbox with emails I will delete first thing Monday morning.
  11. Alter SQL Agent job deployment code to accept the Operator as a parameter to reduce hard coding (and set that parameter once!).
  12. Make a SQL based deployment record to report servers without latest versions of deployed code in place so upgrade list can be presented.
  13. Change the weekly maintenance task so it only deploys the State Checking step if this is the monitoring server.
  14. Update "history maintenance" routines to ensure the ProcessLog is also pruned.
  15. Change the Check State routine to work in parallel to reduce run times.
  16. Build a set of standard alerts - this could be done in conjunction with a set of baselines & data collection
  17. Alter all the local checking routines to ensure they all keep current / history data and also feed latest results into a server summary table (or a view / stored proc)
  18. Alter index defrag routine to pick up the actual sample size used for the stats rebuilds (and update the update stats stored procs)
  19. Merge my ProcessLog tables into one location and one set of source to ensure consistency.
  20. Change my 1-click deployment to replace hand carved paths with a variable.
  21. Alter the fill factor review process to also compare page space used post frag to page fullness 

No comments:

Post a Comment