Case Study #1:
Client reports that their SQL Server downtime (due to maintenance tasks) takes nearly 5 hours to complete, leaving the database inaccessible. If any errors or problems occur, even more time is lost.
ProboData identified that database re-indexing tasks took the majority of the time (3 hours). Database jobs that come with SQL Server will re-index all the indexes on all the tables….regardless of whether they need re-indexing or not. By doing intelligent re-indexing, often one can find that only a few indexes in the database actually need to be re-indexed. ProboData was able to log fragmentation and page density sizes along with the number of corresponding pages for each index. Smaller tables, even if heavily fragmented perform effectively. However, large tables (over 10,000 pages) that are highly fragmented (20% or higher) should be re-indexed.
Database downtime was reduced from 5 hours to 2 hours. After investigation, it was discovered that only a few indexes needed re-indexing…and not every night. Re-indexing task was moved to weekends and only took 30 minutes. Three hours of unnecessary downtime per night were eliminated.
Case Study #2:
The client’s business was downsizing and there were 30+ SQL Servers spread across the entire company. The client was committing to having all the servers hosted by a facility.
ProboData performed an analysis of the architectural environment and located 51 SQL Servers throughout the company – all of which were SQL Enterprise versions. It was determined all servers were running below 10% CPU utilization. Only two servers in a cluster actually need to run the enterprise version. ProboData consolidated all reporting and legacy databases onto one SQL Server and further consolidated more databases (grouped transactional and finance databases together). Virtualized many consolidated SQL Servers into Virtual VMware clusters, which eliminated the need for SQL Enterprise Edition altogether.
Consolidation and virtualization saved company $100,000 in licensing costs.