To me, the biggest blunder is knowing that a problem exists and either ignoring it or procrastinating on the implementing the resolution. This tip focuses on Common DBA project Challenge that could have been prevented.
The reality is that nothing is perfect and as technical professionals, we need to build a realistic solution with the time and budget available, then communicate any potential issues to the business so they are aware of them.
To deliver better application performance, DBAs should consider the following tips:
- Be proactive and align behind end-user experience as a shared objective across the entire IT organization by looking at application performance and the impact that the database has on it continuously, not only when it becomes a major problem.
- Measure performance based not on an infrastructure resources perspective, but on end-user wait times. The wait-time analysis gives DBAs a view into what end-users are waiting for and what the database is waiting for, providing clear visibility into bottlenecks.
- Implement monitoring tools that provide visibility across the entire application stack, including all the infrastructure that supports the database – virtualization layers, database servers, hosts, storage systems, networks, etc.
- Establish historic baselines of application and database performance that look at how applications performed at the same time on the same day last week, and the week before that, to detect any anomalies before they become larger problems.
- Have a common set of goals, metrics, and SLAs across all databases, ideally based on application response times, not only uptime.
- Use tools that provide a single dashboard of performance and the ability to drill down across database technologies and deployment methods, including cloud.
- Document a consistent set of processes for ensuring integrity and security: backup and restore processes, encryption at rest and in transit, detection of anomalies and potential security events in logs, to name a few.
- Establish a strategy, roadmap, and guidelines for moving to the cloud (or not) and for reducing workload costs by moving databases to lower-license-cost versions or open-source alternatives.
- Make sure team members can escape firefighting mode and spend enough time proactively optimizing the performance of the databases and taking care of important maintenance tasks, which can result in significant cost savings and prevent problems in the future.
Database project Operations issue:
- All senior team members on vacation – When you have a major deployment make sure to have your key staff members on site and available to meet the project needs. Do not fool yourself that a junior team member will be able to just push a button and deploy a solution, especially when a minimal amount of testing is conducted. When a problem arises it is the senior team member’s knowledge and expertise that is needed to quickly resolve issues. If all of those people are out on the beach and will be back next week, it makes sense to wait a week for the deployment to have your team onsite and available to address any unexpected issues.
- Putting all of your eggs in 1 basket – When you work through an enterprise upgrade whether it is an application or hardware firmware, do not upgrade all of the systems (including the DR site) at once. Take a step back and be sure to have some systems that are out of sync for a short period of time to migrate to a known stable platform in case an unexpected issue arises.
- Not validating backups on a daily basis – If a serious issue occurs, make sure you have a solid last line of defense. That is a consistent and reliable set of backups on a daily basis. In addition, make sure your backup plan includes retiring tapes on a weekly, monthly or quarterly basis to be able to rollback to some point in time rather than going out of business. Also, check-in with the business to ensure backups are not needed for legal or regulatory needs.
- Not changing passwords – As an administrator, you have the keys to the kingdom and need to recognize the responsibility that you have. As such, make sure your passwords are complex, change them frequently and do not share your passwords.
- Password expiration – This is almost the opposite of the previous bullet. With SQL Server 2005 password policies can be set up for standard logins so the passwords expire and accounts get locked out. When this happens your applications will not be accessible if one of these accounts are in use. As such, setting password expiration is a good idea, just be sure to change the password and coordinate the change with your team.
- Letting the primary filegroup fill up – With the rate of data growth, be sure to either cap your database size or monitor the size on a daily, weekly or monthly basis or permit your databases to automatically grow. In either circumstance, be sure to watch your disk space so that you do not fill up your disk and then have 2 problems (full file group and full disk drive).
- Hot data centers – High temperatures mean failure for servers. The failure could be a controller card or a disk drive, but one sustained spike in the room temperature could be a critical problem that is not fully realized for a three to six month time period. Make sure your temperature is properly regulated, has a backup conditioning system and can notify your team before an issue arises.
Whatever your company’s needs may be, we’ve got you covered. Our team of domestic primary DBAs has a proven depth and range of experience, unsurpassed in the industry. With our highly tailored set of solutions, we’ll make sure you’re only paying for what you need.
For more information, visit our website at https://www.smarterp.com/services/database-and-middleware-management .