В течении доклада мы с вами рассмотрим ряд принципов и техник, которые позволят вашей базе данных справляться с большей нагрузкой. P.S. Все примеры и демо будут проводиться на базе данных MS SQL Server. Все совпадения с другими базами данными случайны, но вполне вероятны :) так что знания, полученные в ходе доклада, могут вам пригодиться даже если вы работаете с другой базой данных.
6. Questions
• How do you know that your database can’t
handle a load?
• Which load your database can handle?
• Is load predictable or unpredictable?
• Does only database can’t handle it?
• Has anything been done to improve this?
• Do you have any information which prove that
database is not efficient?
7. Time for the Questions
• How we came to the current state?
• Were there any particular actions which increase the
load?
• Does load predictable or sporadic?
• Has database started to operate slowly at one
moment or it was a continuous process?
• Which information do we have?
8. What to do
• Collect Information (current state and
historical)
• Talk to people (well.. we are already doing this )
• Environment information (OS, Database, Hypervisor,
etc)
• Server OS logs
• System resources utilization
• Database queries trace
• Any third-party dashboards and logs
19. Solution Variations
• Code or SQL optimization
• Adding more resources to the server (Scale-Up)
• Patching database server or OS
• Distribute database data among several servers
(Scale-Out) (?)
• Switch to another database server (?)
20. So, here is the end of story
well.. actually not..