Hello,[i](I just want to preface my question by saying that I'm not looking for a definitive answer on this: I just want to gauge peoples thoughts and experiences on tackling locking and blocking with hardware changes. Obviously code changes is the obvious choice, but with 3rd party databases that is hard to manage, especially when the developers are insistent that there is nothing they can do.)[/i]I'm managing a customised database based off of the Microsoft Dynamics NAV CRM solution, and we are experiencing a lot of locking and blocking issues. A quick Google search will throw up plenty of results that locking/blocking is a known issue within NAV. Using the Redgate Monitor tool, we have 100's of locking/blocking alerts a day, and the company want to know what changes we can make at a hardware level to help reduce this: the 3rd party development team who customised NAV have recommended that the database is stored on SSDs. However, I'm skeptical with this because the buffer cache hit ratio is rarely, if ever below 100%, and that latency for the read/write on the data disk is rarely above 25ms. The total memory available to the instance is 196GB, and the database itself is 525GB, so I know that SQL does have to retrieve data on occasion.So essentially my question is:[b]Will putting the data file on SSD actually reduce locking blocking when all the data appears to be in memory anyway?[/b]I'm pretty certain that the server itself is OK, it's the database itself that really needs an overhaul. For what it is worth, MAXDOP is set to 1, the CPU's load (2*X5570) are never greater than 40%, and the CPU Length Queue is rarely above 0,005, and never greater than .025, so definitely not CPU bottleneck.Any thoughts appreciated!
↧