At what point do you start getting neverous that your running out of disk space? I realize the answer to this questions is it depends on your rate of growth and whether its linear or exponential, but as a rule of thumb what percentage of available disk space makes you wonder if you're going to have a job tomorrow? I work for a company that seems to operate perpetually between 10% and 15% available disk space and when I say that I don't mean to imply that disk usage is steady. We are adding capacity, but it is being used up nearly as fast as we're adding it. I have always used about 25% to 30% as a general rule of thumb to indicate that I need to expand capacity and I like to see that number at around 75% or better after expansion. I realize times are tight and companies don't like to spend money if they don't have to but this seems to be a risky practice to me. I'm curious to hear what everyone else uses as rule of thumb/best practice/corporate policy and whether their rules match the reality of their work environments. Thanks for your input, best regards, steve h
The task is to make sure you always have enough space. Determining what that minimum is means forecasting your rate of growth and knowing how fast you can add capacity. A small company that can run to Fry's and double their capacity in a couple of hours can run much closer to the razor's edge than a company with ordering or installation procedures that would mean upgrading would take weeks. If you really want a rule of thumb without estimating that, I would say anything less than 10% should make you take action. But that is just a very rough rule of thumb, something to keep in mind the first time you see a new server. A highly active server would likely want a much, much larger margin, while a nearly static server that was virtually read only could be safe with sufficiently less.
Its better to configure some alerts once your disk is about 75%-85% full, so you will not be in trouble. Always make sure that you have sufficient space on the disk for your backup files also. I recommend you to reserve some space for emergency Like, if your DB needs an Index or an indexed view to improve the performance then you should able to create those with no worries.
This is a tricky one if you are acquiring new systems their size will be a shock and their trends unknown. At some point you need to take the system with new hardware and share the load (clustering?). Regarding @Cyborgs answer - setting an alert at 75%-85% will simply trigger every time it checks if you are running at 10%-15% free so you will need to set it differently or get ahead of the game with your adding capacity program. Have you considered any of the database virtualisation solutions? They run from a compressed backup file, sitting between the file and the SQL Server. SQL thinks it has normal mdf and ldf but the app is translating changes back into the backup file. There have been some positive reviews [
http://www.simple-talk.com/sql/sql-tools/brads-sure-guide-to-sql-virtual-restore-/] but I havent tried it personally. There are other products that do this but I dont have their names/providers details at the moment. Disclosure: I am a Friend of RedGate ([
www.red-gate.com]) and have a certain bias to their products. I have however not used the product mentioned above so cannot recommend it as a solution without serious testing against your requirements, it just seems that is would save you some space without changing your infrastructure. :