Re: To noatime or not?

I want it to last as long as I keep this computer . . .which given laptop lifetimes and the need to upgrade every 3 years or so means about 2 more years probably. Given that . . .it's a question of whether noatime will prevent SSD write life from being used up. When you average the number of writes over the whole 240 GB on the drive . . .then you end up with really long times to failure. OTOH, if you're constantly rewriting the same blocks due to modifying the directory entries for access time, etc . . .then conceivably it's a much shorter time since the writes aren't averaged over the whole drive but just a small subset of blocks.

In my research .  .I've seen statements that drive writes are randomized and that the directory blocks are constantly moved around on the SSD by the wear leveling routines in the drive firmware . . .but then there were the counter claims that only data blocks were randomized through wear leveling and not the directory blocks. I was unable to find any definitive statement on it.

My gut feeling is that you're right and the life shortening by not using noatime is minimal . . .but was looking for some sort of definitive statement.

On Mar 11, 2011, at 11:43 AM, Daniel Hazelbaker wrote:

> I would say it depends on how long you want your drive to last.  We just purchased 2 128GB Samsung SSD drives for use with moving large files from one city to another on a weekly basis.  After deep research through Samsung's specs and expected time before writing begins to fail (they actually have this documented) we decided we have plenty of time.  I forget exactly how much data Samsung said you could write to it before it started having issues, but we calculated how long it would last for us (which reversing the calculation gives about 13TB of expected write life-time):


-----------------------------------------------
There are only three kinds of stress; your basic nuclear stress, cooking stress, and A$$hole stress. The key to their relationship is Jello.

neil