On 07/06/11 02:26, Phil Stracchino wrote:

>> Are there any performance implications for SSD's when they get
>> fragmented?
>
> Not enough so to notice.

I'll add this: the spool area is constantly being cleaned out so 
fragmentation never gets much of a chance to build up.

As Phil says, it's mostly irrelevant on SSDs because they don't have any 
seek time to speak of - however it's a _very_ good idea to make sure you 
lay out the disks so that the structure conforms to the underlaying 
block arrangement (there are a number of online tutorials on this) in 
order to avoid erase/write cycles being split over 2 blocks - that 
causes write amplification but doesn't usually matter for spooling 
purposes (especially if you use SLC, which I strongly recommend for this 
kind of application as it generally has far superior write speeds to MLC)

Having run postgres and SSD spooling for a couple of years now, I'd 
recommend using a separate raid10 SSD for the postgres area. The 
spinning media array I'm using at the moment is quite a bottleneck at 
times and SSD drives are a lot cheaper than they were 2 years ago. :)

If you have money to burn, consider using a PCIe SSD, but SATA SSD is 
fast enough for most purposes.




------------------------------------------------------------------------------
EditLive Enterprise is the world's most technically advanced content
authoring tool. Experience the power of Track Changes, Inline Image
Editing and ensure content is compliant with Accessibility Checking.
http://p.sf.net/sfu/ephox-dev2dev
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users