http://blog.nigelpoulton.com/self-tuning-storage/#utm_source=feed&utm_medium=feed&utm_campaign=feed with nigel poulton Tue, 20 Dec 2011 19:37:54 +0000 http://wordpress.org/?v=2.8.6 hourly 1 http://blog.nigelpoulton.com/self-tuning-storage/comment-page-1/#comment-181 Chris M Evans Mon, 14 May 2007 18:50:27 +0000 http://blogs.rupturedmonkey.com/?p=85#comment-181 Right on the money, Nigel.  In fact I'd go as far as to suggest that EMC, HDS and IBM (although I have little knowledge of their products) are amending their arrays to fix inherent problems with their design - like having to have disks on FC-AL loops which then require moving data between physical disks to manage performance.  Rather than developing "self-optimising" arrays, I think they're all putting sticking plasters over the problem. Right on the money, Nigel.  In fact I’d go as far as to suggest that EMC, HDS and IBM (although I have little knowledge of their products) are amending their arrays to fix inherent problems with their design – like having to have disks on FC-AL loops which then require moving data between physical disks to manage performance.  Rather than developing "self-optimising" arrays, I think they’re all putting sticking plasters over the problem.

]]> http://blog.nigelpoulton.com/self-tuning-storage/comment-page-1/#comment-180 nigel Mon, 14 May 2007 11:47:13 +0000 http://blogs.rupturedmonkey.com/?p=85#comment-180 Hi Ewan, I hear what your saying.  sounds similar to the conversation around one of my previous posts - <a href="http://blogs.rupturedmonkey.com/?p=56" rel="nofollow">http://blogs.rupturedmonkey.com/?p=56</a> Thanks for the feedback Nigel Hi Ewan,
I hear what your saying.  sounds similar to the conversation around one of my previous posts – http://blogs.rupturedmonkey.com/?p=56
Thanks for the feedback
Nigel

]]> http://blog.nigelpoulton.com/self-tuning-storage/comment-page-1/#comment-179 Ewan Mon, 14 May 2007 11:39:05 +0000 http://blogs.rupturedmonkey.com/?p=85#comment-179 I don't really see why some storage vendors are putting so much complexity into tuning storage systems, personally I don't have time to work out how few ms per write is required for each system connected to our SAN, or how much cache should be allocated per drive.I just want to be able to say "All drives attached to production environment A, high priority", and "All drives attached to test environments B, C, D, low priority", then let the array sort itself out.Even a simple scale of priority from 1-10 would probably be too much, high medium and low would probably cover 90% of peoples requirements and need so much less effort from storage administrators everywhere. I don’t really see why some storage vendors are putting so much complexity into tuning storage systems, personally I don’t have time to work out how few ms per write is required for each system connected to our SAN, or how much cache should be allocated per drive.I just want to be able to say "All drives attached to production environment A, high priority", and "All drives attached to test environments B, C, D, low priority", then let the array sort itself out.Even a simple scale of priority from 1-10 would probably be too much, high medium and low would probably cover 90% of peoples requirements and need so much less effort from storage administrators everywhere.

]]>