ProLiant Servers (ML,DL,SL)
cancel
Showing results for 
Search instead for 
Did you mean: 

HP SSA Performance question

 
j1_bailey
Occasional Contributor

HP SSA Performance question

I have a ProLiant DL380 G7 Red Hat Enterprise Linux (RHEL) server with an Smart Array 410i controller.

On Array B (RAID5, 3+1, 450GB drives), I'd like to extend the array onto two 450GB drives.

 

Are there any metrics to how long array extnend would take?  Does it normally take 15minutes per GB or is that a worse-case scenario?

 

Also, when starting an array extension, does it affect the performance of the existing array or the server (cpu, memory, I/O response time?  Does it slow performance down?  Are there any reports of losing data or data being inaccessible on the targetted array? 

 

My customer and I must be well-informed of this before proceeding.

 

Thanks!

 

P.S. This thread has been moved from ProLiant Servers - Netservers to ProLiant Servers (ML,DL,SL). -HP Forum Moderator

2 REPLIES
scharchouf
Trusted Contributor

Re: HP SSA Performance question

Yes it's normal to take 15minutes per GB :)

 

check this document you can find on it all responses for you questions :)

 

HP Smart Array controller technology

I work for HP
A quick resolution to technical issues for your HP Enterprise products is just a click away HP Support Center Knowledge-base
See Self Help Post for more details

waaronb
Respected Contributor

Re: HP SSA Performance question

There will be some performance impact any time you're doing something on the array like extending it, changing stripe sizes, migrating raid levels, etc.

You can configure the priority of those operations in the general config settings for the controller. If you want it to finish quicker, you can set them to high priority but that will slow down normal access. Or if you don't care if it takes even longer, set it to low priority so it won't have as much impact.

I made the mistake of adding some drives to an array and making some RAID changes (from RAID 10 to RAID 50) on some of our development systems a couple weeks ago, and it was having a noticeable impact on the virtual systems hosted on that machine. I changed it to lower priority and that helped, even though it meant I had to wait longer to expand the array once the level migration finished.