Want to protect your cyber security and still get fast solutions? Ask a secure question today.Go Premium

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 37
  • Last Modified:

I have a test the load and stress in SAN environment using both FC and iSCSI

I am working a SAN testing to test the load and stress usuing both FC and iSCSI  on the Axiom 600 hardware
In order to support Windows 03 & 08 , Linux , Vmware and XenApps (host )
HBA Type both Fiber and ethernet  
0
pakuma1154
Asked:
pakuma1154
3 Solutions
 
kevinhsiehCommented:
The industry standard package is Iometer. http://www.iometer.org/ . You can test reads and writes with different ratios and IO sizes.
0
 
driskolltCommented:
In addition to what kevinsieh said...

Before you test - Know your workload.  Is it mostly small/random, or large/sequential I/O.

Most production systems are going to be random.  It also helps to know your average read/write size of the applications.  For the most part 4k-8k read/writes is a pretty good bet.  Data Warehouse applications will probably be higher.

IOPS is the most important factor when measuring any type of disk subsystem performance.  Response time is also incredibly important, but high response times are usually a product of either host issues (HBA Queue Depth incorrectly set) , network issues, or RAID config (not right RAID type, not enough disks).

Measuring by MB/s is not a good way of measuring - good MB/s from testing is always easily attainable by increasing the I/O size and doing sequential I/O.  MB/s is a fool's way of measuring disk performance.  MB/s can also be easily calculated by multiplying IOPS * I/O Size anyway.
0
 
DavidPresidentCommented:
Iometer is a decent throughput and I/O generator .. but by no means is it a thorough testing mechanism.  What makes or breaks a SAN is how it performs under stress.   Generate some LIPs, bus resets, inject media ECC errors, change link speeds.

DVT testing is vital..  (Data Validation Testing).  So what if you can pump data, but is data getting corrupted? Are frames getting lost when things are in stress?  If you are using redundant paths, is load balancing and/or path failover and path failback working?

You need to capture log files and see what happens with recovered & unrecovered errors, and look at loop A/B errors, warnings, and such.

You can get some of this information from a decent SAN switch, or a great deal more with some software and an analyzer, but it really depends on your budget and what your goals are to provide some help on a good bang-for-the-buck solution.
0
 
EE_AutoDeleterCommented:
I've requested that this question be deleted for the following reason:

                           
No comment has been added to this question in more than 21 days, so it is now classified as abandoned and is now flagged for deletion.


If there is a valid solution, please OBJECT and indicate the comments that are, or would otherwise lead to, a solution.


Use the specific format https:#axxxxxxxx for comment ID(s).


Also, please don't object simply because the author did not respond to your comment. While we understand this is frustrating, unfortunately we cannot force the author to return to the question. Unless you feel you have presented a valid, verifiable solution we'll simply delete the question.


Experts-Exchange Auto Deleter
0
 
DavidPresidentCommented:
All responses were valid as this was just a question on what software to use. Split points evenly between all experts.
0

Featured Post

VIDEO: THE CONCERTO CLOUD FOR HEALTHCARE

Modern healthcare requires a modern cloud. View this brief video to understand how the Concerto Cloud for Healthcare can help your organization.

Tackle projects and never again get stuck behind a technical roadblock.
Join Now