Activity › Forums › Storage & Archiving › RAID 50 loses half the expected speed?
-
RAID 50 loses half the expected speed?
Chris Murphy replied 12 years, 3 months ago 5 Members · 16 Replies
-
Alex Gerulaitis
January 22, 2014 at 10:48 pm[Bob Zelin] “if you merge both of them as a RAID 50 (or 60), shouldn’t the speed INCREASE when it appears as a single volume ?”
The easiest tests would be:
– on a different host system to see if the problem is machine-specific
– to borrow an R680 or 1882x and perform similar testsUntil then, it’s shooting in the dark.
P.S. Understanding the problem was never the issue, getting the relevant information was.
We did find out that 800MB/s is R380’s performance ceiling, only to later find out there was a test done on two of them, with similar results – although we do not know if both cards were working in PCIe 8x mode.
We still don’t know how that RAID0 test was done – a single stripe set all done in R380, two stripe sets then soft-striped in the OS (with one or two R380s). We still don’t know the specs of the host machine. We don’t know what other tests the OP did.
I’ve asked if the OP contacted ATTO about it, and heard no response.
We don’t know if the system hits memory or CPU utilization ceilings during speed testing possibly pointing to configuration or performance problems.
The root cause of the slowdown is likely the OS – but we won’t know unless we we have all the relevant info about the system, and until proper tests are done.
-
Bob Zelin
January 24, 2014 at 1:00 pmAn R680 or an Areca 1882x test is pointless because –
1) these are 3g chassis with 3g drives
and
2) the client does not want to spend any money.So the only question here is as follows –
no matter what host card you have, no matter what chassis you have, no matter what drives you have –
if they are all the same generation (3g, 6g, whatever) –if you use TWO RAID cards on the same host computer (Mac Pro) connected to TWO RAID arrays (RAID0, RAID5, whatever), and stripe the two together (RAID 10, 50, 60) – SHOULD THE SPEED INCREASE ?
Yes, or no ?And for the record, if you go with a single new ATTO R680 or Areca 1882x (with of course 6g chassis, and 6g drives), then yes – of course, the performance will be dramatically greater than any outcome of the R380 tests. Only one problem here – the client does NOT want to spend one penny on anything new. He just wants more performance with the equipment he currently owns.
Bob Zelin
Rescue 1, Inc.
maxavid@cfl.rr.com -
Alex Gerulaitis
January 24, 2014 at 6:53 pmIndividual drives’ link speeds (3G or 6G) don’t matter. Drives don’t saturate these link speeds.
Neither does the backplane: a 4-lane 3G SAS connection means 12Gbs in full duplex, i.e. roughly 1.2GB/s line speed per connection, and he has two of them: 2.4GB/s.
Testing this setup with a different (faster) controller will determine whether the bottleneck is in the controller or the host system.
-
Chris Murphy
January 28, 2014 at 6:41 amI agree with the earlier questions and I’ll add some of my own:
What computer make/model/RAM?
What benchmarking tool?
Rerun the test on the raid50 volume, while running the Terminal command: ‘top -s5 -n10 -o cpu’ it will take 5 seconds for this to accumulate data, give it 10-15 seconds and take a screen shot (command-shift-4 then spacebar, hover-highlight the terminal window, mouse click)
Rerun the test on a single raid5 set (i.e. break the raid50, run the test on one of the raid5 arrays without software raid), while running the same Terminal command above and screen shot after 10-15 seconds for settling.
In Disk Utility when the raid0 set was created, what was the chunk size? I’m pretty sure the default is 32KB?
In Disk utility created the raid0 set with two hw raid5s, make sure the chunk size is a lot bigger or a lot smaller than whatever you used before. Dollars to donuts it was at the default of 32KB which is normally fine, but 32KB translates into a pile of IOs round robin between those two raid5s that’s totally unnecessary and might just be amping the kernel doing a lot of work it doesn’t need to do. So I’d try something almost obscene like 1MB if the chunk size goes that high and then redo your tests. I’m assuming your average file size is well above 1MB anyway for this raid50?
-
Chris Murphy
January 28, 2014 at 6:44 amOh and for that matter I’d like to know what the chunk size is for the two R380 raid5 sets too.
Reply to this Discussion! Login or Sign Up