Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Gents, We have a database server (Oracle), its raid array is read intensive... Raid Controller type:

04:00.0 RAID bus controller: Hewlett-Packard Company Smart Array G6 controllers (rev 01)
07:00.0 RAID bus controller: Hewlett-Packard Company Smart Array G6 controllers (rev 01)

Here are the values we are seeing via iostat:

cciss/c1d0     4933.00    788848.00      7088.00     788848       7088

File system is ext3... Can/should we expect greater performance from this setup?

Anything specific you would recommend I check/do?


share|improve this question

closed as not a real question by thinice, Lucas Kauffman, Chopper3, womble, ewwhite Aug 8 '12 at 16:15

It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center.If this question can be reworded to fit the rules in the help center, please edit the question.

We're missing some of the details. Which server model is this? I'm assuming the controller is a Smart Array P410i, however, the number type and RAID arrangement of the disks would also be helpful. – ewwhite Aug 1 '12 at 22:03
@ewwhite cciss0: HP Smart Array P410i Controller. Logical drives: 2 Raid 1(1+0) – CMag Aug 1 '12 at 22:10
What are the drives? 10k? 15k? – David Aug 2 '12 at 0:49
mmmm, not sure... lets assume 10K – CMag Aug 2 '12 at 21:54

You don't say what controller you have but most G6 servers come with a P410i, which has a PCIe 2.O x8 connector. So that could, at least theoretically, deliver 4GBps Either way, so if we take typical 4k reads and writes, 100% from cache and without overhead you'd be able to read and write that same block ~1m times per second both ways.

Of course this is all entirely theoretical because you won't want to actually do that but then you've only given us maybe 2% of the information we'd actually need to help you but it does answer your question as written.

share|improve this answer
Now you've told us some (small) information then even if you've got a pair of 146GB 15krpm disks in that box then ignoring controller and disk caching the most you've get our a R1 (it's not a R10 if you only have 2 disks) will be ~2x180 IOPS, so 360 for random read/write. So if you're seeing nearly 5k IOPS you must be getting a hell of a lot of successful cache hits – Chopper3 Aug 1 '12 at 22:54
Typo sorry, 360 for read, 180 for write. – Chopper3 Aug 1 '12 at 23:00
Thanks!!! So how do you make a correlation between 2x180 and 8000 IOPS :) So what I can read from this is... we are maxing the IOPS on this controller... hardware bottleneck is the answer – CMag Aug 1 '12 at 23:54
where can i get the first number you are saying? The 2x180 IOPS? Is it the cat /proc/driver/cciss/cciss1 ? or somewhere else, or just a known hardware spec – CMag Aug 1 '12 at 23:55
That's the max guarunteed non-cached random read/write capability of those disks, I buy them too. – Chopper3 Aug 2 '12 at 6:50

Not the answer you're looking for? Browse other questions tagged or ask your own question.