All of lore.kernel.org
 help / color / mirror / Atom feed
* rbd speed: write vs read
@ 2011-05-06 14:14 Fyodor Ustinov
  2011-05-06 16:08 ` Sage Weil
  0 siblings, 1 reply; 4+ messages in thread
From: Fyodor Ustinov @ 2011-05-06 14:14 UTC (permalink / raw)
  To: ceph-devel

Hi!

oops. This my misconfiguration.

WBR,
     Fyodor.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: rbd speed: write vs read
  2011-05-06 14:14 rbd speed: write vs read Fyodor Ustinov
@ 2011-05-06 16:08 ` Sage Weil
  2011-05-06 18:01   ` Fyodor Ustinov
  0 siblings, 1 reply; 4+ messages in thread
From: Sage Weil @ 2011-05-06 16:08 UTC (permalink / raw)
  To: Fyodor Ustinov; +Cc: ceph-devel

On Fri, 6 May 2011, Fyodor Ustinov wrote:
> Hi!
> 
> oops. This my misconfiguration.

Out of curiousity, what was the misconfiguration?

Doing a streaming read in 10MB chunks when RBD is striped over 4MB 
objects won't be terribly efficient.  You should get the best sync read 
performance when the object sizes match up.  You'll get better performance 
if whatever is sitting above the block layer dispatches multiple read 
requests in parallel (i.e. something smarter than dd, like a file system).

sage

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: rbd speed: write vs read
  2011-05-06 16:08 ` Sage Weil
@ 2011-05-06 18:01   ` Fyodor Ustinov
  0 siblings, 0 replies; 4+ messages in thread
From: Fyodor Ustinov @ 2011-05-06 18:01 UTC (permalink / raw)
  To: Sage Weil; +Cc: ceph-devel

On 05/06/2011 07:08 PM, Sage Weil wrote:
> On Fri, 6 May 2011, Fyodor Ustinov wrote:
>> Hi!
>>
>> oops. This my misconfiguration.
> Out of curiousity, what was the misconfiguration?
I have 20 OSD server. 10 in one vlan and 10 in other (btw, question: mon 
server can listen on all ip addressessimultaneously?).
Because our NOC, 10 servers have anisotropic configuration - 800Mbps in 
one direction and 150Mbps to another.
> Doing a streaming read in 10MB chunks when RBD is striped over 4MB
> objects won't be terribly efficient.  You should get the best sync read
> performance when the object sizes match up.  You'll get better performance
> if whatever is sitting above the block layer dispatches multiple read
> requests in parallel (i.e. something smarter than dd, like a file system).
Hmm. But it's linear read. I do not know how it works internally, but 
did reading 10MB by 4MB chunk can not be executed in parallel?

WBR,
     Fyodor.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* rbd speed: write vs read
@ 2011-05-06 13:58 Fyodor Ustinov
  0 siblings, 0 replies; 4+ messages in thread
From: Fyodor Ustinov @ 2011-05-06 13:58 UTC (permalink / raw)
  To: ceph-devel

Hi!

root@stb1:~# dd if=/dev/zero of=/dev/rbd0 bs=10240000
^C13217+0 records in
13217+0 records out
135342080000 bytes (135 GB) copied, 1135.75 s, 119 MB/s

root@stb1:~# dd if=/dev/rbd0 of=/dev/null bs=10240000
^C930+0 records in
929+0 records out
9512960000 bytes (9.5 GB) copied, 394.773 s, 24.1 MB/s

It's my misconfiguration?

P.S. 20 OSD servers, 2*1G ether.

WBR,
     Fyodor.

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2011-05-06 18:01 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-05-06 14:14 rbd speed: write vs read Fyodor Ustinov
2011-05-06 16:08 ` Sage Weil
2011-05-06 18:01   ` Fyodor Ustinov
  -- strict thread matches above, loose matches on Subject: below --
2011-05-06 13:58 Fyodor Ustinov

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.