[DRBD-user] DRBD and Direct Block IO
Ross S. W. Walker
rwalker at medallion.com
Fri Jan 12 02:20:47 CET 2007
> -----Original Message-----
> From: drbd-user-bounces at lists.linbit.com
> [mailto:drbd-user-bounces at lists.linbit.com] On Behalf Of Ross
> S. W. Walker
> Sent: Thursday, January 11, 2007 7:48 PM
> To: drbd-user at lists.linbit.com
> Subject: [DRBD-user] DRBD and Direct Block IO
> I am having some serious performance issues with a kernel module for
> iSCSI Enterprise Target and drbd 0.7.21 on CentOS 4.4.
> The kernel module (source code listed below) basically takes a data
> request issued over iSCSI translates it to a bio request that is then
> carried out synchronously to the device below.
> When run against an MD/LVM/SD device I don't have any problems, I get
> performance to be expected, but when run against a drbd
> 0.7.21 device it
> chokes down hard. For example when doing seq 64K block write direct to
> device I can get 112 MB/s sustained, when I have drbd in the
> middle that
> throughput drops to 10 MB/s.
> Can anybody help explain this or point out a serious flaw in the code
> below that would cause this. I would rather solve the problem (if it
> can) then try to run version 8 beta in production (if it will
> even solve
> my problem).
<snip code sample>
I forgot to mention probably the most important piece of information,
this drbd replica is currently running Standalone with no peer, so there
is actually no replication happening at this time.
This e-mail, and any attachments thereto, is intended only for use by
the addressee(s) named herein and may contain legally privileged
and/or confidential information. If you are not the intended recipient
of this e-mail, you are hereby notified that any dissemination,
distribution or copying of this e-mail, and any attachments thereto,
is strictly prohibited. If you have received this e-mail in error,
please immediately notify the sender and permanently delete the
original and any copy or printout thereof.
More information about the drbd-user