From mboxrd@z Thu Jan 1 00:00:00 1970 From: Goswin von Brederlow Subject: Re: -z not on RAID0 / MPIO hurdle Date: Mon, 17 Aug 2009 09:32:56 +0200 Message-ID: <87zl9yvqs7.fsf@frosties.localdomain> References: <66781b10908121144n1f053b57led29391534614b7@mail.gmail.com> <87d4705uxm.fsf@frosties.localdomain> <66781b10908140129t4fc0947ek1e6f211758a2e969@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <66781b10908140129t4fc0947ek1e6f211758a2e969@mail.gmail.com> (mark delfman's message of "Fri, 14 Aug 2009 09:29:39 +0100") Sender: linux-raid-owner@vger.kernel.org To: mark delfman Cc: Goswin von Brederlow , linux-raid@vger.kernel.org List-Id: linux-raid.ids mark delfman writes: > Hi Goswin, > > I am not sure that this would work well for us... we would have you > use LVM to create individual block targerts, then use MD to create > RAIDs and then LVM again to create pools. Last I tried raid on lvm would mess with the idle I/O determination. The raid rebuild would cause I/O on the device mapper and then raid things the user is doing I/O and slows the raid rebuild down to the minimum speed. > We would ideally stay to LVM on all our raids (whether 0 / 6 etc). so > would like to keep it in there... > > Does any know if there is a reason -z doesn=E2=80=99t work with RAID0= (i mean > is it designed to not work because of abc reason, or simply its never > been needed etc)? > > > Thank you for you help > > > Mark D > On Thu, Aug 13, 2009 at 3:06 AM, Goswin von Brederlow wrote: >> mark delfman writes: >> >>> Hi.... I am hopeful that the group could provide me with some advic= e >>> regarding limiting the drive size on a RAID0. =C2=A0I know and use = the =E2=80=93z >>> option for other raid levels, but this seems to have no effect on >>> RAID0. >>> >>> I am sure many will ask =E2=80=98why bother=E2=80=99 or =E2=80=98us= e fdisk=E2=80=99 but the problem is: >>> >>> We need to have some preserved information on each drive =E2=80=93 = can be only >>> a few MB at the end of the drive for example >>> In the past we have simply used Fdisk to partition, save our data o= n >>> Sdx2 and used Mdadm on Sdx1 =C2=A0(which works fine) >>> But we are looking to move to dual channel SAS drives and use DM MP= IO >>> >>> DM MPIO works fine also, but not on the partitions, only the whole >>> drive. =C2=A0So when we MDadm DMx it overwrites the SdX2 partition = etc. >>> >>> The answer was that we used the =E2=80=93z and limited the drive si= ze and >>> simply used DD to write our data to the end LBA=E2=80=99s.... >>> >>> But on RAID0 the =E2=80=93z does not seem to function, so I cant th= ink of a >>> way around this =EF=81=8C =C2=A0(other than changing the LBA range = in the drive and >>> this is restricted I think) >>> >>> >>> Does anyone have any ideas? =C2=A0I appreciate any help >> >> LVM and --setphysicalvolumesize on pvcreate? >> >> MfG >> =C2=A0 =C2=A0 =C2=A0 =C2=A0Goswin >> MfG Goswin -- To unsubscribe from this list: send the line "unsubscribe linux-raid" i= n the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html