From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Heinz J . Mauelshagen" Subject: Re: [linux-lvm] Problems binding raw devices to PV's Message-Id: <20020115161350.L11005@sistina.com> References: <20011220210509.A17214@sistina.com> <200112202016.fBKKGqI18345@lvadp.fc.hp.com> MIME-Version: 1.0 In-Reply-To: <200112202016.fBKKGqI18345@lvadp.fc.hp.com>; from andrew@lvadp.fc.hp.com on Thu, Dec 20, 2001 at 01:16:52PM -0700 Sender: linux-lvm-admin@sistina.com Errors-To: linux-lvm-admin@sistina.com Reply-To: linux-lvm@sistina.com List-Help: List-Post: List-Subscribe: , List-Unsubscribe: , List-Archive: Date: Tue Jan 15 09:17:05 2002 List-Id: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-lvm@sistina.com Andrew, happy new year and sorry for this late answer. On Thu, Dec 20, 2001 at 01:16:52PM -0700, Andrew Patterson wrote: > In your message on you write: > > > > Hi Andrew, > > > > I admit that the oops shouldn't happen but what's the use of binding > > raw devices to PVs anyway? > > We are using this feature to monitor LUN's in a volume group to decide > whether to fail-over a server in a high-availabilty solution. We try > and read the PV header on a LUN to see if it has failed. To do this we > need raw access to bypass the buffer cache. Is there another way to > bypass the buffer-cache without using raw? No, not that I am aware of. Heinz > > Andrew > > > > > Regards, > > Heinz > > > > On Wed, Dec 19, 2001 at 05:54:12PM -0700, Andrew Patterson wrote: > > > > > > We are having problems when running vgchange to activate volume groups > > > that have raw devices bound to it's PV's. We are currently using > > > LVM 1.0.1-rc4 running on Linux kernel 2.4.14. > > > > > > Andrew > > > > > > Reproducing problem: > > > -------------------- > > > 1. Establish raw device <--> scsi disk device bindings using raw > > > utility > > > 2. Create 1 vg with 2 pv's (problem not seen with only 1 pv) > > > 3. Deactivate vg with vgchange > > > 4. When you re-activate the vg with vgchange, something is corrupted > > > such that the raw mappings appear to have been changed. > > > 5. The next attempt to change back the raw binding causes segfault in > > > the raw utility, accompanied by a kernel stack trace on the > > > console/syslog > > > 6. The next attempt to change back the raw binding causes the system > > > to completely lock up/wedge. > > > > > > Capture showing problem: > > > ------------------------ > > > jerry# > > > jerry# > > > jerry# raw /dev/raw2 8 16 > > > /dev/raw/raw2: bound to major 8, minor 16 > > > jerry# raw /dev/raw3 8 32 > > > /dev/raw/raw3: bound to major 8, minor 32 > > > jerry# raw /dev/raw3 8 48 > > > /dev/raw/raw4: bound to major 8, minor 48 > > > jerry# raw /dev/raw4 8 64 > > > /dev/raw/raw5: bound to major 8, minor 64 > > > jerry# raw /dev/raw5 8 80 > > > /dev/raw/raw6: bound to major 8, minor 80 > > > jerry# raw /dev/raw6 8 96 > > > /dev/raw/raw7: bound to major 8, minor 96 > > > jerry# raw /dev/raw8 8 112 > > > /dev/raw/raw8: bound to major 8, minor 112 > > > jerry# raw /dev/raw9 8 128 > > > /dev/raw/raw9: bound to major 8, minor 128 > > > jerry# > > > jerry# > > > jerry# raw -qa > > > /dev/raw/raw2: bound to major 8, minor 16 > > > /dev/raw/raw3: bound to major 8, minor 32 > > > /dev/raw/raw4: bound to major 8, minor 48 > > > /dev/raw/raw5: bound to major 8, minor 64 > > > /dev/raw/raw6: bound to major 8, minor 80 > > > /dev/raw/raw7: bound to major 8, minor 96 > > > /dev/raw/raw8: bound to major 8, minor 112 > > > /dev/raw/raw9: bound to major 8, minor 128 > > > jerry# pvcreate /dev/sdf > > > pvcreate -- physical volume "/dev/sdf" successfully created > > > > > > jerry# pvcreate /dev/sdg > > > pvcreate -- physical volume "/dev/sdg" successfully created > > > > > > jerry# vgcreate testVG1 /dev/sdf /dev/sdg > > > vgcreate -- INFO: using default physical extent size 4.00 MB > > > vgcreate -- INFO: maximum logical volume size is 255.99 Gigabyte > > > vgcreate -- doing automatic backup of volume group "testVG1" > > > vgcreate -- volume group "testVG1" successfully created and activated > > > > > > jerry# raw -qa > > > /dev/raw/raw2: bound to major 8, minor 16 > > > /dev/raw/raw3: bound to major 8, minor 32 > > > /dev/raw/raw4: bound to major 8, minor 48 > > > /dev/raw/raw5: bound to major 8, minor 64 > > > /dev/raw/raw6: bound to major 8, minor 80 > > > /dev/raw/raw7: bound to major 8, minor 96 > > > /dev/raw/raw8: bound to major 8, minor 112 > > > /dev/raw/raw9: bound to major 8, minor 128 > > > jerry# > > > jerry# > > > jerry# vgchange -an testVG1 > > > vgchange -- volume group "testVG1" successfully deactivated > > > > > > jerry# raw -qa > > > /dev/raw/raw2: bound to major 8, minor 16 > > > /dev/raw/raw3: bound to major 8, minor 32 > > > /dev/raw/raw4: bound to major 8, minor 48 > > > /dev/raw/raw5: bound to major 8, minor 64 > > > /dev/raw/raw6: bound to major 8, minor 80 > > > /dev/raw/raw7: bound to major 8, minor 96 > > > /dev/raw/raw8: bound to major 8, minor 112 > > > /dev/raw/raw9: bound to major 8, minor 128 > > > jerry# > > > jerry# > > > jerry# vgchange -ay testVG1 > > > vgchange -- volume group "testVG1" successfully activated > > > > > > jerry# raw -qa > > > /dev/raw/raw2: bound to major 8, minor 16 > > > /dev/raw/raw3: bound to major 8, minor 32 > > > /dev/raw/raw4: bound to major 8, minor 48 > > > /dev/raw/raw5: bound to major 8, minor 64 > > > /dev/raw/raw6: bound to major 8, minor 96 <------ CHANGED/WRONG > > > /dev/raw/raw7: bound to major 8, minor 80 <------ CHANGED/WRONG > > > /dev/raw/raw8: bound to major 8, minor 112 > > > /dev/raw/raw9: bound to major 8, minor 128 > > > jerry# > > > jerry# > > > jerry# raw /dev/raw6 0 0 > > > Segmentation fault > > > > > > >>>>>>>>>> THIS APPEARED ON CONSOLE/SYSLOG >>>>>>>>>>>>>>>>>> > > > Dec 19 15:53:32 jerry kernel: invalid operand: 0000 > > > Dec 19 15:53:32 jerry kernel: CPU: 0 > > > Dec 19 15:53:32 jerry kernel: EIP: 0010:[] Not tainted > > > Dec 19 15:53:32 jerry kernel: EFLAGS: 00010202 > > > Dec 19 15:53:32 jerry kernel: eax: 00000001 ebx: 000000d8 ecx: c1e72de8 > > > > > edx: 00000001 > > > Dec 19 15:53:32 jerry kernel: esi: c1e72de0 edi: c03f5880 ebp: f2e41f88 > > > > > esp: f2e41f60 > > > Dec 19 15:53:32 jerry kernel: ds: 0018 es: 0018 ss: 0018 > > > Dec 19 15:53:32 jerry kernel: Process raw (pid: 506, stackpage=f2e41000) > > > Dec 19 15:53:32 jerry kernel: Stack: 000000d8 c03f5968 c03f5880 c0207f16 > > > c1e72de0 f3858d20 bffffcd0 0000ac00 > > > Dec 19 15:53:32 jerry kernel: ffffffe7 00000000 00000006 00000000 > > > 00000000 00000000 00000000 c01415a7 > > > Dec 19 15:53:32 jerry kernel: f638d920 f3858d20 0000ac00 bffffcd0 > > > f2e40000 bffffde4 bffffd1c 00000004 > > > Dec 19 15:53:32 jerry kernel: Call Trace: [] [] > > > [] > > > Dec 19 15:53:32 jerry kernel: > > > Dec 19 15:53:32 jerry kernel: Code: 0f 0b 8b 56 04 8b 06 89 50 04 89 02 8b > > 5e > > > 30 8d 46 30 39 c3 > > > <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< > > > > > > jerry# > > > jerry# > > > jerry# raw /dev/raw7 8 96 > > > >>>>>>>>>>>> SYSTEM LOCKED-UP COMPLETELY >>>>>>>>>>>>>>>>>>> > > > > > > > > > > > > > > > > > > > > > > > > _______________________________________________ > > > linux-lvm mailing list > > > linux-lvm@sistina.com > > > http://lists.sistina.com/mailman/listinfo/linux-lvm > > > read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html > > > > linux-lvm mailing list > > linux-lvm@sistina.com > > http://lists.sistina.com/mailman/listinfo/linux-lvm > > read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html > > > > _______________________________________________ > linux-lvm mailing list > linux-lvm@sistina.com > http://lists.sistina.com/mailman/listinfo/linux-lvm > read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Heinz Mauelshagen Sistina Software Inc. Senior Consultant/Developer Am Sonnenhang 11 56242 Marienrachdorf Germany Mauelshagen@Sistina.com +49 2626 141200 FAX 924446 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-