From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mx1.redhat.com (ext-mx14.extmail.prod.ext.phx2.redhat.com [10.5.110.19]) by int-mx10.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP id p5TJlM1U027278 for ; Wed, 29 Jun 2011 15:47:22 -0400 Received: from ndmsnpf02.ndc.nasa.gov (ndmsnpf02.ndc.nasa.gov [198.117.0.122]) by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id p5TJlG7Z019249 for ; Wed, 29 Jun 2011 15:47:16 -0400 Received: from ndjsppt05.ndc.nasa.gov (ndjsppt05.ndc.nasa.gov [198.117.1.104]) by ndmsnpf02.ndc.nasa.gov (Postfix) with ESMTP id 93E18148116 for ; Wed, 29 Jun 2011 14:47:15 -0500 (CDT) Received: from ndjshub06.ndc.nasa.gov (ndjshub06.ndc.nasa.gov [198.117.4.165]) by ndjsppt05.ndc.nasa.gov (8.14.4/8.14.4) with ESMTP id p5TJlFom020928 for ; Wed, 29 Jun 2011 14:47:15 -0500 From: "Stamper, Brian P. (ARC-D)[Stinger Ghaffarian Technologies Inc. (SGT Inc.)]" Date: Wed, 29 Jun 2011 14:47:13 -0500 Message-ID: In-Reply-To: Content-Language: en Content-Type: multipart/alternative; boundary="_000_CA30CF5111CACbrianpstampernasagov_" MIME-Version: 1.0 Subject: Re: [linux-lvm] Fedora 15 lvm root device not found Reply-To: LVM general discussion and development List-Id: LVM general discussion and development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , List-Id: To: LVM general discussion and development --_000_CA30CF5111CACbrianpstampernasagov_ Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable I was able to boot this system by taking an initramfs from another Fedora 1= 5 test system and putting it in /boot and referencing it in grub.conf. How= ever, I need to find out why an initramfs I build on the system is unable t= o boot. When I'm in dracut, I was able to do some checking. Using the lvm commands= , I was able to activate and deactivate the vg, display the lvs, but not us= e them. I would get "device not found", and in /dev/mapper, the lvm device= s were not there. There were other lvm devices there, but not from my prim= ary system lvm. The only thing that stands out about it is that it is ext4= , not ext3. Comparing the decompiled initrds, there's no modules that stand out. Sas.o= ut is a listing of the WORKING initrd, tfs.out is the NOT WORKING initrd. [root@testfs ramdisks]# diff sas.out tfs.out 718d717 < ./lib/firmware/ql2500_fw.bin 1248,1249d1246 < ./etc/ld.so.conf.d/openmotif-64.conf < ./etc/ld.so.conf.d/kernel-2.6.38.7-30.fc15.x86_64.conf 1252c1249 < ./etc/ld.so.conf.d/kernel-2.6.34.8-68.fc13.x86_64.conf --- > ./etc/ld.so.conf.d/qt-x86_64.conf 1262d1258 < ./etc/modprobe.d/openfwwf.conf 1278a1275 > ./etc/udev/rules.d/40-multipath.rules 1317a1315 > ./lib64/libcryptsetup.so.1.2.0 1322a1321 > ./lib64/libcap.so 1407d1405 < ./lib64/libcryptsetup.so.1.1.0 1419a1418 > ./lib64/libnss_files-2.12.2.so 1480a1480 > ./usr/bin/checkisomd5 1491d1490 < ./usr/lib64/libnss_files.so 1517a1517 > ./usr/share/plymouth/themes/default.plymouth Nothing there looks relevant. Libcryptsetup needs updating on the test sys= tem I pulled the initrd from. The q-logic firmware is irrelevant. There a= re a couple of older kernels on the test system. Libcap.so is a dependency= of legato backup software. Libnss_files should have nothing to do with ac= cessing an LVM. So I'm stuck. My system is up, but only if I build an initrd on another sy= stem and copy it over, meaning next time I update the kernel it will not bo= ot. Nothing's jumping out at me as obviously missing modules from the brok= en initrd. Where else can I look? -Brian On 6/29/11 1:47 AM, "Brian Stamper" wrote: I just upgraded a Fedora system from 13 to 15, and now I can't get it to bo= ot. I'm getting dumped to a dracut shell and getting the error dm-linear d= evice lookup failed. In the dracut output with debug on, I see dracut scan= my drives, and I see it find my 3 LVs, but it then tries to do something w= ith them, I get the error, and it then fails out with root device not found= and drops me into dracut. Here's the odd thing, booting off the rescue dvd my volumes are fine. The = LVM can be mounted no problem. This makes me think that there's something missing from the kernel or initr= d. I can't really figure out what though. I've gone through the output of= lsmod when running from the rescue dvd. I've even build a few ramdisks in= cluding all the modules that I thought might be related (dm, lvm, mdraid, d= mraid, multipath). I tried building them with mkinitrd -with=3Dmodule, and= I tried again with dracut -add=3Dmodules. No change in results. I'm not sure where to go from here. I feel like the system can be rescued,= but I don't know how to work backward through dracut to find out why it's = unable to find my root volume. Any help would be appreciated. Cross-posti= ng to dm-devel. Thanks, -Brian --_000_CA30CF5111CACbrianpstampernasagov_ Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Re: [linux-lvm] Fedora 15 lvm root device not found
I was able to boot this system by taking an initramfs from another Fedora 1= 5 test system and putting it in /boot and referencing it in grub.conf. &nbs= p;However, I need to find out why an initramfs I build on the system is una= ble to boot.

When I’m in dracut, I was able to do some checking.  Using the l= vm commands, I was able to activate and deactivate the vg, display the lvs,= but not use them.  I would get “device not found”, and in= /dev/mapper, the lvm devices were not there.  There were other lvm de= vices there, but not from my primary system lvm.  The only thing that = stands out about it is that it is ext4, not ext3.

Comparing the decompiled initrds, there’s no modules that stand out. =  Sas.out is a listing of the WORKING initrd, tfs.out is the NOT WORKIN= G initrd.

[root@testfs ramdisks]# diff sas.out tfs.out
718d717
< ./lib/firmware/ql2500_fw.bin
1248,1249d1246
< ./etc/ld.so.conf.d/openmotif-64.conf
< ./etc/ld.so.conf.d/kernel-2.6.38.7-30.fc15.x86_64.conf
1252c1249
< ./etc/ld.so.conf.d/kernel-2.6.34.8-68.fc13.x86_64.conf
---
> ./etc/ld.so.conf.d/qt-x86_64.conf
1262d1258
< ./etc/modprobe.d/openfwwf.conf
1278a1275
> ./etc/udev/rules.d/40-multipath.rules
1317a1315
> ./lib64/libcryptsetup.so.1.2.0
1322a1321
> ./lib64/libcap.so
1407d1405
< ./lib64/libcryptsetup.so.1.1.0
1419a1418
> ./lib64/libnss_files-2.12.2.so
1480a1480
> ./usr/bin/checkisomd5
1491d1490
< ./usr/lib64/libnss_files.so
1517a1517
> ./usr/share/plymouth/themes/default.plymouth

Nothing there looks relevant.  Libcryptsetup needs updating on the tes= t system I pulled the initrd from.  The q-logic firmware is irrelevant= .  There are a couple of older kernels on the test system.  Libca= p.so is a dependency of legato backup software.  Libnss_files should h= ave nothing to do with accessing an LVM.

So I’m stuck.  My system is up, but only if I build an initrd on= another system and copy it over, meaning next time I update the kernel it = will not boot.  Nothing’s jumping out at me as obviously missing= modules from the broken initrd.  Where else can I look?

-Brian


On 6/29/11 1:47 AM, "Brian Stamper" <brian.p.stamper@nasa.gov> wrote:

I just upgraded a Fedora system from 13 to = 15, and now I can’t get it to boot.  I’m getting dumped to= a dracut shell and getting the error dm-linear device lookup failed.  = ;In the dracut output with debug on, I see dracut scan my drives, and I see= it find my 3 LVs, but it then tries to do something with them, I get the e= rror, and it then fails out with root device not found and drops me into dr= acut.

Here’s the odd thing, booting off the rescue dvd my volumes are fine.=  The LVM can be mounted no problem.

This makes me think that there’s something missing from the kernel or= initrd.  I can’t really figure out what though.  I’v= e gone through the output of lsmod when running from the rescue dvd.  = I’ve even build a few ramdisks including all the modules that I thoug= ht might be related (dm, lvm, mdraid, dmraid, multipath).  I tried bui= lding them with mkinitrd —with=3Dmodule, and I tried again with dracu= t —add=3Dmodules.  No change in results.

I’m not sure where to go from here.  I feel like the system can = be rescued, but I don’t know how to work backward through dracut to f= ind out why it’s unable to find my root volume.  Any help would = be appreciated.  Cross-posting to dm-devel.

Thanks,
-Brian
--_000_CA30CF5111CACbrianpstampernasagov_--