--- Begin Message ---
Package: initrd-tools
Version: Lenny (0.1.84.2)
Severity: important
I just upgraded a fully up-to-date Etch system to Lenny on a guest in
VMware Server 1. When I reboot into the 2.6.26-1-686 kernel I get stuck
at: "Begin: Waiting for root file system ..." during boot. It
eventually times-out to initramfs busybox prompt. When I poke around in
there:
echo $ROOT = /dev/sda1 # As expected/desired
ls -l /dev = missing sd* # This is bad!
When I try my previous kernel 2.6.18-6-686, it works fine.
When I copied the file system from the SCSI to an IDE virtual disk,
removed the SCSI and ran only the IDE, both above kernels work fine.
I tried the following (and possibly other things I'm forgetting) on the
SCSI disk but nothing except the above worked:
aptitude reinstall linux-image-2.6-686
update-initramfs -k all -u
dpkg-reconfigure linux-image-2.6.26-1-686
So I am thinking that something that the initrd-tools are failing to do
is messing this up, so that you can't boot into 2.6.26-1-686 as a VMware
guest using an LSI SCSI virtual disk. Note that using this SCSI virtual
device is the VMware Workstation 5.5.9 "recommended" default when
creating a new VM set for Linux, Ubuntu.
Possibly related, I'm also getting errors like this from this guest VM
under the 2.6.18-6-686 kernel, once or twice an hour (pretty low low
machine):
kernel: mptscsih: ioc0: attempting task abort! (sc=dfbc4500)
kernel: sd 0:0:0:0:
kernel: command: Write(10): 2a 00 03 a4 13 5f 00 00 48 00
kernel: mptbase: ioc0: IOCStatus(0x004b): SCSI IOC Terminated
kernel: mptscsih: ioc0: task abort: SUCCESS (sc=dfbc4500)
I think that upgrading to the latest kernel would fix those, except due
to this bug I can't. (Well, since I switched to IDE I can and did.)
Work-Arounds
------------
1) Revert to a previous kernel (e.g. 2.6.18-6-686 works for me)
2) Don't use SCSI disks in VMware VMs when running Debian (which I've
previously found to be a Good Idea anyway, but forgot this time).
3) Convert an existing host from SCSI to IDE on (tedious: briefly, the
only way I know how to do that is here:
http://lists.netisland.net/archives/plug/plug-2008-12/msg00191.html)
--
Later,
JP
----------------------------|:::======|-------------------------------
JP Vossen, CISSP |:::======| jp{at}jpsdomain{dot}org
My Account, My Opinions |=========| http://www.jpsdomain.org/
----------------------------|=========|-------------------------------
"Microsoft Tax" = the additional hardware & yearly fees for the add-on
software required to protect Windows from its own poorly designed and
implemented self, while the overhead incidentally flattens Moore's Law.
--- End Message ---
--- Begin Message ---
On Sat, Dec 20, 2008 at 06:50:11PM -0500, JP Vossen wrote:
> Package: initrd-tools
> Version: Lenny (0.1.84.2)
> Severity: important
>
> I just upgraded a fully up-to-date Etch system to Lenny on a guest in
> VMware Server 1. When I reboot into the 2.6.26-1-686 kernel I get stuck
> at: "Begin: Waiting for root file system ..." during boot. It
> eventually times-out to initramfs busybox prompt. When I poke around in
> there:
> echo $ROOT = /dev/sda1 # As expected/desired
> ls -l /dev = missing sd* # This is bad!
>
> When I try my previous kernel 2.6.18-6-686, it works fine.
>
> When I copied the file system from the SCSI to an IDE virtual disk,
> removed the SCSI and ran only the IDE, both above kernels work fine.
>
> I tried the following (and possibly other things I'm forgetting) on the
> SCSI disk but nothing except the above worked:
> aptitude reinstall linux-image-2.6-686
> update-initramfs -k all -u
> dpkg-reconfigure linux-image-2.6.26-1-686
>
> So I am thinking that something that the initrd-tools are failing to do
> is messing this up, so that you can't boot into 2.6.26-1-686 as a VMware
> guest using an LSI SCSI virtual disk. Note that using this SCSI virtual
> device is the VMware Workstation 5.5.9 "recommended" default when
> creating a new VM set for Linux, Ubuntu.
>
> Possibly related, I'm also getting errors like this from this guest VM
> under the 2.6.18-6-686 kernel, once or twice an hour (pretty low low
> machine):
> kernel: mptscsih: ioc0: attempting task abort! (sc=dfbc4500)
> kernel: sd 0:0:0:0:
> kernel: command: Write(10): 2a 00 03 a4 13 5f 00 00 48 00
> kernel: mptbase: ioc0: IOCStatus(0x004b): SCSI IOC Terminated
> kernel: mptscsih: ioc0: task abort: SUCCESS (sc=dfbc4500)
>
> I think that upgrading to the latest kernel would fix those, except due
> to this bug I can't. (Well, since I switched to IDE I can and did.)
>
>
> Work-Arounds
> ------------
> 1) Revert to a previous kernel (e.g. 2.6.18-6-686 works for me)
> 2) Don't use SCSI disks in VMware VMs when running Debian (which I've
> previously found to be a Good Idea anyway, but forgot this time).
> 3) Convert an existing host from SCSI to IDE on (tedious: briefly, the
> only way I know how to do that is here:
> http://lists.netisland.net/archives/plug/plug-2008-12/msg00191.html)
device names are not stable.
use UUID as bootarg if you need a reliable root.
closing as not a bug.
--
maks
--- End Message ---