Discussion:
Supermicro cd/dvd inst error: ARCH_201603 device did not show up after 30 seconds
(too old to reply)
David C. Rankin
2016-09-21 03:24:32 UTC
Permalink
All,

First let me explain I have search through the forums, StackExchange site
(SuperUser & Linux Unix StackExhange) and read the problems with similar title,
90% of which were USB install and related to failure to have the proper iso name
(e.g ARCH_201603) as the image name, the other 10% were related to
/dev/disk/by-label not being created and suggest the creation of the directory
and symlink of the install device (e.g. /dev/sr0 to
/dev/disk/by-label/ARCH_201603) and exiting the maintenance shell to have the
install pickup and continue. Neither of these are a solution.

I have one Supermicro server (similar but slightly older) where the install
went without issue (H8QM8-2 w/4 Opteron 8360SE). This server is the
H8DME-2-LS006 Rev 1.10 w/2 Opteron 8382 processors. (specs:
http://www.supermicro.com/Aplus/motherboard/Opteron2000/MCP55/H8DME-2.cfm)

Initially, the boot would hang during pci activation with the line

pci 000:00:1a.0: [1a:1e:1100] type -- class 0x060000

I reset the bios to the failsafe defaults, still no joy.

The Hardware Detection Tool showed a 3Leaf Systems, Inc.|unknown device at
that address, the details of which were:

Vendor : 3Leaf Systems, Inc.
Product : unknown
Class : Host bridge
Location : 00:1a.0
PCI ID : 1a1e:1100[0000:0000]
Latency : 0
Module : unknown

Looking though all the bios options, trying one at a time, I found an option
for something related to AQUA (? This board an AQUA were apparently the
brainchild of 3Leaf, Inc. now out of business
http://www.eetimes.com/document.asp?doc_id=1172150 which allowed you to link 32
Opteron processors into a single server)

So disabling AQUA completely in the bios allowed the box to boot normally all
the way though:

:: running hook [archiso_pxe_nfs]
:: Mounting '/dev/disk/by-label/ARCH_201603' to '/run/archiso/bootmnt'
Waiting 30 seconds for devices....

then the error and I'm dropped to the maintenance shell (e.g. [rootfs /]# ).
I've tried the solutions shown at the beginning but I still can't proceed with
the remainder of the install. When I create the /dev/disk/by-label directory,
create the symlink, and exit the maintenance shell, I get the following error:

Buffer I/O error on dev sr0 logical block 0, async page read
sr 6:0:1:0 [sr0] tag#0 UNKNOWN(0x2003) Result:hostbyte-0x00 driverbyte=0x08
sr 6:0:1:0 [sr0] tag#0 Sense Key : 0x4 [current]
sr 6:0:1:0 [sr0] tag#0 ASC=0x8 ASCQ=0x3
sr 6:0:1:0 [sr0] tag#0 CDB: opcode: 0x28 28 00 00 00 00 00 00 00 02 00
blk_update_request: I/O error, dev sr0, sector 0
isofs_fill_super: bread failed, dev=sr0, iso_blknum=16, block=16

So I need help from the smart folks here. I'm not entirely sure what this
error is telling me other than there is some UNKNOWN result that prevents the
archiso mount. With this AQUA setting completely disabled in the bios, I don't
know if this error is related to that setting, but I suspect it is. What could I
try to get past this error? Is there some kernel parameter I coud try to pass at
boot that might help? Has anyone else seen this or found another solution?

Thanks for any advise you can give..
--
David C. Rankin, J.D.,P.E.
Dave via arch-general
2016-09-21 03:30:19 UTC
Permalink
Did you check this?

Installation problems: I/O erro: dev sr0 / Installation / Arch Linux Forums
https://bbs.archlinux.org/viewtopic.php?id=195763


On Tue, Sep 20, 2016 at 11:24 PM, David C. Rankin <
Post by David C. Rankin
All,
First let me explain I have search through the forums, StackExchange site
(SuperUser & Linux Unix StackExhange) and read the problems with similar title,
90% of which were USB install and related to failure to have the proper iso name
(e.g ARCH_201603) as the image name, the other 10% were related to
/dev/disk/by-label not being created and suggest the creation of the directory
and symlink of the install device (e.g. /dev/sr0 to
/dev/disk/by-label/ARCH_201603) and exiting the maintenance shell to have the
install pickup and continue. Neither of these are a solution.
I have one Supermicro server (similar but slightly older) where the install
went without issue (H8QM8-2 w/4 Opteron 8360SE). This server is the
http://www.supermicro.com/Aplus/motherboard/Opteron2000/MCP55/H8DME-2.cfm)
Initially, the boot would hang during pci activation with the line
pci 000:00:1a.0: [1a:1e:1100] type -- class 0x060000
I reset the bios to the failsafe defaults, still no joy.
The Hardware Detection Tool showed a 3Leaf Systems, Inc.|unknown device at
Vendor : 3Leaf Systems, Inc.
Product : unknown
Class : Host bridge
Location : 00:1a.0
PCI ID : 1a1e:1100[0000:0000]
Latency : 0
Module : unknown
Looking though all the bios options, trying one at a time, I found an option
for something related to AQUA (? This board an AQUA were apparently the
brainchild of 3Leaf, Inc. now out of business
http://www.eetimes.com/document.asp?doc_id=1172150 which allowed you to link 32
Opteron processors into a single server)
So disabling AQUA completely in the bios allowed the box to boot normally all
:: running hook [archiso_pxe_nfs]
:: Mounting '/dev/disk/by-label/ARCH_201603' to '/run/archiso/bootmnt'
Waiting 30 seconds for devices....
then the error and I'm dropped to the maintenance shell (e.g. [rootfs /]# ).
I've tried the solutions shown at the beginning but I still can't proceed with
the remainder of the install. When I create the /dev/disk/by-label directory,
Buffer I/O error on dev sr0 logical block 0, async page read
sr 6:0:1:0 [sr0] tag#0 UNKNOWN(0x2003) Result:hostbyte-0x00 driverbyte=0x08
sr 6:0:1:0 [sr0] tag#0 Sense Key : 0x4 [current]
sr 6:0:1:0 [sr0] tag#0 ASC=0x8 ASCQ=0x3
sr 6:0:1:0 [sr0] tag#0 CDB: opcode: 0x28 28 00 00 00 00 00 00 00 02 00
blk_update_request: I/O error, dev sr0, sector 0
isofs_fill_super: bread failed, dev=sr0, iso_blknum=16, block=16
So I need help from the smart folks here. I'm not entirely sure what this
error is telling me other than there is some UNKNOWN result that prevents the
archiso mount. With this AQUA setting completely disabled in the bios, I don't
know if this error is related to that setting, but I suspect it is. What could I
try to get past this error? Is there some kernel parameter I coud try to pass at
boot that might help? Has anyone else seen this or found another solution?
Thanks for any advise you can give..
--
David C. Rankin, J.D.,P.E.
David C. Rankin
2016-09-21 05:06:54 UTC
Permalink
Post by Dave via arch-general
Did you check this?
Installation problems: I/O erro: dev sr0 / Installation / Arch Linux Forums
https://bbs.archlinux.org/viewtopic.php?id=195763
Thank you for the link,

There is much discussion but little solution there. This seems like a bug that
has plagued whoever it should go to for quite some time. I don't think this is a
dao/tao issue. I get the same error with ARCH_201603 (which I've installed Arch
from several times w/o issue) and ARCH 201609 and openSuSE 13.1. All cd's were
md5/shasum verified (before and after burning), all burning was done from the
command line per
https://wiki.archlinux.org/index.php/Optical_disc_drive#Burning_an_ISO_image_to_CD.2C_DVD.2C_or_BD
(the ARCH_201609 was to DVD since it finally exceeded the overburn of my discs).

So the forum thread discusses in 43 posts at least a dozen theories as to why
this continues to reoccur, but it doesn't seem that anybody actually has a
handle as to the exact cause. Me, I've probably installed arch 30 times over the
past 7 years and I've never seen this before. The post
https://bbs.archlinux.org/viewtopic.php?pid=1608432#p1608432 probably captures
the current thinking/experimentation as it stands now. I'll try and experiment
with a couple of additional ATA and SATA based cd/dvd drives and if I find
anything new I'll add this to the forum thread.

If anyone has any additional thoughts on this, I (and apparently many others)
would appreciate your thoughts. I can't believe my search didn't turn up that
formum thread, (StartPage is usually dead on). However, I was surprised at the
litany of different times and situations this has apparently occurred over the
past year or so. Glad to know there is still a little mystery left in computing :)
--
David C. Rankin, J.D.,P.E.
Mauro Santos via arch-general
2016-09-21 10:43:05 UTC
Permalink
Have you tried with the latest install ISO?

You also seem to have physical access to the hardware, any special
reason why you have to use optical media instead of a usb flash drive?
--
Mauro Santos
David C. Rankin
2016-10-10 18:30:06 UTC
Permalink
Post by Mauro Santos via arch-general
Have you tried with the latest install ISO?
Yes, that was my first attempt.
Post by Mauro Santos via arch-general
You also seem to have physical access to the hardware, any special
reason why you have to use optical media instead of a usb flash drive?
No, no reason other than convenience. This particular box is an older
SuperMicro H8DME-2 board with a pair of quad-core Opterons. I'll make another
run at it today and tomorrow.

I'll first attempt a PXE boot from an existing Arch server attempting to serve
the install .iso mounted loop via tftpd. My only quandary there is whether to
symlink the lpxelinux.0 contained in the install to /srv/tftp/pxelinux.0, or
whether the use the example style /srv/tftp/arch.menu discussed at
https://wiki.archlinux.org/index.php/PXE#Existing_PXE_Server (my PXE setup is
only for the install.iso, so I was intending to link
/archiso/arch/boot/syslinux/lpxelinux.0 /srv/tftp/pxelinux.0, but am unsure if
that will serve the install iso as intended)

We'll see how that goes. Failing that, I'll write the iso to USB and try that
way and report back. Thanks Mauro.
--
David C. Rankin, J.D.,P.E.
LoneVVolf
2016-10-10 20:26:17 UTC
Permalink
Post by David C. Rankin
This particular box is an older
SuperMicro H8DME-2 board with a pair of quad-core Opterons.
I just realised I have a similar motherboard (Supermicro H8DAE-2 ) with
the same chipset and also a pair of quadcore opterons on it.

Check these 2 options in the AMI Bios :

Advanced Chipset Control > Northbridge
Set IOMMU MODE to 1 GB
(unless you have an agp videocard in this system, then it should be set
to "AGP present" )


Processor & Clock Options
Set Secure Virtual Machine Mode to Enabled


Without those settings in my experience linux 64-bit kernel is very
unreliable on this system.
With those settings it's been running rock steady for 7 years now.
(it's my main workstation pc ).

LW
David C. Rankin
2016-10-10 22:49:47 UTC
Permalink
Post by LoneVVolf
Advanced Chipset Control > Northbridge
Set IOMMU MODE to 1 GB
(unless you have an agp videocard in this system, then it should be set to "AGP
present" )
Processor & Clock Options
Set Secure Virtual Machine Mode to Enabled
Without those settings in my experience linux 64-bit kernel is very unreliable
on this system.
With those settings it's been running rock steady for 7 years now.
(it's my main workstation pc ).
LW
Got both set, thanks! I think there is something funny with the arch install
iso. For some reason, I'm getting a failure waiting on the install to be mounted
at /run/archiso/bootmnt

The strange things is it says:

Waiting 30 seconds....
Error: '/dev/disk/by-label/ARCH_201603' device did not show up after 30 seconds.

WTF?? 201603, this is the 201610 install medial

Sure enough

# ls -al /dev/disk/by-label

ARCHISO_EFI -> ../../sdb2
ARCH_201610 -> ../../sdb1

Why the hell is it trying to mount 'ARCH_201603' when 'ARCH_201610' is the
actual label? I can manually mount /dev/disk/by-label/ARCH_201610 to
/run/archiso/bootmnt but I don't know how to make the install continue from that
point??

This was with the USB installer. Funny, the arch iso from both 20160903 and
20161001 both fail looking for the 'ARCH_201603' (this was a fresh USB drive
too, no prior iso on it)

"get a rope... somebody done screwed something up..."
--
David C. Rankin, J.D.,P.E.
Mauro Santos via arch-general
2016-10-10 23:29:42 UTC
Permalink
Post by David C. Rankin
Post by LoneVVolf
Advanced Chipset Control > Northbridge
Set IOMMU MODE to 1 GB
(unless you have an agp videocard in this system, then it should be set to "AGP
present" )
Processor & Clock Options
Set Secure Virtual Machine Mode to Enabled
Without those settings in my experience linux 64-bit kernel is very unreliable
on this system.
With those settings it's been running rock steady for 7 years now.
(it's my main workstation pc ).
LW
Got both set, thanks! I think there is something funny with the arch install
iso. For some reason, I'm getting a failure waiting on the install to be mounted
at /run/archiso/bootmnt
Waiting 30 seconds....
Error: '/dev/disk/by-label/ARCH_201603' device did not show up after 30 seconds.
WTF?? 201603, this is the 201610 install medial
Sure enough
# ls -al /dev/disk/by-label
ARCHISO_EFI -> ../../sdb2
ARCH_201610 -> ../../sdb1
Why the hell is it trying to mount 'ARCH_201603' when 'ARCH_201610' is the
actual label? I can manually mount /dev/disk/by-label/ARCH_201610 to
/run/archiso/bootmnt but I don't know how to make the install continue from that
point??
This was with the USB installer. Funny, the arch iso from both 20160903 and
20161001 both fail looking for the 'ARCH_201603' (this was a fresh USB drive
too, no prior iso on it)
"get a rope... somebody done screwed something up..."
You could try to inspect the boot entry. If I recall correctly the
install iso uses grub so you can check and edit the entry before booting.

If indeed the boot entries are not correct then you should file a bug
report.
--
Mauro Santos
David C. Rankin
2016-10-10 23:39:38 UTC
Permalink
Post by Mauro Santos via arch-general
You could try to inspect the boot entry. If I recall correctly the
install iso uses grub so you can check and edit the entry before booting.
If indeed the boot entries are not correct then you should file a bug
report.
Thanks Mauro,

I'm picking though the iso right now. I'll let you know what turns up. The
install works fine other than looking for disk/by-label/ARCH_201603.
--
David C. Rankin, J.D.,P.E.
David C. Rankin
2016-10-11 00:18:26 UTC
Permalink
Post by David C. Rankin
Thanks Mauro,
I'm picking though the iso right now. I'll let you know what turns up. The
install works fine other than looking for disk/by-label/ARCH_201603.
Oh my God, what moron forgot and left the 201603 install CD in the CD drive a
couple of weeks ago and then tried to boot the 20161001 installer from USB. No
wonder the installer was looking for ARCH_201603. The bizarre part was actually
looking with 'ls -al /dev/disk/by-label' only showed ARCH_201610 and
ARCHISO_EFI, lending to the confusion.

Sorry for the noise. The USB install works fine, the CD/DVD install still fails
with the issue discussed in
https://bbs.archlinux.org/viewtopic.php?pid=1608432#p1608432

Now to chastise the idiot that left 201603 CD in the drive :). If these servers
didn't sound like a 747 at full-takeoff, I might have actually noticed the CD
drive clicking :)

Thanks again Mauro, LoneWolf, Dave and the rest.
--
David C. Rankin, J.D.,P.E.
Loading...