View previous topic :: View next topic |
Author |
Message |
knirirr n00b
Joined: 20 Jan 2010 Posts: 39
|
Posted: Tue Jan 29, 2013 12:07 pm Post subject: Megaraid hassle |
|
|
The recent update to udev-197 required that I rebuild the kernel on various machines (taking the opportunity to upgrade to the newest gentoo-sources, 3.6.11), one of which is no longer bootable afterwards. It has the following hardware:
Code: |
*-pci:1
description: PCI bridge
product: 80332 [Dobson] I/O processor (B-Segment Bridge)
vendor: Intel Corporation
physical id: 0.2
bus info: pci@0000:01:00.2
version: 06
width: 32 bits
clock: 33MHz
capabilities: pci pciexpress msi pm pcix normal_decode bus_master cap_list
*-storage
description: RAID bus controller
product: MegaRAID
vendor: LSI Logic / Symbios Logic
physical id: b
bus info: pci@0000:03:0b.0
logical name: scsi2
version: 01
width: 32 bits
clock: 66MHz
capabilities: storage pm bus_master cap_list
configuration: latency=32
resources: iomemory:d8ff0000-d8ffffff irq:37
*-processor:0 UNCLAIMED
description: SCSI Processor
product: 1x2 SCSI BP
vendor: PE/PV
physical id: 0.6.0
bus info: scsi@2:0.6.0
version: 1.0
serial: 1
configuration: ansiversion=2
*-processor:1 UNCLAIMED
description: SCSI Processor
product: PV22XS
vendor: DELL
physical id: 1.6.0
bus info: scsi@2:1.6.0
version: E.17
configuration: ansiversion=3
*-disk:0
description: SCSI Disk
product: LD0 RAID1 69880R
vendor: MegaRAID
physical id: 2.0.0
bus info: scsi@2:2.0.0
logical name: /dev/sda
version: 350O
size: 68GiB (73GB)
capabilities: partitioned partitioned:dos
configuration: ansiversion=2 sectorsize=512 signature=0009fdd2
*-volume:0
description: EXT3 volume
vendor: Linux
physical id: 1
bus info: scsi@2:2.0.0,1
logical name: /dev/sda1
logical name: /
version: 1.0
serial: 722c866e-0b8f-465b-9fb6-ed72734fe558
size: 60GiB
capacity: 60GiB
capabilities: primary bootable journaled extended_attributes large_files recover ext3 ext2 initialized
configuration: created=2011-03-31 16:50:47 filesystem=ext3 label=/ modified=2013-01-29 11:49:40 mount.fstype=ext3 mount.options=rw,relatime,errors=continue,barrier=1,data=writeback mounted=2013-01-29 11:49:40 state=mounted
*-volume:1
description: Linux swap volume
physical id: 2
bus info: scsi@2:2.0.0,2
logical name: /dev/sda2
version: 1
serial: dcd8f6f1-edb9-420e-9aca-822b82ec425f
size: 8424MiB
capacity: 8424MiB
capabilities: primary nofs swap initialized
configuration: filesystem=swap pagesize=4096
*-disk:1
description: SCSI Disk
product: LD1 RAID5 38900R
vendor: MegaRAID
physical id: 2.1.0
bus info: scsi@2:2.1.0
logical name: /dev/sdb
version: 350O
size: 1502GiB (1613GB)
capabilities: partitioned partitioned:dos
configuration: ansiversion=2 sectorsize=512 signature=000a16a1
*-volume:0
description: EXT3 volume
vendor: Linux
physical id: 1
bus info: scsi@2:2.1.0,1
logical name: /dev/sdb1
version: 1.0
serial: a7b5286d-178b-481b-b8b3-734dac778c8c
size: 1502GiB
capacity: 1578GiB
capabilities: primary journaled extended_attributes large_files ext3 ext2 initialized
configuration: created=2011-04-01 09:39:14 filesystem=ext3 label=/home modified=2013-01-28 12:27:51 mounted=2013-01-28 11:59:50 state=clean
*-volume:1
description: Linux filesystem partition
physical id: 2
bus info: scsi@2:2.1.0,2
logical name: /dev/sdb2
capacity: 299GiB
capabilities: primary
*-volume:2
description: Empty partition
physical id: 3
bus info: scsi@2:2.1.0,3
logical name: /dev/sdb3
capacity: 1007GiB
capabilities: primary nofs
|
If Megaraid support is not added then, unsurprisingly, the machine won't recognise the root device /dev/sda1. If it's compiled as a module then the system hangs on boot at around this point:
Code: |
....
:: Scanning for megaraid...megaraid loaded.
:: Scanning for megaraid_mbox...megaraid_mm, megaraid_mbox loaded.
|
If I compile these various megaraid modules into the kernel instead then the result is a hang at "Switching to clocksource tsc".
The system boots perfectly to the install cd so I am currently using Genkernel to build a kernel to the same configuration in the hope that it might boot with that. Other than that I don't have any ideas, so if anyone else can see anything I may have missed then please let me know.
Edit: The install CD kernel did not work. I'm a bit stuck here, for clearly I need the megaraid driver but can't boot with it! |
|
Back to top |
|
|
nanoczar Apprentice
Joined: 10 Dec 2002 Posts: 192 Location: Henderson, TN
|
Posted: Tue Jan 29, 2013 5:19 pm Post subject: |
|
|
Just for clarification, do you have a RAID setup? The fact that it's looking for sda1 is strange if you do...at least it's not that way on my RAID system. _________________ Don't fight the LaTeX. WORK with the LaTeX!! Become one with the LaTeX!!! - JIC |
|
Back to top |
|
|
knirirr n00b
Joined: 20 Jan 2010 Posts: 39
|
Posted: Tue Jan 29, 2013 6:32 pm Post subject: |
|
|
nanoczar wrote: | Just for clarification, do you have a RAID setup? The fact that it's looking for sda1 is strange if you do...at least it's not that way on my RAID system. |
Yes. IIRC the system has one external RAID array which has been configured to present a small sda and large sdb; the former is used for the OS and the latter for data.
I've managed to get it booting using the boot CD configuration and the additional command line of clocksource=acpi_pm, although I didn't have any luck with my previous kernel configuration. So, it's sort of working, but I am not really sure what the problem is. |
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54237 Location: 56N 3W
|
Posted: Tue Jan 29, 2013 6:51 pm Post subject: |
|
|
knirirr,
Code: | description: RAID bus controller
product: MegaRAID
vendor: LSI Logic / Symbios Logic |
Its not a real raid controller. Do you use dmraid to use the software raid functionality ? _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
knirirr n00b
Joined: 20 Jan 2010 Posts: 39
|
Posted: Tue Jan 29, 2013 8:19 pm Post subject: |
|
|
NeddySeagoon wrote: |
Its not a real raid controller. Do you use dmraid to use the software raid functionality ? |
In what way is it not a real controller? I'm not using software RAID on this device, and megactl reports the following:
Code: |
a0 PERC 4/DC chan:2 ldrv:2 batt:good
a0d0 68GiB RAID 1 1x2 optimal
a0d1 1TiB RAID 5 1x12 optimal
hot spares : a0c1t8
a0c0t0 68GiB a0d0 online
a0c0t1 68GiB a0d0 online
a0c1t1 136GiB a0d1 online
a0c1t2 136GiB a0d1 online
a0c1t3 136GiB a0d1 online
a0c1t4 136GiB a0d1 online
a0c1t5 136GiB a0d1 online
a0c1t8 136GiB hotspare
a0c1t9 136GiB a0d1 online
a0c1t10 136GiB a0d1 online
a0c1t11 136GiB a0d1 online
a0c1t12 136GiB a0d1 online
a0c1t13 136GiB a0d1 online
a0c1t14 136GiB a0d1 online
a0c1t15 136GiB a0d1 online
|
|
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54237 Location: 56N 3W
|
Posted: Tue Jan 29, 2013 10:04 pm Post subject: |
|
|
knirirr,
I take that back - that looks real raid to me.
Do you have Code: | CONFIG_DEVTMPFS=y
CONFIG_DEVTMPFS_MOUNT=y | in your kernel. Genkernel will set those options.
I don't expect genkernel to work for you without a lot of help as it will probably leave out the megaraid drivers you need.
Can you put your kernel .config file on a pastebin and post the output of lspci and lspci -n please. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
knirirr n00b
Joined: 20 Jan 2010 Posts: 39
|
Posted: Wed Jan 30, 2013 9:08 am Post subject: |
|
|
NeddySeagoon wrote: |
Do you have Code: | CONFIG_DEVTMPFS=y
CONFIG_DEVTMPFS_MOUNT=y | in your kernel. Genkernel will set those options.
|
Yes, these are both enabled.
NeddySeagoon wrote: |
Can you put your kernel .config file on a pastebin and post the output of lspci and lspci -n please. |
Here's the config for the old kernel:
http://pastebin.com/CkC0Qxy4
I couldn't get this config. to boot when I used it to compile 3.6.11; I'd get a hang when loading the megaraid drivers, or that clocksource error (see earlier post) if they were compiled into the kernel.
Here's one I generated with Genkernel using the boot CD's config:
http://pastebin.com/pPZ22SyF
This one I got to boot by specifying clocksource on boot:
Code: |
# cat /proc/cmdline
root=/dev/ram0 real_root=/dev/sda1 clocksource=acpi_pm
|
However, I'm not sure why that works.
I've had to rebuild again overnight as I realised that I'd forgotten to build the iptables module, so I hope it will come up again after doing that... |
|
Back to top |
|
|
Wallsandfences Guru
Joined: 29 Mar 2010 Posts: 378
|
Posted: Wed Jan 30, 2013 11:44 am Post subject: |
|
|
There's not much I can add, but did you give the clocksource boot hang some time (minutes)? I had that hangup during the first reboot after a fresh kernel a few weeks ago, and all I did was waiting a few minutes, and then it booted, and does so ever since.
Rüdiger |
|
Back to top |
|
|
knirirr n00b
Joined: 20 Jan 2010 Posts: 39
|
Posted: Wed Jan 30, 2013 11:50 am Post subject: |
|
|
Wallsandfences wrote: | There's not much I can add, but did you give the clocksource boot hang some time (minutes)? I had that hangup during the first reboot after a fresh kernel a few weeks ago, and all I did was waiting a few minutes, and then it booted, and does so ever since.
|
I left it overnight after one hang and the machine stopped at not being able to read /dev/sda1. So, I presume that it tried for some time to load the modules and moved on after failing, only to find that the RAID array was inaccessible due to the lack of modules. |
|
Back to top |
|
|
|