| 1 |
On Tue, 2011-03-22 at 00:37 -0100, Jorge Manuel B. S. Vicetto wrote: |
| 2 |
> I haven't seen anyone mention that the |
| 3 |
> latest udev versions react *very* badly to CONFIG_SYSFS_DEPRECATED_V2. |
| 4 |
> So be sure to check if you disable that as iirc udev will refuse to |
| 5 |
> create the proper device nodes if that kernel option is active. |
| 6 |
|
| 7 |
Jorge, thanks VERY MUCH for the heads-up on CONFIG_SYSFS_DEPRECATED_V2. |
| 8 |
Making sure that this, and CONFIG_SYSFS_DEPRECATED are off in my kernel |
| 9 |
config has at least gotten the system to a terminal prompt - no single |
| 10 |
user mode and no kernel panic. Nvidia support is broken, so no X, but |
| 11 |
that's always a hassle with new kernels and can probably be fixed, |
| 12 |
assuming that support for my ancient Nvidia card is still available in |
| 13 |
the portage tree and compatible with kernel 2.6.36. I've done this many |
| 14 |
times before and I'll deal with it later for this upgrade. |
| 15 |
|
| 16 |
Device mapper configuration seems to be broken, which is where I'm at |
| 17 |
now. Kernel config settings for RAID autodetect and device mapper |
| 18 |
support are exactly as before, but the RAID subsystem appears to be |
| 19 |
stumbling on a device name conflict, which has prevented the LVM VG from |
| 20 |
being constituted. |
| 21 |
|
| 22 |
Looking into the kernel log file for the 2.6.36 boot up ... |
| 23 |
|
| 24 |
Mar 24 12:18:53 [kernel] [ 1.229464] md: Autodetecting RAID arrays. |
| 25 |
Mar 24 12:18:53 [kernel] [ 1.281388] md: Scanned 2 and added 2 devices. |
| 26 |
Mar 24 12:18:53 [kernel] [ 1.281566] md: autorun ... |
| 27 |
Mar 24 12:18:53 [kernel] [ 1.281739] md: considering sdc1 ... |
| 28 |
Mar 24 12:18:53 [kernel] [ 1.281918] md: adding sdc1 ... |
| 29 |
Mar 24 12:18:53 [kernel] [ 1.282096] md: adding sdb1 ... |
| 30 |
Mar 24 12:18:53 [kernel] [ 1.282273] md: created md0 |
| 31 |
Mar 24 12:18:53 [kernel] [ 1.282453] md: bind<sdb1> |
| 32 |
Mar 24 12:18:53 [kernel] [ 1.282639] md: bind<sdc1> |
| 33 |
Mar 24 12:18:53 [kernel] [ 1.282826] md: running: <sdc1><sdb1> |
| 34 |
Mar 24 12:18:53 [kernel] [ 1.283219] md: personality for level 1 is not loaded! |
| 35 |
Mar 24 12:18:53 [kernel] [ 1.283401] md: do_md_run() returned -22 |
| 36 |
Mar 24 12:18:53 [kernel] [ 1.283581] md: md0 still in use. |
| 37 |
Mar 24 12:18:53 [kernel] [ 1.283756] md: ... autorun DONE. |
| 38 |
|
| 39 |
This is fubar! There are two RAID arrays which support the missing LVM |
| 40 |
volume group, and they're seriously out of kilter here. The correct |
| 41 |
configuration, booting from my old kernel, is: |
| 42 |
|
| 43 |
$ cat /proc/mdstat |
| 44 |
Personalities : [raid1] |
| 45 |
md1 : active raid1 sdc1[0] sdd1[1] |
| 46 |
36146112 blocks [2/2] [UU] |
| 47 |
|
| 48 |
md0 : active raid1 sdb1[1] sda1[0] |
| 49 |
117218176 blocks [2/2] [UU] |
| 50 |
|
| 51 |
The Linux RAID subsystem is trying to pair drives drives from two |
| 52 |
different RAID arrays into a single md! I see the possibility for |
| 53 |
serious data corruption here :-( Although it looks as if the RAID |
| 54 |
subsystem may have bailed on the RAID array, sensing that something was |
| 55 |
awry. |
| 56 |
|
| 57 |
The root of this problem is that on the old kernel, there are both |
| 58 |
a /dev/hda1 and a /dev/sda1. The former is a partition on an old PATA |
| 59 |
drive, while the latter is a proper component of md0, but when |
| 60 |
everything becomes /dev/sdNx, there's an obvious conflict and the RAID |
| 61 |
subsystem is getting confused and is obviously not seeing it's sda1. |
| 62 |
|
| 63 |
I suspect that this is why the main VG isn't getting set up. What can |
| 64 |
be done about this? |
| 65 |
|
| 66 |
-- |
| 67 |
Lindsay Haisley | "Fighting against human creativity is like |
| 68 |
FMP Computer Services | trying to eradicate dandelions" |
| 69 |
512-259-1190 | (Pamela Jones) |
| 70 |
http://www.fmp.com | |
Archives