| 1 |
Mark wrote: |
| 2 |
|
| 3 |
>Cross posting is not very polite. How should we, on this list, know if |
| 4 |
>someone has helped you yet? |
| 5 |
> |
| 6 |
> |
| 7 |
By looking on the other groups it was posted to. |
| 8 |
|
| 9 |
>This is the same as being in a grocery store, when a new counter row is |
| 10 |
>opened and a big lady bumps everyone aside just to be the first one in |
| 11 |
>the row. How rude. |
| 12 |
> |
| 13 |
> |
| 14 |
It's not even close to being the same thing. If you don't want to reply |
| 15 |
to a question on the basis that it is cross-posted, feel free not to reply. |
| 16 |
|
| 17 |
>Mark |
| 18 |
> |
| 19 |
>Joe Harvell wrote: |
| 20 |
> |
| 21 |
> |
| 22 |
>>I am in the process of installing Gentoo on an x86 from the 2006.0 |
| 23 |
>>installer CD using the 2.4 kernel profile. I created an LVM2 volume |
| 24 |
>>group with logical volumes / filesystems for '/', '/usr', '/var', |
| 25 |
>>'/opt', and '/home'. I had no problems mounting those filesystems and |
| 26 |
>>then chrooting into that environment and performing the base install. |
| 27 |
>>I compiled the 2.4.32-r4 kernel from gentoo-sources with no problems, |
| 28 |
>>and created the initial ram disk by hand so that I could bring up the |
| 29 |
>>LVM2 volumes. |
| 30 |
>> |
| 31 |
>>The problem I am having is when I boot into the initrd. I run vgscan |
| 32 |
>>with no apparent problems, and it finds the volume group. When I run |
| 33 |
>>'vgchange -ay <vgname>, I notice three problems: |
| 34 |
>> |
| 35 |
>>1. There are entries in /dev/mapper/<vgname>-<lvname>, but no |
| 36 |
>>symlinks in the form /dev/<vgname>/<lvname> --> |
| 37 |
>>/dev/mapper/<vgname>/<lvname>. I run 'lvm vgmknodes' and it remedies |
| 38 |
>>this. |
| 39 |
>> |
| 40 |
>>2. I see five complaints from 'vgchange -ay <vgname>' (one for each |
| 41 |
>>LV) as follows: |
| 42 |
>> |
| 43 |
>>on stderr: device-mapper: table ioctl failed: Invalid argument |
| 44 |
>>on console (from dm-mod driver): device-mapper: one of name or uuid |
| 45 |
>>must be supplied, cmd (12) |
| 46 |
>> |
| 47 |
>>I ran vgchange under strace and here is the following snippet: |
| 48 |
>> |
| 49 |
>>open("/dev/mapper/control", O_RDWR|O_LARGEFILE) = 9 |
| 50 |
>>ioctl(9, DM_DEV_STATUS, 0x8109958) = -1 ENXIO (No such device or |
| 51 |
>>address) |
| 52 |
>>ioctl(9, DM_DEV_STATUS, 0x8109958) = -1 ENXIO (No such device or |
| 53 |
>>address) |
| 54 |
>>ioctl(9, DM_DEV_STATUS, 0x8109958) = -1 ENXIO (No such device or |
| 55 |
>>address) |
| 56 |
>>mlockall(MCL_CURRENT|MCL_FUTURE) = 0 |
| 57 |
>>brk(0x8936000) = 0x8936000 |
| 58 |
>>brk(0x8b1e000) = 0x8b1e000 |
| 59 |
>>getpriority(PRIO_PROCESS, 0) = 20 |
| 60 |
>>setpriority(PRIO_PROCESS, 0, -18) = 0 |
| 61 |
>>ioctl(9, DM_DEV_STATUS, 0x8109958) = -1 ENXIO (No such device or |
| 62 |
>>address) |
| 63 |
>>ioctl(9, DM_DEV_STATUS, 0x8109958) = -1 ENXIO (No such device or |
| 64 |
>>address) |
| 65 |
>>ioctl(9, DM_DEV_STATUS, 0x8109958) = -1 ENXIO (No such device or |
| 66 |
>>address) |
| 67 |
>>ioctl(9, DM_DEV_STATUS, 0x8109958) = -1 ENXIO (No such device or |
| 68 |
>>address) |
| 69 |
>>ioctl(9, DM_DEV_STATUS, 0x8109958) = -1 ENXIO (No such device or |
| 70 |
>>address) |
| 71 |
>>ioctl(9, DM_DEV_STATUS, 0x8109958) = -1 ENXIO (No such device or |
| 72 |
>>address) |
| 73 |
>>ioctl(9, DM_DEV_STATUS, 0x8109958) = -1 ENXIO (No such device or |
| 74 |
>>address) |
| 75 |
>>ioctl(9, DM_DEV_STATUS, 0x8109958) = -1 ENXIO (No such device or |
| 76 |
>>address) |
| 77 |
>>ioctl(9, DM_DEV_STATUS, 0x8109958) = -1 ENXIO (No such device or |
| 78 |
>>address) |
| 79 |
>>stat64("/dev/ide/host0/bus0/target0/lun0/part10", |
| 80 |
>>{st_mode=S_IFBLK|0600, st_rdev=makedev(3, 10), ...}) = 0 |
| 81 |
>>stat64("/dev/ide/host0/bus0/target0/lun0/part11", |
| 82 |
>>{st_mode=S_IFBLK|0600, st_rdev=makedev(3, 11), ...}) = 0 |
| 83 |
>>stat64("/dev/ide/host0/bus0/target0/lun0/part12", |
| 84 |
>>{st_mode=S_IFBLK|0600, st_rdev=makedev(3, 12), ...}) = 0 |
| 85 |
>>ioctl(9, DM_DEV_CREATE, 0x8109958) = 0 |
| 86 |
>>ioctl(9, DM_TABLE_STATUS, 0x8109958) = -1 EINVAL (Invalid argument) |
| 87 |
>>write(2, " ", 2) = 2 |
| 88 |
>>write(2, "device-mapper: table ioctl faile"..., 51) = 51 |
| 89 |
>>write(2, "\n", 1) = 1 |
| 90 |
>> |
| 91 |
>>3. When I try to mount the filesystem in that LV, it fails. I tried |
| 92 |
>>with different filesystem types and it's the same result. From within |
| 93 |
>>the 2.4 kernel based initrd, it fails. If I reboot into the 2.6 |
| 94 |
>>kernel based installer environment, no problem. |
| 95 |
>> |
| 96 |
>>This looks like an incompatibility between the lvm2 and/or |
| 97 |
>>device-mapper library code and the kernel device-mapper code. Here |
| 98 |
>>are the versions installed in the new installation and its initrd: |
| 99 |
>> |
| 100 |
>>lvm2-2.02.05 |
| 101 |
>>device-mapper-1.02.03 |
| 102 |
>>gentoo-sources-2.4.32-r4 |
| 103 |
>> |
| 104 |
>>livecd 2.4 # ls -l /etc/make.profile |
| 105 |
>>lrwxrwxrwx 1 root root 51 May 11 18:09 /etc/make.profile -> |
| 106 |
>>/usr/portage/profiles/default-linux/x86/no-nptl/2.4 |
| 107 |
>> |
| 108 |
>> |
| 109 |
>> |
| 110 |
|
| 111 |
-- |
| 112 |
gentoo-admin@g.o mailing list |
Archives