Bug 1241774

After some investigation, I found kpartx (from device-mapper-multipath) set the uuid in read_sun_pt() in kpartx/sun.c. It read some blocks in the multipath device and get the 8 partition information.

$9 = {
{start_cylinder = 0, num_sectors = 0},
{start_cylinder = 0, num_sectors = 1078071040},
{start_cylinder = 0, num_sectors = 351010560},
{start_cylinder = 0, num_sectors = 0}, 
{start_cylinder = 0, num_sectors = 0}, 
{start_cylinder = 0, num_sectors = 0}, 
{start_cylinder = 0, num_sectors = 0}, 
{start_cylinder = 1056964608, num_sectors = 2507796480}}

The partitions with num_sectors = 0 are skipped, and get the following partition list including the 3 partitions.

(gdb) print slices[0]
$16 = {start = 0, size = 1000000, major = 0, minor = 0}
(gdb) print slices[1]
$17 = {start = 0, size = 16771860, major = 0, minor = 0}
(gdb) print slices[2]
$18 = {start = 1012095, size = 15759765, major = 0, minor = 0}

slices[0] is the partition 2 in the disk with the sun label.
slices[1] is the whole disk.
slices[2] is the partition 8 in the disk with the sun label.

slices[1] is ignored since it's overlapped with slices[2].

kpartx uses the index of slices[] array for the partition number in uuid. Precisely, it's index+1. 

For the partition 2, part1-xxx
For the partition 8, part3-xxx

So, the partition number in uuid doesn't match the partition number (2 and 8) in the disk with sun label.

I was not sure which component it's fixed in. util-linux or device-mapper-multipath. If it's fixed in device-mapper-multipath, the patch should be like this.

diff -up multipath-tools-130222/kpartx/sun.c.original multipath-tools-130222/kpartx/sun.c
--- multipath-tools-130222/kpartx/sun.c.original        2015-07-10 17:32:47.763111142 +0900
+++ multipath-tools-130222/kpartx/sun.c 2015-07-10 17:33:27.363111142 +0900
@@ -82,8 +82,6 @@ read_sun_pt(int fd, struct slice all, st
        for(i=0, n=0; i<SUN_DISK_MAXPARTITIONS; i++) {
                s = &l->partitions[i];
 
-               if (s->num_sectors == 0)
-                       continue;
                if (n < ns) {
                        sp[n].start = offset +
                                be32_to_cpu(s->start_cylinder) * be16_to_cpu(l->nsect) * be16_to_cpu(l->ntrks);


I'll change the component into device-mapper-multipath.

I've fixed the issue where kpartx is not correctly naming the partition device.  However, as you mentioned in Comment 2, this appears to be unrelated to the blkid issue.  Specifically, the blkid issue occurs exactly the same on a regular scsi device, without multipath being involved at all.

kpartx now correctly counts the zero sector sized partitions when determining that partition number, and names the partition devices correctly.

Reproduced on device-mapper-multipath-0.4.9-77.el7
1. # rpm -qa | grep multipath
device-mapper-multipath-0.4.9-77.el7.x86_64
device-mapper-multipath-libs-0.4.9-77.el7.x86_64

2.# multipath -ll
mpatha (360fff19abdd9552f8a36e5355226ba27) dm-0 EQLOGIC ,100E-00         
size=50G features='0' hwhandler='0' wp=rw
|-+- policy='service-time 0' prio=1 status=active
| `- 3:0:0:0 sdb 8:16 active ready running
`-+- policy='service-time 0' prio=1 status=enabled
  `- 4:0:0:0 sdc 8:32 active ready running

3.# fdisk -l /dev/mapper/mpatha

Disk /dev/mapper/mpatha (Sun disk label): 64 heads, 32 sectors, 51210 cylinders
Units = sectors of 1 * 512 bytes

             Device Flag    Start       End    Blocks   Id  System
/dev/mapper/mpatha2  u  104775680 104878080     51200   82  Linux swap
/dev/mapper/mpatha3             0 104878080  52439040    5  Whole disk
/dev/mapper/mpatha8         45056    345666    150305   83  Linux native

4.# parted /dev/mapper/mpatha
GNU Parted 3.1
Using /dev/mapper/mpatha
Welcome to GNU Parted! Type 'help' to view a list of commands.
(parted) print                                                            
Model: Linux device-mapper (multipath) (dm)
Disk /dev/mapper/mpatha: 53.7GB
Sector size (logical/physical): 512B/512B
Partition Table: sun
Disk Flags: 

Number  Start   End     Size    File system  Flags
 8      23.1MB  177MB   154MB
 2      53.6GB  53.7GB  52.4MB

5.# ls -l /dev/mapper/mpatha*
lrwxrwxrwx. 1 root root 7 Jul 16 15:14 /dev/mapper/mpatha -> ../dm-0
lrwxrwxrwx. 1 root root 7 Jul 16  2016 /dev/mapper/mpatha1 -> ../dm-1
lrwxrwxrwx. 1 root root 7 Jul 16  2016 /dev/mapper/mpatha2 -> ../dm-2

6.# blkid -po udev /dev/mapper/mpatha1
ID_FS_UUID=dc3cae86-81ea-406e-9207-05a857264a0c
ID_FS_UUID_ENC=dc3cae86-81ea-406e-9207-05a857264a0c
ID_FS_TYPE=xfs
ID_FS_USAGE=filesystem

7.# blkid -po udev /dev/mapper/mpatha2
ID_FS_UUID=AiQMMf-nFYA-Kita-GweC-NuPx-ScQl-i3WOHl
ID_FS_UUID_ENC=AiQMMf-nFYA-Kita-GweC-NuPx-ScQl-i3WOHl
ID_FS_VERSION=LVM2\x20001
ID_FS_TYPE=LVM2_member
ID_FS_USAGE=raid



Verified on device-mapper-multipath-0.4.9-93.el7
1. # rpm -qa | grep multipath 
device-mapper-multipath-libs-0.4.9-93.el7.x86_64
device-mapper-multipath-0.4.9-93.el7.x86_64
device-mapper-multipath-sysvinit-0.4.9-93.el7.x86_64
device-mapper-multipath-debuginfo-0.4.9-93.el7.x86_64

2.# multipath -ll
mpatha (36001405848565fc8934460cac7148f3d) dm-0 LIO-ORG ,st72-io         
size=10G features='0' hwhandler='0' wp=rw
|-+- policy='service-time 0' prio=1 status=active
| `- 3:0:0:0  sdb 8:16 active ready running
`-+- policy='service-time 0' prio=1 status=enabled
  `- 4:0:0:0  sdc 8:32 active ready running

3.# fdisk -l /dev/mapper/mpatha

Disk /dev/mapper/mpatha (Sun disk label): 64 heads, 32 sectors, 10240 cylinders
Units = sectors of 1 * 512 bytes

             Device Flag    Start       End    Blocks   Id  System
/dev/mapper/mpatha2  u   20869120  20971520     51200   82  Linux swap
/dev/mapper/mpatha3             0  20971520  10485760    5  Whole disk
/dev/mapper/mpatha8        346112    435678     44783   83  Linux native

4.# parted /dev/mapper/mpatha
GNU Parted 3.1
Using /dev/mapper/mpatha
Welcome to GNU Parted! Type 'help' to view a list of commands.
(parted) print                                                            
Model: Linux device-mapper (multipath) (dm)
Disk /dev/mapper/mpatha: 10.7GB
Sector size (logical/physical): 512B/512B
Partition Table: sun
Disk Flags: 

Number  Start   End     Size    File system  Flags
 8      177MB   223MB   45.9MB
 2      10.7GB  10.7GB  52.4MB

5.# ls -l /dev/mapper/mpatha*
lrwxrwxrwx. 1 root root 7 Jul 16 15:34 /dev/mapper/mpatha -> ../dm-0
lrwxrwxrwx. 1 root root 7 Jul 16 12:00 /dev/mapper/mpatha1 -> ../dm-1
lrwxrwxrwx. 1 root root 7 Jul 16 15:34 /dev/mapper/mpatha2 -> ../dm-6
lrwxrwxrwx. 1 root root 7 Jul 16 12:00 /dev/mapper/mpatha3 -> ../dm-7
lrwxrwxrwx. 1 root root 8 Jul 16 15:34 /dev/mapper/mpatha8 -> ../dm-10

 the partition 2 is /dev/mapper/mpatha2, and the partition 8 is /dev/mapper/mpatha8.

6.# # blkid -po udev /dev/mapper/mpatha2
ID_PART_ENTRY_SCHEME=sun
ID_PART_ENTRY_TYPE=0x82
ID_PART_ENTRY_FLAGS=0x1
ID_PART_ENTRY_NUMBER=2
ID_PART_ENTRY_OFFSET=20869120
ID_PART_ENTRY_SIZE=102400
ID_PART_ENTRY_DISK=253:0

7.# blkid -po udev /dev/mapper/mpatha8
ID_PART_ENTRY_SCHEME=sun
ID_PART_ENTRY_TYPE=0x83
ID_PART_ENTRY_NUMBER=8
ID_PART_ENTRY_OFFSET=346112
ID_PART_ENTRY_SIZE=89566
ID_PART_ENTRY_DISK=253:0


Test result: the kpartx generated devices have the same partition number as the actual partition number

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://rhn.redhat.com/errata/RHBA-2016-2536.html