Bug 926053

Summary: journalctl shows error about duplicate /dev/disk/by-id/ with hardware raid HP P420i
Product: [openSUSE] openSUSE Tumbleweed Reporter: Philippe Condé <conde.philippe>
Component: BasesystemAssignee: Thomas Blume <thomas.blume>
Status: RESOLVED INVALID QA Contact: E-mail List <qa-bugs>
Severity: Normal    
Priority: P5 - None CC: bwiedemann, chcao, conde.philippe, hare, rmilasan, systemd-maintainers, thomas.blume, tiwai
Version: 201503*   
Target Milestone: ---   
Hardware: HP   
OS: openSUSE 13.2   
Whiteboard:
Found By: --- Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: file 1 sda
file 2 sdb
file 3 sdc
output of /lib/udev/scsi_id
KInfoCenter
udevadm output

Description Philippe Condé 2015-04-06 13:51:07 UTC
User-Agent:       Mozilla/5.0 (X11; Linux x86_64; rv:34.0) Gecko/20100101 Firefox/34.0 SeaMonkey/2.31 Lightning/3.6b3
Build Identifier: 

Since kernel 19.3 on a HP proliant ML350 with a hardware raid unit P420i. there are three logical disks defined. Each logical disks with 4 partitions
In journalctl I receive each 5 minutes a lot of warnings/errors

"Apr 04 19:17:58 hpprol2 systemd[3098]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/target0:0:0/0
Apr 04 19:17:58 hpprol2 systemd[3098]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/target0:0:0/0
Apr 04 19:17:58 hpprol2 systemd[3098]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart1.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 04 19:17:58 hpprol2 systemd[3098]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart3.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 04 19:17:58 hpprol2 systemd[3098]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart4.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 04 19:17:58 hpprol2 systemd[3098]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart4.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 04 19:17:58 hpprol2 systemd[3098]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart1.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 04 19:17:58 hpprol2 systemd[3098]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart2.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 04 19:17:58 hpprol2 systemd[3098]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart3.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 04 19:17:58 hpprol2 systemd[3098]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart3.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 04 19:17:58 hpprol2 systemd[3098]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart2.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 04 19:17:58 hpprol2 systemd[3098]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart1.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 04 19:17:58 hpprol2 systemd[3098]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart4.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 04 19:23:08 hpprol2 systemd[3098]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/target0:0:0/0"

executing hwinfo --partition show that partion id 90 and 95 have the same  /dev/disk/by-id/
"hwinfo --partition
90: None 00.0: 11300 Partition                                  
  [Created at block.414]
  Unique ID: bdUI.SE1wIdpsiiC
  Parent ID: R7kM.067DVheHAD3
  SysFS ID: /class/block/sda/sda1
  Hardware Class: partition
  Model: "Partition"
  Device File: /dev/sda1
  Device Files: /dev/sda1, /dev/disk/by-id/scsi-0HP_LOGICAL_VOLUME_00000000-part1, /dev/disk/by-id/scsi-3600508b1001c99233458581ffb65cc88-part1, /dev/disk/by-id/scsi-SHP_LOGICAL_VOLUME_0014380280B60D0-part1, /dev/disk/by-id/wwn-0x600508b1001c99233458581ffb65cc88-part1, /dev/disk/by-path/pci-0000:03:00.0-scsi-0:0:0:0-part1, /dev/disk/by-uuid/ab7656ab-41e6-41e8-aa16-bbb4ca3c643e
  Config Status: cfg=no, avail=yes, need=no, active=unknown
  Attached to: #89 (Disk)
.
.
95: None 00.0: 11300 Partition
  [Created at block.414]
  Unique ID: h4pj.SE1wIdpsiiC
  Parent ID: uI_Q.067DVheHAD3
  SysFS ID: /class/block/sdb/sdb1
  Hardware Class: partition
  Model: "Partition"
  Device File: /dev/sdb1
  Device Files: /dev/sdb1, /dev/disk/by-id/scsi-0HP_LOGICAL_VOLUME_01000000-part1, /dev/disk/by-id/scsi-3600508b1001cd3e7527deb3b931a6639-part1, /dev/disk/by-id/scsi-SHP_LOGICAL_VOLUME_0014380280B60D0-part1, /dev/disk/by-id/wwn-0x600508b1001cd3e7527deb3b931a6639-part1, /dev/disk/by-label/srv, /dev/disk/by-path/pci-0000:03:00.0-scsi-0:0:0:1-part1, /dev/disk/by-uuid/bf2e6374-cee0-4c18-96b5-851f821ac807
  Config Status: cfg=no, avail=yes, need=no, active=unknown
  Attached to: #94 (Disk)"

as you can see above the same id "/dev/disk/by-id/scsi-SHP_LOGICAL_VOLUME_0014380280B60D0-part1" is the sam efor the two partitions on two different logical disks

Regards
Philippe Condé

Reproducible: Always

Steps to Reproduce:
1.boot system
2.wait 5 minutes
3.journalctl is filled with the warnings
4.after 5 minutes the warnings occur again
Actual Results:  
a lot of warnings

Expected Results:  
No warnings
Comment 1 Takashi Iwai 2015-04-07 16:24:10 UTC
Can you check that the older kernel (3.18?) showed the different device paths, to confirm that it's a kernel regression?
Comment 2 Philippe Condé 2015-04-07 20:00:08 UTC
Hello,

I have only two kernel-xen :
kernel-xen-3.19.2-1.1.x86_64
kernel-xen-devel-3.19.3-1.1.x86_64
kernel-xen-devel-3.19.2-1.1.x86_64
kernel-xen-3.19.3-1.1.x86_64

The problem occurs also with kernel 3.19.2-1.1.
I didn't find an older kernel 3.18 on OpenSuse site.

I'm practically sure that this didn't occur with kernel 3.18: I check regularly the journalctl due to a problem with some sensors and I didn't see these warnings before.
DO you have a link with a old kernel? I can download it and install it

Regards
Philippe Condé
Comment 3 Takashi Iwai 2015-04-08 07:04:08 UTC
You can find kernel packages for 3.18 in OBS home:tiwai:kernel:3.18 repo.
But I didn't build kernel-xen, so I enabled it now but it takes some time until the build finishes.  If you'll need kernel-xen, check the repo later (after one or two hours).
Comment 4 Philippe Condé 2015-04-09 20:35:11 UTC
Hello,

I tested it with your kernel 3.18 and the problem occurs also. I found this strange so I reinstalled the GA kernel 3.16.7.7 and here also the problem occurred.
I think that this is not kernel related but another component (systemd?)

I had a deeper look in journalctl and I found that the warnings are preceded every time by this ACPI Error related to the sensors AFAIK

Apr 09 21:46:56 hpprol2 kernel: ACPI Error: SMBus/IPMI/GenericSerialBus write requires Buffer of length 66, found length 32 (20141107/exfield-418)
Apr 09 21:46:56 hpprol2 kernel: ACPI Error: Method parse/execution failed [\_SB_.PMI0._PMM] (Node ffff8802d6cc4220), AE_AML_BUFFER_LIMIT (20141107/psparse-536)
Apr 09 21:46:56 hpprol2 kernel: ACPI Exception: AE_AML_BUFFER_LIMIT, Evaluating _PMM (20141107/power_meter-338)
Apr 09 21:46:56 hpprol2 systemd[2093]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/target0:0:0/0
Apr 09 21:46:56 hpprol2 systemd[1]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart1.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/target0
Apr 09 21:46:56 hpprol2 systemd[2093]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart1.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 09 21:46:56 hpprol2 systemd[1]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart4.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/target0
Apr 09 21:46:56 hpprol2 systemd[2093]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart4.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 09 21:46:56 hpprol2 systemd[2093]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart2.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 09 21:46:56 hpprol2 systemd[2093]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart3.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/targ
Apr 09 21:46:56 hpprol2 systemd[1]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart2.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/target0
Apr 09 21:46:56 hpprol2 systemd[1]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0\x2dpart3.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/target0
Apr 09 21:46:56 hpprol2 systemd[2093]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/target0:0:0/0
Apr 09 21:46:56 hpprol2 systemd[1]: Device dev-disk-by\x2did-scsi\x2dSHP_LOGICAL_VOLUME_0014380280B60D0.device appeared twice with different sysfs paths /sys/devices/pci0000:00/0000:00:02.2/0000:03:00.0/host0/target0:0:0/0:0:
Apr 09 21:46:56 hpprol2 kernel:  sdc: sdc1 sdc2 sdc3 sdc4


The last line is also related to the ACPI errors. 
So I disabled the service for lm-sensor but the warnings still occur.
I removed then packages openipmi and freeipmi but the warnings remain even after reboot

Regards
Philippe Condé
Comment 5 Philippe Condé 2015-05-17 07:11:40 UTC
Hello,

this problem remains with kerenl 4.0.1.1
The device by-uuid seems correct.
command blkid gives
hpprol2:~ # blkid 
/dev/sda1: UUID="ab7656ab-41e6-41e8-aa16-bbb4ca3c643e" TYPE="swap" PARTUUID="000aab75-01"
/dev/sda2: LABEL="OpenSuse" UUID="0d169d88-fe6d-4ca8-bebf-916cae9ec3d7" TYPE="ext4" PARTUUID="000aab75-02"
/dev/sda3: LABEL="usr" UUID="0e2de715-8f0b-4206-842c-2a1f4c4e5669" TYPE="ext4" PARTUUID="000aab75-03"
/dev/sda4: LABEL="home" UUID="dc7675a1-24a2-4ec2-a3a1-458af1b9d1e0" TYPE="ext4" PARTUUID="000aab75-04"
/dev/sdb1: LABEL="srv" UUID="bf2e6374-cee0-4c18-96b5-851f821ac807" TYPE="ext4" PARTUUID="0005f844-01"
/dev/sdb2: LABEL="var" UUID="b4849ee9-4a6c-4304-bc8d-cdca4f113408" TYPE="ext4" PARTUUID="0005f844-02"
/dev/sdb3: LABEL="local" UUID="89036c77-2a68-4ad2-9dfd-f950288600fa" TYPE="ext4" PARTUUID="0005f844-03"
/dev/sdb4: LABEL="opt" UUID="55cafd49-1996-4599-8af6-46fbaf61b5eb" TYPE="ext4" PARTUUID="0005f844-04"
/dev/sdc1: LABEL="vmroot" UUID="f14d0615-94ce-4f37-b491-4cefc8b83ff0" TYPE="ext4" PARTUUID="0001a1e7-01"
/dev/sdc2: LABEL="vmswap" UUID="e29476d3-e6a3-4083-ac21-12999fcb072c" TYPE="swap" PARTUUID="0001a1e7-02"
/dev/sdc3: LABEL="vmuser" UUID="506b37dd-4412-4398-abd1-970f223592a1" TYPE="ext4" PARTUUID="0001a1e7-03"
/dev/sdc4: LABEL="vmhome" UUID="0ddfdd63-b1a1-446b-8518-119273bdeff8" TYPE="ext4" PARTUUID="0001a1e7-04"

and 
hpprol2:~ # lsblk -f
NAME   FSTYPE LABEL    UUID                                 MOUNTPOINT
sda                                                         
├─sda1 swap            ab7656ab-41e6-41e8-aa16-bbb4ca3c643e [SWAP]
├─sda2 ext4   OpenSuse 0d169d88-fe6d-4ca8-bebf-916cae9ec3d7 /
├─sda3 ext4   usr      0e2de715-8f0b-4206-842c-2a1f4c4e5669 /usr
└─sda4 ext4   home     dc7675a1-24a2-4ec2-a3a1-458af1b9d1e0 /home
sdb                                                         
├─sdb1 ext4   srv      bf2e6374-cee0-4c18-96b5-851f821ac807 /srv
├─sdb2 ext4   var      b4849ee9-4a6c-4304-bc8d-cdca4f113408 /var
├─sdb3 ext4   local    89036c77-2a68-4ad2-9dfd-f950288600fa /local
└─sdb4 ext4   opt      55cafd49-1996-4599-8af6-46fbaf61b5eb /opt
sdc                                                         
├─sdc1 ext4   vmroot   f14d0615-94ce-4f37-b491-4cefc8b83ff0 
├─sdc2 swap   vmswap   e29476d3-e6a3-4083-ac21-12999fcb072c 
├─sdc3 ext4   vmuser   506b37dd-4412-4398-abd1-970f223592a1 
└─sdc4 ext4   vmhome   0ddfdd63-b1a1-446b-8518-119273bdeff8 

I think that the by-id generation uses some hardware information that is taken from the 4 disks  that  are in a raid 5 configuration and receive the same info for the 3 logical disks

Is it possible changing the default use of by-id by by_uuid? 
In Yast  ==> /etc/sysconfig  ==> system==> Yast2 ==> storage ==> DEVICE_NAMES I changed the "id" in "label" but this doesn't change anything for the errors in journalctl. Do I need doing manually some rebuild?
  
Regards
Philippe
Comment 6 Takashi Iwai 2015-05-28 12:58:19 UTC
Change the component, as it's apparently no kernel issue.
Comment 7 Bernhard Wiedemann 2015-05-29 04:40:54 UTC
Is any multipath involved? Check
ps ax|grep multipath

could be related to udev or be hw-specific?
Comment 8 Philippe Condé 2015-05-29 05:02:44 UTC
(In reply to Bernhard Wiedemann from comment #7)
> Is any multipath involved? Check
> ps ax|grep multipath
> 
> could be related to udev or be hw-specific?

No there is no multipath involved.
I think that the generation of the dev-ID uses some hardware information but the hardware raid 5 probably send the same information for the 3 logical disks that I defined.

Regards
Philippe
Comment 9 Thomas Blume 2015-05-29 06:15:54 UTC
(In reply to Philippe Condé from comment #8)
> (In reply to Bernhard Wiedemann from comment #7)
> > Is any multipath involved? Check
> > ps ax|grep multipath
> > 
> > could be related to udev or be hw-specific?
> 
> No there is no multipath involved.
> I think that the generation of the dev-ID uses some hardware information but
> the hardware raid 5 probably send the same information for the 3 logical
> disks that I defined.
> 

Can you please show the output of:

/lib/udev/scsi_id --export --whitelisted -d /dev/sda
/lib/udev/scsi_id --export --whitelisted -d /dev/sdb
Comment 10 Philippe Condé 2015-05-29 16:16:04 UTC
Hello, 

here the output for the 3 logical disks: As you can see the field ID_SCSI_SERIAL is the same for the 3 disks


hpprol2:~ # /lib/udev/scsi_id --export --whitelisted -d /dev/sda
ID_SCSI=1
ID_VENDOR=HP
ID_VENDOR_ENC=HP\x20\x20\x20\x20\x20\x20
ID_MODEL=LOGICAL_VOLUME
ID_MODEL_ENC=LOGICAL\x20VOLUME\x20\x20
ID_REVISION=4.68
ID_TYPE=disk
ID_SERIAL=3600508b1001c99233458581ffb65cc88
ID_SERIAL_COMPAT=
ID_SERIAL_SHORT=600508b1001c99233458581ffb65cc88
ID_WWN=0x600508b1001c9923
ID_WWN_VENDOR_EXTENSION=0x3458581ffb65cc88
ID_WWN_WITH_EXTENSION=0x600508b1001c99233458581ffb65cc88
ID_SCSI_SERIAL=0014380280B60D0 

hpprol2:~ # /lib/udev/scsi_id --export --whitelisted -d /dev/sdb
ID_SCSI=1
ID_VENDOR=HP
ID_VENDOR_ENC=HP\x20\x20\x20\x20\x20\x20
ID_MODEL=LOGICAL_VOLUME
ID_MODEL_ENC=LOGICAL\x20VOLUME\x20\x20
ID_REVISION=4.68
ID_TYPE=disk
ID_SERIAL=3600508b1001cd3e7527deb3b931a6639
ID_SERIAL_COMPAT=
ID_SERIAL_SHORT=600508b1001cd3e7527deb3b931a6639
ID_WWN=0x600508b1001cd3e7
ID_WWN_VENDOR_EXTENSION=0x527deb3b931a6639
ID_WWN_WITH_EXTENSION=0x600508b1001cd3e7527deb3b931a6639
ID_SCSI_SERIAL=0014380280B60D0
 
hpprol2:~ # /lib/udev/scsi_id --export --whitelisted -d /dev/sdc
ID_SCSI=1
ID_VENDOR=HP
ID_VENDOR_ENC=HP\x20\x20\x20\x20\x20\x20
ID_MODEL=LOGICAL_VOLUME
ID_MODEL_ENC=LOGICAL\x20VOLUME\x20\x20
ID_REVISION=4.68
ID_TYPE=disk
ID_SERIAL=3600508b1001c21d488c454a46bc39f22
ID_SERIAL_COMPAT=
ID_SERIAL_SHORT=600508b1001c21d488c454a46bc39f22
ID_WWN=0x600508b1001c21d4
ID_WWN_VENDOR_EXTENSION=0x88c454a46bc39f22
ID_WWN_WITH_EXTENSION=0x600508b1001c21d488c454a46bc39f22
ID_SCSI_SERIAL=0014380280B60D0 

Regards
Philippe
Comment 11 Thomas Blume 2015-06-01 06:54:40 UTC
Ok, thanks for the feedback.
So, the problem is ID_SCSI_SERIAL.
I'm just wondering how this could have changed with the kernel version.
WOuld it be possible to also send this output from a previous kernel that didn't show this bug?
Comment 12 Philippe Condé 2015-06-01 15:32:04 UTC
Hello,

I found the errors after some update of tumbleweed. I thought first that is was related to the new kernel delivered at this time (kernel 3.19). But I tested with a kernel 3.18 and the errors were also present: maybe this is not related to the kernel but to something else which is used by all kernels

As far as I remember this occurred end of march 2015

Regards
Philippe
Comment 13 Thomas Blume 2015-06-02 06:45:33 UTC
(In reply to Philippe Condé from comment #12)
> Hello,
> 
> I found the errors after some update of tumbleweed. I thought first that is
> was related to the new kernel delivered at this time (kernel 3.19). But I
> tested with a kernel 3.18 and the errors were also present: maybe this is
> not related to the kernel but to something else which is used by all kernels
> 
> As far as I remember this occurred end of march 2015
> 

Ok, maybe something has changed in the scsi_id code.
Robert, are there any patches for /lib/udev/scsi_id in recent tumbleweed versions that could explain this?
Comment 14 Robert Milasan 2015-06-02 06:50:52 UTC
No, I haven't touch scsi_id for tumbleweed nor there are many changes in scsi_id now a days.

Can you try sg_inq:

/usr/bin/sg_inq -p di --export /dev/sda
/usr/bin/sg_inq -p di --export /dev/sdb
/usr/bin/sg_inq -p di --export /dev/sdc

Adding Hannes to the bug, this seems more his 'area'.
Comment 15 Robert Milasan 2015-06-02 06:59:01 UTC
Sorry, better try this:

for disk in sda sdb sdc; do
     sg_inq /dev/$disk > disk_$disk.log 2>&1
done

This way we include, or we should unit_serial_number. Please attach the 3 file created.
Comment 16 Philippe Condé 2015-06-02 21:21:03 UTC
Created attachment 636509 [details]
file 1 sda
Comment 17 Philippe Condé 2015-06-02 21:21:47 UTC
Created attachment 636510 [details]
file 2 sdb
Comment 18 Philippe Condé 2015-06-02 21:22:17 UTC
Created attachment 636511 [details]
file 3 sdc
Comment 19 Philippe Condé 2015-06-02 21:24:27 UTC
Hello,

Here the result

hpprol2:~ # /usr/bin/sg_inq -p di --export /dev/sda
SCSI_IDENT_LUN_NAA=600508b1001c99233458581ffb65cc88
SCSI_IDENT_LUN_VENDOR=00000000

hpprol2:~ # /usr/bin/sg_inq -p di --export /dev/sdb
SCSI_IDENT_LUN_NAA=600508b1001cd3e7527deb3b931a6639
SCSI_IDENT_LUN_VENDOR=01000000

hpprol2:~ # /usr/bin/sg_inq -p di --export /dev/sdc
SCSI_IDENT_LUN_NAA=600508b1001c21d488c454a46bc39f22
SCSI_IDENT_LUN_VENDOR=02000000

3 files also attached
Regards
Philippe
Comment 20 Robert Milasan 2015-06-03 06:18:32 UTC
OK, so, from the results this is not udev nor scsi_id, but possibly kernel or hardware (not really sure).

Thomas, please check it yourself too and think of adding Hannes to the bug.
Comment 21 Thomas Blume 2015-06-03 08:02:05 UTC
(In reply to Robert Milasan from comment #20)
> OK, so, from the results this is not udev nor scsi_id, but possibly kernel
> or hardware (not really sure).
> 
> Thomas, please check it yourself too and think of adding Hannes to the bug.

Hm, the attachments show:

-->--
standard INQUIRY:
  PQual=0  Device_type=0  RMB=0  LU_CONG=0  version=0x05  [SPC-3]
  [AERC=0]  [TrmTsk=0]  NormACA=0  HiSUP=1  Resp_data_format=2
  SCCS=0  ACC=0  TPGS=0  3PC=0  Protect=0  [BQue=0]
  EncServ=0  MultiP=0  [MChngr=0]  [ACKREQQ=0]  Addr16=0
  [RelAdr=0]  WBus16=0  Sync=0  [Linked=0]  [TranDis=0]  CmdQue=1
    length=56 (0x38)   Peripheral device type: disk
 Vendor identification: HP      
 Product identification: LOGICAL VOLUME  
 Product revision level: 4.68
 Unit serial number: 0014380280B60D0 
--<--

Seems the standard inquiry gets codepage 0x80, e.g.:

# sg_inq -p 0x80 /dev/sdde
VPD INQUIRY: Unit serial number page
[...]


According to the scsi_id code:

        /*
         * Get page 0, the page of the pages. By default, try from best to
         * worst of supported pages: 0x83 then 0x80.
         */

this should only be done if no better codepage is available.

Philippe, can you please attach the output of:

/lib/udev/scsi_id --export --whitelisted -p 0x80 -d /dev/sda

/lib/udev/scsi_id --export --whitelisted -p 0x83 -d /dev/sda

/lib/udev/scsi_id --export --whitelisted -p pre-spc3-83 -d /dev/sda
Comment 22 Philippe Condé 2015-06-04 06:32:21 UTC
Hello,

Here the output asked. i executed the same command on the 3 disks
see file b926053_scsi_id.txt

as far as i see the next line is the same for the 3 disks
ID_SCSI_SERIAL=0014380280B60D0 

Regards 
Philippe
Comment 23 Philippe Condé 2015-06-04 06:33:42 UTC
Created attachment 636666 [details]
output of /lib/udev/scsi_id
Comment 24 Thomas Blume 2015-06-05 06:09:03 UTC
(In reply to Philippe Condé from comment #23)
> Created attachment 636666 [details]
> output of /lib/udev/scsi_id

Ok, scsi_id is correctly using page 0x83.
Seems that ID_SCSI_SERIAL is unusable for udev, at least in hardware raid setups.

Hannes, can we get a comment on this?
Comment 25 Robert Milasan 2015-06-05 06:16:00 UTC
Who's using and why is using ID_SCSI_SERIAL?
Comment 26 Thomas Blume 2015-06-05 06:42:53 UTC
(In reply to Robert Milasan from comment #25)
> Who's using and why is using ID_SCSI_SERIAL?

Ah, Robert, you are right, it isn't ID_SCSI_SERIAL.
It is ID_SERIAL_COMPAT from codepage 0x80.
See this rule, it calls explicitely page 0x80:

-->--
# scsi compat links for ATA devices
KERNEL=="sd*[!0-9]", ENV{ID_BUS}=="ata", PROGRAM="scsi_id --whitelisted --replace-whitespace -p0x80 -d $devnode", RESULT=="?*"
, ENV{ID_SCSI_COMPAT}="$result", SYMLINK+="disk/by-id/scsi-$env{ID_SCSI_COMPAT}"
KERNEL=="sd*[0-9]", ENV{ID_SCSI_COMPAT}=="?*", SYMLINK+="disk/by-id/scsi-$env{ID_SCSI_COMPAT}-part%n"
--<--
Comment 30 Hannes Reinecke 2015-06-05 07:24:08 UTC
But in any case: This is a firmware issue.
The 'hpsa' driver is responsible for retrieving page 0x80 from the HBA firmware.
As the driver doesn't modify the returned contents the firmware returns incorrect information.

So it's more an issue for HP.
Comment 33 Thomas Blume 2015-06-05 08:06:43 UTC
Ok, thanks Robert, Hannes.

Philippe, your raid controllers firmware does soemthing wrong.
Either it wrongly reports the bus to be ata (ENV{ID_BUS}=="ata") or it reports a wrong ID_SERIAL for your disks.
Can you please check wheter there is a firmware update for your raid controller available?
Comment 36 Philippe Condé 2015-06-05 14:24:17 UTC
Hello,

the system is a HP proliant ML350p gen8 with a hardware raid controller p420i
I checked on HP site for this system and opensuse:

The only firmware controller is at version 1.28 (opensuse 11.4)
But my current firmware is at version 4.68.


for SLES 12 I find a more recent firmware (6.34) but the fixes are related to raid6 and other system (blade systems)

see
http://h20565.www2.hp.com/hpsc/swd/public/detail?sp4ts.oid=5195931&swItemId=MTX_13fcbd530be24670b72b5c96fe&swEnvOid=4175#tab5 

If you think that this should solve the problem I can try to install this version but I'm not sure that SLES 12 version is compatible with Opensuse Tumbleweed 

Regards
Philippe
Comment 37 Philippe Condé 2015-06-05 15:03:54 UTC
Created attachment 636906 [details]
KInfoCenter

Hello

See attached kinfo.txt
This is the output from KInfoCenter: I find this for the device information ==> scsi

I don't see any sata?

Regards
Philippe
Comment 38 Thomas Blume 2015-06-08 07:07:53 UTC
(In reply to Philippe Condé from comment #37)
> Created attachment 636906 [details]
> KInfoCenter
> 
> Hello
> 
> See attached kinfo.txt
> This is the output from KInfoCenter: I find this for the device information
> ==> scsi
> 
> I don't see any sata?

It depends what udev sees.
Can you please provide the ouptut of:

udevadm info /dev/sda
udevadm info /dev/sdb
udevadm info /dev/sdc
Comment 39 Philippe Condé 2015-06-08 19:03:08 UTC
Created attachment 637074 [details]
udevadm output

Hello,

Here the output.

Remark that I have two sata devices 
- 1 dvdrom
- 1 dvdram

But udevadm gives scsi as far as I see for sda, sdb and sdc
Regards
Philippe
Comment 42 Thomas Blume 2015-06-09 07:50:55 UTC
(In reply to Philippe Condé from comment #39)
> Created attachment 637074 [details]
> udevadm output
> 
> Hello,
> 
> Here the output.
> 
> Remark that I have two sata devices 
> - 1 dvdrom
> - 1 dvdram
> 
> But udevadm gives scsi as far as I see for sda, sdb and sdc
> Regards
> Philippe

Ok, thanks, now it's clear.
It isn't a wrong identification of your disks as ata devices.
It is a wrong value in codepage 0x80.

/lib/udev/rules.d/58-scsi-sg3_symlink.rules

is using SCSI_IDENT_SERIAL, which takes the value of 'Unit serial number' from codepage 0x80.

The value of 'Unit serial number' must be uniqe for a disk.
I your case, it is not.
I'd recommend to report this as a firmware bug to HP.

As an interim solution, you might just comment out the problematic rule, so that the symlinks don't get created (they are broken anyway since they can point to any of your scsi disks).
Comment 43 Philippe Condé 2015-06-09 09:16:46 UTC
Hello,

I checked in /etc/udev/rules.d but I don't have the 58-scsi-sg3_symlink.rules

I find only
40-xen.rules
55-iscan.rules
55-libsane.rules
56-sane-backends-autoconfig.rules
60-libpisock.rules
70-persistent-net.rules

I'll try to open an incident by HP

Regards
Philippe
Comment 45 Philippe Condé 2015-06-13 06:39:09 UTC
Hello,

I updated the firmware op the raid controller to the last version (6.34) but this doesn't help.

I have seen that I didn't read carefully your last comment and was not looking in the correct library.

I have then changed the file /lib/udev/rules.d/58-scsi-sgr_symlink.rules 
I commented first the line
ENV{SCSI_IDENT_SERIAL}=="?*", ENV{DEVTYPE}=="disk", SYMLINK+="disk/by-id/scsi-S$env{SCSI_VENDOR}_$env{SCSI_MODEL}_$env{SCSI_IDENT_SERIAL}"
 and rebooted the system but some error messages were still present  

I commented then also the next line:
ENV{SCSI_IDENT_SERIAL}=="?*", ENV{DEVTYPE}=="partition", SYMLINK+="disk/by-id/scsi-S$env{SCSI_VENDOR}_$env{SCSI_MODEL}_$env{SCSI_IDENT_SERIAL}-part%n"

After reboot the error messages have disappear. 

One question: are there some drawback with these settings?

Many thanks in advance
Philippe
Comment 46 Thomas Blume 2015-06-15 09:10:28 UTC
(In reply to Philippe Condé from comment #45)
> Hello,
> 
> I updated the firmware op the raid controller to the last version (6.34) but
> this doesn't help.
> 
> I have seen that I didn't read carefully your last comment and was not
> looking in the correct library.
> 
> I have then changed the file /lib/udev/rules.d/58-scsi-sgr_symlink.rules 
> I commented first the line
> ENV{SCSI_IDENT_SERIAL}=="?*", ENV{DEVTYPE}=="disk",
> SYMLINK+="disk/by-id/scsi-
> S$env{SCSI_VENDOR}_$env{SCSI_MODEL}_$env{SCSI_IDENT_SERIAL}"
>  and rebooted the system but some error messages were still present  
> 
> I commented then also the next line:
> ENV{SCSI_IDENT_SERIAL}=="?*", ENV{DEVTYPE}=="partition",
> SYMLINK+="disk/by-id/scsi-
> S$env{SCSI_VENDOR}_$env{SCSI_MODEL}_$env{SCSI_IDENT_SERIAL}-part%n"
> 
> After reboot the error messages have disappear. 

Ok
 
> One question: are there some drawback with these settings?
> 

It would only be problematic if something would be using these symlinks (for example mounts or lvm).
But per default they are not used, so you can savely omit them.
Comment 47 Thomas Blume 2015-07-13 12:27:56 UTC
Since the open questions are clarified, I'm closing the bug.