dmesg error:partially open device

[mark_238]

hi, this is what i get from dmesg.

vmunix: SCSI: Attempt to access partially open device -- dev: bc032000

hpux B.11.00

[Maz_2]

Hi,
I found the following for you....
Hope this solves your problem.

Periodically on some systems you will see messages like the following appearing
in /var/adm/syslog/syslog.log:

vmunix: SCSI: Attempt to access partially open device --
dev: bc094000
vmunix: SCSI: Attempt to access partially open device --
dev: bc094000
vmunix: SCSI: Attempt to access partially open device --
dev: bc096000

RESOLUTION
The kernel SCSI driver when it opens a device sends the INQUIRY,
START_STOP_UNIT, TEST_UNIT_READY and READ_CAPACITY SCSI commands to start
the device and get the characteristics and the capacity of the device.

If any of the above SCSI commands fail and the device is a raw device the
kernel will print a message like the ones above. This is done so commands
like diskinfo and mediainit can still open devices even though they may not
be ready.

An example would be a CDROM drive without a CDROM in it. For the diskinfo
command some of the SCSI commands will fail because there is no media (e.g.
READ_CAPACITY cannot return the amount of data on the CDROM if there is no
CDROM in the drive) but the INQUIRY command will work (i.e. even if the
CDROM does not have disk in it it will still reply with information to an
INQUIRY command). The mount command will not cause these errors because
the kernel mounts the CDROM, no programs attempt to open the device file.

Disk arrays can also create these messages. Disk arrays can have luns created
that have no device associated with them. See Document A5469587 for
information about this message and EMC BCVs, other disk arrays can also
cause these messages to be seen.

The attempt to access partially open device messages typically do not
indicate any problems that need attention, unless the device being complained
about SHOULD have a real device associated with it. In that case since
the disk array is not reporting some of the information needed by HPUX
the problem would most likely reside in the array.

[T G Manikandan]

//*

What do partially open device SCSI messages in syslog.log mean? DocId: AUSRCKBRC00000687 Updated: 4/13/00 9:19:43 PM

PROBLEM
Periodically on some systems you will see messages like the following appearing
in /var/adm/syslog/syslog.log:

Jan 11 16:06:01 hp1 vmunix: SCSI: Attempt to access partially open device --
dev: bc094000
Jan 11 16:06:01 hp1 vmunix: SCSI: Attempt to access partially open device --
dev: bc094000
Jan 11 16:06:01 hp1 vmunix: SCSI: Attempt to access partially open device --
dev: bc096000
RESOLUTION
The kernel SCSI driver when it opens a device sends the INQUIRY,
START_STOP_UNIT, TEST_UNIT_READY and READ_CAPACITY SCSI commands to start
the device and get the characteristics and the capacity of the device.

If any of the above SCSI commands fail and the device is a raw device the
kernel will print a message like the ones above. This is done so commands
like diskinfo and mediainit can still open devices even though they may not
be ready.

An example would be a CDROM drive without a CDROM in it. For the diskinfo
command some of the SCSI commands will fail because there is no media (e.g.
READ_CAPACITY cannot return the amount of data on the CDROM if there is no
CDROM in the drive) but the INQUIRY command will work (i.e. even if the
CDROM does not have disk in it it will still reply with information to an
INQUIRY command). The mount command will not cause these errors because
the kernel mounts the CDROM, no programs attempt to open the device file.

Disk arrays can also create these messages. Disk arrays can have luns created
that have no device associated with them. See Document A5469587 for
information about this message and EMC BCVs, other disk arrays can also
cause these messages to be seen.

The attempt to access partially open device messages typically do not
indicate any problems that need attention, unless the device being complained
about SHOULD have a real device associated with it. In that case since
the disk array is not reporting some of the information needed by HPUX
the problem would most likely reside in the array.

*//

[Robert-Jan Goossens]

Hi,

take a look at next link,

Europe;
http://www4.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&admit=-938907319+1048585440427+28353475&docId=200000064900353

U.S.
http://support2.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000064900353

Robert-Jan.

[T G Manikandan]

Also,
bc from lsdev it indicates a disk

bc032000

Then it is c3t2d0 which is a disk which is sending those SCSI status messages

[Sanjay_6]

Hi,

Have you tried the scsi subsystem patch.

Here is the latest one for 11.0


http://www1.itrc.hp.com/service/patch/patchDetail.do?patchid=PHKL_28496&context=hpux:800:11:00

Hope this helps.

regds