Trying to use USB as a backup media - supported? and writlck

[Jenwae]

Hi,

I'm running OopenVMS 8.3-1H1 on rx3600. The machine has an USB port on the front.
I'm trying to use the USB media (memory stick/flash drive) to backup data to, replacing the conventional DDS tape.

Is this possible, and supported?

When I plug in the USB stick, it is recognized by the system. I did a $Show Dev, and see the following:

$8$DNA1: (RND1) Online wrtlck

I did a $show dev $8$dna1/full, and it says:

Disk $8$DNA1: (RND1), device type USB Flash Memory, is online, file-oriented device, shareable, served to cluster via MSCP Server, error logging is enabled.

When I try to initialize the flash drive, I encountered error:

$INIT $8$DNA1: backup

%INIT-F-WRITLCK, write lock error

Can anyone help please?

Thank you.

[Hoff · Accepted solution]

AFAIK, there's no support for arbitrary USB devices.

The drive was reported as write-locked, so the error reported by BACKUP is to be expected.

What does USM tool report about the device? (The system manager's utilities reference manual, IIRC, and there's a section there on USM and USB details that's worth a read.)

Try a different USB flash drive; different vendor and different model. I've hit various differences with different flash drives. Some work. Some don't.

Check that the EFI firmware level is current for this box too, as there have been various EFI updates around USB support.

Make sure you don't get tangled with the CD/DVD drive or a $8$DNA1: device elsewhere in the cluster, too.

[Steven Schweda]

> Try a different USB flash drive [...]

For example, on my zx2000:

IT $ show devi /full dna0:

Disk IT$DNA0:, device type USB 2.0 Flash Disk , is online, file-oriented
device, shareable, available to cluster, error logging is enabled.

Error count 0 Operations completed 0
Owner process "" Owner UIC [SYSTEM]
Owner process ID 00000000 Dev Prot S:RWPL,O:RWPL,G:R,W
Reference count 0 Default buffer size 512

IT $ show logi mtools_u
"MTOOLS_U" = "DNA0:" (LNM$PROCESS_TABLE)

IT $ mtools -c mdir u:
MTOOLS_U (_IT$DNA0:) mounted.
Volume in drive U has no label
Volume Serial Number is 176C-7C90
Directory for U:/

No files
2 025 017 344 bytes free

_IT$DNA0: dismounted.

IT $ mtools -c mcopy sylogin.com u:
MTOOLS_U (_IT$DNA0:) mounted.
_IT$DNA0: dismounted.

IT $ mtools -c mdir u:
MTOOLS_U (_IT$DNA0:) mounted.
Volume in drive U has no label
Volume Serial Number is 176C-7C90
Directory for U:/

sylogin com 6003 2009-09-22 16:04 sylogin.com
1 file 6 003 bytes
2 025 009 152 bytes free

_IT$DNA0: dismounted.
IT $

Or, if you're anti-FAT:

IT $ init DNA0: f2gb

IT $ mount DNA0: f2gb
%MOUNT-I-MOUNTED, F2GB mounted on _IT$DNA0:

IT $ dire /date /prot /size DNA0:[000000]

Directory DNA0:[000000]

000000.DIR;1 1 22-SEP-2009 16:08:16.33 (RWED,RWED,RE,E)
BACKUP.SYS;1 0 22-SEP-2009 16:08:16.33 (RWED,RWED,RE,)
BADBLK.SYS;1 0 22-SEP-2009 16:08:16.33 (RWED,RWED,RE,)
BADLOG.SYS;1 0 22-SEP-2009 16:08:16.33 (RWED,RWED,RE,)
BITMAP.SYS;1 62 22-SEP-2009 16:08:16.33 (RWED,RWED,RE,)
CONTIN.SYS;1 0 22-SEP-2009 16:08:16.33 (RWED,RWED,RE,)
CORIMG.SYS;1 0 22-SEP-2009 16:08:16.33 (RWED,RWED,RE,)
GPT.SYS;1 96 22-SEP-2009 16:08:16.33 (RWED,RWED,RE,E)
INDEXF.SYS;1 104 22-SEP-2009 16:08:16.33 (RWED,RWED,RE,)
SECURITY.SYS;1 1 22-SEP-2009 16:08:16.33 (RWED,RWED,RE,)
VOLSET.SYS;1 0 22-SEP-2009 16:08:16.33 (RWED,RWED,RE,)

Total of 11 files, 264 blocks.

That's a 2GB USB flash EEPROM gizmo from
MicroCenter. A 1GB unit from the same place
also worked. I haven't tried anything
bigger, but what could go wrong?

[Jenwae]

Hi Steven and Hoff,

Thank you for your excellent guidance. I went out to buy a HP pendrive, and plugged it in, and it works. I could initialize, mount and copy data files to the new USB media.

I guess VMS has not consistent support of USB drives from different vendors, like you said.

The issue I have now is that the device entries now remain even though I have unplugged the flash drive from the USB port, that is, when I do a $show dev d, the flash drive entries can still be seen.

$8$DNA1: (RND1) Online wrtlck 0
$8$DNA2: (RND1) Online 0

DNA1 is the original flash drive that could not be initialized and mounted.
DNA2 is the one I got from HP.

I'm concerned that the next time I plug in the HP flash drive, a new entry will be created, and cannot be removed from the system.

Do you by any chance know how to resolve this?

Thank you.

[Hoff]

Concern? Fear? Um, OK. Any proof?

(Worst case, the box leaks the pool memory for a unit control block (UCB) and such, and the UCB and related data structures are small allocations in the grand scheme. Given the memory on a typical Integrity server, you can leak vast numbers of UCBs before you get in trouble with the amount of pool. Look at the numbers of MB and BG devices, for instance.)

And as for your concern, please go try it. Prove there is a problem. See if you can show a "device leak", or see if the existing device gets re-used.

Then if you do get a "device leak" and if the concern warrants the extra effort, then use the USM tool that was mentioned earlier to look to clean up the devices. And please also make a formal report to HP, if you can show a "device leak"; a USB USM device management bug.

The USM stuff is available in the system management utilities manual: http://www.hp.com/go/openvms/doc

Please do not assume that USB devices are the same. That all CD and DVD devices and SCSI devices are identical. They are not. Best case, random devices are mostly compatible. But they're seldom if anything closer than mostly-compatible. And if the device does not work, then swap the device for a different vendor and different model.

[Steven Schweda]

> I'm concerned [...]

> And as for your concern, please go try it.

Exactly. How much does it cost to run the
experiment?

Around here, for example, I plugged my 2GB
gizmo into one hole in a four-port hub, then
unplugged it, and then plugged it into
another hole in the same hub:

UCM> show event /since = today
Date Time Type Priority Component
--------------------------------------------------------------------------------
23-SEP-2009 22:27:46.69 DRIVER NORMAL HUBDRIVER
Message: Find a driver for DeviceClass/DeviceSubClass = 0x0/0x0

23-SEP-2009 22:27:46.69 UNKNOWN NORMAL UCM UNKNOWN DEVICE
Message: VENDOR_ID = 2316
PRODUCT_ID = 4096
RELEASE_NUMBER = 4352
BUS_NUMBER = 1
PATH = 2.4.0.0.0.0
DEVICE_CLASS = 0
DEVICE_SUB_CLASS = 0
DEVICE_PROTOCOL = 0
NUMBER_OF_INTERFACES = 1
NUMBER_OF_CONFIGURATIONS = 1
SERIAL_NUMBER = A400000000022251
MANUFACTURER_STRING = USB 2.0
PRODUCT_STRING = Flash Disk
CONFIGURATION_NUMBER = 0.

23-SEP-2009 22:27:46.69 DRIVER NORMAL HUBDRIVER
Message: Find a driver for InterfaceClass/InterfaceSubClass/Protocol = 0x8/0x6/0x50

23-SEP-2009 22:27:46.77 UCM CRITICAL ucm_config_request
Message: %SYSTEM-W-DEVEXISTS, device unit already exists

23-SEP-2009 22:27:46.78 DRIVER NORMAL HUBDRIVER
Message: Configured device DNA4 using driver SYS$DNDRIVER:

23-SEP-2009 22:28:36.57 DECONFIGURED NORMAL HUBDRIVER
Message: Deconfiguring device on bus 1 at port 4 bus tier 2 usb address
4

23-SEP-2009 22:28:57.47 DRIVER NORMAL HUBDRIVER
Message: Find a driver for DeviceClass/DeviceSubClass = 0x0/0x0

23-SEP-2009 22:28:57.47 UNKNOWN NORMAL UCM UNKNOWN DEVICE
Message: VENDOR_ID = 2316
PRODUCT_ID = 4096
RELEASE_NUMBER = 4352
BUS_NUMBER = 1
PATH = 2.2.0.0.0.0
DEVICE_CLASS = 0
DEVICE_SUB_CLASS = 0
DEVICE_PROTOCOL = 0
NUMBER_OF_INTERFACES = 1
NUMBER_OF_CONFIGURATIONS = 1
SERIAL_NUMBER = A400000000022251
MANUFACTURER_STRING = USB 2.0
PRODUCT_STRING = Flash Disk
CONFIGURATION_NUMBER = 0.

23-SEP-2009 22:28:57.47 DRIVER NORMAL HUBDRIVER
Message: Find a driver for InterfaceClass/InterfaceSubClass/Protocol = 0x8/0x6/0x50


23-SEP-2009 22:28:57.55 UCM CRITICAL ucm_config_request
Message: %SYSTEM-W-DEVEXISTS, device unit already exists

23-SEP-2009 22:28:57.56 DRIVER NORMAL HUBDRIVER
Message: Configured device DNA4 using driver SYS$DNDRIVER:

UCM>


The first time, "PATH = 2.4.0.0.0.0". The
second time, "PATH = 2.2.0.0.0.0". (It's
similar in the other two holes, "PATH =
2.1.0.0.0.0" and "PATH = 2.3.0.0.0.0".) It
keeps coming up as "device DNA4". Perhaps it
takes that "SERIAL_NUMBER" stuff seriously.

And that's with "Universal Serial Bus
Configuration Manager, Version V1.0" on VMS
V7.3-2 (Alpha). It could only be better now.

[Ian Miller.]

Steven, interesting, which USB card are you using on your alphaserver V7.3-2 system?

[Steven Schweda]

> [...] which USB card are you using on your
> alphaserver V7.3-2 system?

It's some cheap ($10) Chinese thing with an
NEC µPD720100AGM chip on it. I have this
entry in my
SYS$COMMON:[SYSEXE]SYS$USER_CONFIG.DAT:

! Added on 07-JUL-2006 17:39:14 manually by SMS.
!
device = "NEC OHCI USB Controller"
name = OH
driver = SYS$OHCIDRIVER
adapter = PCI
id = 0x00351033
boot_class = DN
boot_flags = HW_CTRL_LTR, UNIT_0
flags = PORT, BOOT
!!! ASSOC = UCM, USB
end_device

It looks like what's in SYS$CONFIG.DAT on a
V8.3 system, except for that one line being
commented out. (The facts are lost in
antiquity, but I assume that the older OS
complained with that line left in.)

At least some of this was discussed on
comp.os.vms when it was news.

When shopping for cheap Chinese USB cards
with an NEC chip, look for ones with the
3.3V-OK slot in the edge connector, or else
you won't be able to use it in a newer
machine with no 5V PCI slots.

My newer cards with the µPD720101F1 (newer
chip, smaller package) seem to work, too,
so anything with a "D720100A" or "D720101"
in the number should be ok.

This one's in an XP1000, by the way. I
believe that I may recall some discussion of
support (or at least functionality) having
been added at some time for the built-in USB
in an XP1000, but by that time I had a
boat-load of these NEC-chip cards spread all
around, so I never needed to try using the
built-in USB.

[Hoff]

Configuring and using the (entirely unsupported) USB connections present on the AlphaStation XP1000 box on OpenVMS Alpha V8.3:

http://labs.hoffmanlabs.com/node/1410