10-28-2011 11:41 AM
I have been using 1TB USB hard drives for system backups on an HP rx2600 (1.40GHz/1.5MB) running VMS 8.3-1H1.
This has been working fine for years. VMS Backup creates the savesets and verifies them - no problems.
The saveset file sizes tend to run up to between 50,000,000 and 64,000,000 blocks.
I've recently tried to copy some of the savesets to another drive and find the DNA device hangs as "Device Busy".
The "device busy" hang happens every time I try to copy the backup saveset using either BACKUP or COPY commands.
This same hang has happened on several different USB drives, so I don't think it is the drives themselves causing the problem.
The DNA device error count remains at 0.
When this device hang happens, I have to reboot to handle it, but even the system shutdown fails to complete, requiring power-cycling to get the system back up again.
Recently, I tried doing a restore from one of the backups and again got one of these "device busy" hangs. (Luckily, I had already recovered the files I needed before the hang occurred.)
Does anyone have any suggestions regarding this "device busy" situation?
10-28-2011 10:08 PM
if this happens again, force a system crashdump and log a call with the HP Support Center.
'Device busy' typically means, that an IO operation had been started at the device driver level and did not finish, either successfully or by returning some kind of error.
10-31-2011 12:45 PM
Thanks for your reply.
I don't have an HP support contract so I don't think I can log a call with them.
I put the USB external HD in question onto a Windows system and ran Seagate diagnostics on it and found it has 4 bad blocks, all in one relatively small area about 1/2-way in on the drive (at least in terms of logical blocks).
The funny part is, the bad blocks don't appear to be in the file that I was trying to copy when the IO hung as busy.
Perhaps the DNA driver in VMS 8.3-1H1 doesn't have the necessary code to handle bad block processing?
It appears the DN driver for VMS 8.4 is newer. I'll try accessing the USB drive on VMS 8.4 and see if I get any different resulsts.
10-31-2011 02:01 PM
you can do IO tracing using LDDRIVER, see
This way you can probably find out, which IO at which LBN causes the problem.
11-05-2011 08:20 AM
My experience with a zx6000 (essentially the same as an rx2600) system is that USB is working worse with OpenVMS I64 V8.4 than with V8.3-1H1.
Of course, this may not be the case for your specific problem.
I would recommend to install all available patches, but I suppose this might be a problem for you if you have no support contract.
11-11-2011 06:17 AM
Thanks for your suggestions.
For this situation, VMS 8.4 seems to handle things a bit better.
Instead of the DNA device just hanging as busy (requiring a reboot), VMS 8.4 reports a fatal drive error and cleanly exits back to DCL.
The strange thing is that I put the same external USB drive on an AlphaServer 1200 that's had a USB card installed and although it has a much-slower transfer rate, it appears to read the file without ANY errors.
Perhaps there's something up with the USB implementation on VMS-IA64 that's handled better on Alpha?
I'm going to test this a bit more to see if these results are consistent, not just a fluke.