cancel
Showing results for 
Search instead for 
Did you mean: 

dfu abort

SOLVED
Go to solution
Jack Trachtman
Super Advisor

dfu abort

dfu v3.1-1
2-node VMS V7.3-2 cluster

DFU aborts as seen in the attachment.

Anything I can do about this?

Thanks
21 REPLIES
Thomas Ritter
Respected Contributor

Re: dfu abort

We run DFU 3.0 all the time under vms 7.3-2. Looks like a bug has been found. How long has v3.1-1 been running ?



Karl Rohwedder
Honored Contributor

Re: dfu abort

Just checked with DFU V3.1-1 on V7.3-2 on 2 different disks, runs fine, but these were no shadowed disks.

regards Kalle
Kris Clippeleyr
Honored Contributor

Re: dfu abort

Hi,
Verified DFU 3.1-1 on VMS 7.3-1 and 7.3-2 with shadowed disks. All ok here.
Maybe Ton Dorland can shed some light?
Kris (aka Qkcl)
I'm gonna hit the highway like a battering ram on a silver-black phantom bike...
Jack Trachtman
Super Advisor

Re: dfu abort

I'm finding more disks on this cluster where DFU fails.

Is there someone I can send info to for a
review, understanding that DFU is not officially supported?
Jack Trachtman
Super Advisor

Re: dfu abort

Looks like Ton Dorland is the author.

Does anyone know if he still supports DFU
and how to contact him?
Thomas Ritter
Respected Contributor

Re: dfu abort

Jack, for what its worth... we have had system crashes with DFU when using the undelete function. And the crashes occured days after the file was undeleted. The HP advice was not to use DFU for undeletes.
Thomas Ritter
Respected Contributor

Re: dfu abort

Jack, maybe worth trying the dfu verify with some qualifiers.

Maybe $ dfu /verify/lock

Also would you also check the disk using

$ ANALYZE/DISK/SHADOW DSA102:

and try

$ dfu verify DSA102: in place of the logical name.



Jur van der Burg
Respected Contributor

Re: dfu abort

I think this is a bug in dfu when it tries to access a longword at the last byte of a page (look at the failng address: 00BBC003). I vaguely collect that it is a compiler issue (I did the research a long time ago). Ton would know for sure, and I'm also sure that it's been fixed in the latest version.

Jur.
David B Sneddon
Honored Contributor

Re: dfu abort

Jur,

I believe V3.1-1 is the latest...

Dave
Ian Miller.
Honored Contributor

Re: dfu abort

I think DFU is not officially supported. I wish it was and was included on every VMS system.
You may be able to report the bug via support channels but HP don't have to fix it.

Hopefully they will.
____________________
Purely Personal Opinion
Jack Trachtman
Super Advisor

Re: dfu abort

I had upgraded DFU from v2.7a to v3.1-1 this week and was reviewing its functionality. The disk VERIFY option seems to find more (minor) things wrong with directories than ANALYZE/DISK does, which I found interesting, so I put DFU in a script that would run it against all disks on a system, and copied that script to all our VMS hosts to try it. That's when I started seeing it abort on some disks (I've run across at least 4 disks now that cause it to abort).

Tom: I've tried your suggestions, but I see no change. I've also played with quota settings because the problem seems to occur on "large" disks.

Jur: Your analysis may be the clue needed.

My problem now is I don't have a contact to notify of this problem.
Ian Miller.
Honored Contributor

Re: dfu abort

I found with DFU VERIFY that specifying the /DIR qualfier gave more consistant results.
Also using /LOCK helps if you can.
____________________
Purely Personal Opinion
Jur van der Burg
Respected Contributor

Re: dfu abort

Folks, I know there's a fix in DFU V3.1-3 as that's what I'm running now for some time. When Ton Dorland left HP he handed me the sources where this was fixed. We have been discussing about integrating it in VMS when I was a member of VMS engineering, and I did the work to make that possible, but higher management was against the idea.

I do have an Alpha kit for that version, and I will see that I can make it available one of the next days.

I'm not supporting DFU, but I know Ton personally and will ask him about any new work he may have done in the past since I acquired the sources. Don't ask HP for support, it's freeware after all.

Jur
lddriver at digiater dot nl
TonDorland
Occasional Visitor

Re: dfu abort

Folks,

these issues have been solved in V3.1-3. That is the last published version.
My apologies for replying this late, but I am no longer working in the VMS business, so I do not actively folow OpenVMS threads.

Regards,

Ton Dorland
Jess Goodman
Esteemed Contributor

Re: dfu abort

Jur,

I reported to Tom by email two other bugs in V3.1-1. I'm hoping you can confirm that they are fixed in V3.1-3. I know Tom had found the problems and was working on fixing them for the next version.

One bug was that when you do a DFU SEARCH with /CREATE=SINCE= (or any other way of specifying a date-time limit) you may get files that were created up to 7 minutes before the given time. Probably this is due to not properly comparing the low-order 32 bits of a 64 bit binary datetime value.

The second bug I reported was with DFU UNDELETE (without /LIST) not finding all files it should. Tom eventually tracked this down any replied to me (quoted below so you can ask him about it):

"Well I found the bug. It had something to do with the command-line processing which could cause the result from parsing something like *.COM into a string which was not terminated correctly. Therefore my routine which compares a deleted file name with the *.COM would say 'no match'. Due to the coding the problem will not show up if you either use /LIST or if you use *.COM without semicolon."
I have one, but it's personal.
Jur van der Burg
Respected Contributor

Re: dfu abort

These bugs are fixed in the latest sources that I just got from Ton. I'm going to rebuild and repackage dfu, when it's ready I'll announce it here.

Jur.
Dave Gudewicz
Valued Contributor

Re: dfu abort

Many thanks Jur and Ton. The OVMS community appreciates your effort and help with DFU.

Dave...
Jur van der Burg
Respected Contributor
Solution

Re: dfu abort

Ok, DFU V3.2 is ready to go.

http://www.digiater.nl/dfu

Jur.
Steven Schweda
Honored Contributor

Re: dfu abort

> http://www.digiater.nl/dfu

The Web page (HTML) title needs some work.

Minor feature request:

DFU /VERSION (or similar).
Jur van der Burg
Respected Contributor

Re: dfu abort

>The Web page (HTML) title needs some work.

Fixed.

>Minor feature request:
>DFU /VERSION (or similar).

Well, I'm not gonna change that, it announces itself already with the version. I'm already very busy with LDdriver, and dfu just interrupted that work.

Jur.
Jack Trachtman
Super Advisor

Re: dfu abort

I have installed DFU V3.2 on all my VMS nodes and now see *no* problems with the DFU VERIFY/DIR cmd!

Thanks all!