VxVM x LVM

[Marcia Ferreira_2]

The SO was reinstalled and VxVM was not installed. Can anybody tell how we remove a Veritas logical volume disk from SO, (without installing VXVM) and make this disk available to LVM ???

Thanks

[James R. Ferguson]

Hi:

I believe that to "release" a VxVM disk from its control, that you are going to have to install (at least) the base VxVM product.

Regards!

...JRF...

[Hai Nguyen_1 · Accepted solution]

Marcia,

If you do not care about the data on the disk, you can reformat for new use.

Hai

[Marcia Ferreira_2]

OK Hai.

I used the command "dd", and it solved the problem.

Thanks to everybody

MF

[James R. Ferguson]

Hi Marcia:

While I am not very familar with VxVM, the brute force 'dd' of an LVM disk would not only remove the LVM information but leave critical configuration files (notably '/etc/lvmtab') in an inconsistent state. In my opinion, you may have done yourself a dis-service.

Regards!

...JRF...

[Marcia Ferreira_2]

Hi James

I agree with you, but in this case, /etc/lvmtab hasn??t any pointer to these disks, and "dd" ended the VxVM pointer.
Thanks
MF

[Hai Nguyen_1]

Marcia,

To be safe, I recommend using mediainit (man 1m mediainit for more info) later on.

Hai

[Marcia Ferreira_2]

OK ..

[Frank Slootweg]

Please do *not* use mediainit. mediainit tries to do a low-level format and may make the disk less reliable, but you will not know, one way or the other.

[Stefan Farrelly]

Frank - are you still spreading rumours not to use mediainit ? I thought we'd been here before.

HP-UX 11i also comes with mediainit included as part of the OS-Core.SYS-ADMIN fileset, seeing as 11i is very new why on earth would HP include it if it wasnt recommended to use it ?

From the man page for mediainit on 11i it says;

"Most HP hard disks are formatted and exhaustively tested at the factory by use of a process more thorough but also more time-consuming than appropriate for mediainit. However, mediainit is still valuable for ensuring the integrity of the media after factory shipment, formatting with the correct interleave factor, and sparing any blocks which may have become defective since original factory testing was performed."

[Trond Haugen]

James, Marcia; I belive had this been a LVM disk that was in lvmtab you would be in trouble, but VxVM does not have a lvmtab equivalent. So there should be no problems.
Also I would have thought that a 'pvcreate -f' would have done the trick.
Also remember to remove any /dev/vx/.. files for the disk.

Regards,
Trond

[Frank Slootweg]

Stefan, I am not spreading 'rumours', but giving *facts*. Fact is that mediainit *does* try to do a low-level format and a LLF may make the disk less reliable. So doing a mediainit, especially when it is *not needed*, is a bad idea.

As to why 'not-recommended' stuff is still included in HP-UX:

mediainit is *hardly* ever needed, but there *are* some very *exceptional* uses, *not* for *hard* disks, which is why HP still supplies it. This very exceptional use is: partitioning DDS tapes ('never' used) and low-level formatting diskettes (nowadays most diskettes are LLFed during production). (Some people also use it to be able to use disks which have the wrong sector size. For what it is worth, I consider that also inappropriate use, for the same reason as given above.)

As to your quote from the manual page: Yes, I know what it says. However also note that it talks quite a lot about CS/80 and SS/80 which are protocols which are obsolete for some 15+ years, i.e. this is a very old manual page, which has not been adapted to current needs.

I think this problem can only be solved by user education, which is not easy, if not 'impossible' to do in a manual page. If it was easy, we wouldn't be having this discussion.

In case you want to contact me 'off-line': frank_slootweg@hp.com

[Stefan Farrelly]

Frank - if there was a possible problem using mediainit on disks then HP wouldnt have shipped it with 11i - or they would have changed it, or its manpage, to make clear any warnings, or prevent it running on newer disks. None of this has happened.

If you can produce an HP official document stating not to use it then Im all ears.

[Frank Slootweg]

Stefan, please re-read my response.

I explained what did and did not happen and why. There is nothing more I can add.

And no, as far I know, there is no "HP official document stating not to use it". If there was, we wouldn't be having this discussion. However, if you think that the absence of (official) negative advice means that it is recommended/not_dangerous/etc., then I can only say that I do not agree with that reasoning.

If you want to continue this, then please contact me by e-mail because this looks more like a fight than a discussion. (You have my e-mail address. I don't have yours.)

[Patrick Wallek]

Stefan,

Would you believe the advice if it came from Bill Hassell?

Read his response in this thread:
http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0xdbbb3a7b3682d611abdb0090277a778c,00.html

[Bill Hassell]

Just a note to second Frank's explanation. mediainit tells the firmware to do the formatting, and it does whatever is in the firmware. I agree with Frank that the man page is out of date--as are many man pages. Getting old docs updated is not an easy task in a large company like HP. There are higher priorities. We still still document and ship tcio and ftio, yet the disks and tapes they are designed for have not been manufactured for more than a decade.

It's really hard to get people to work on fixing/removing/documenting obsolete stuff, and even harder to get it funded. I'm still waiting for the labs to document SpeedyBoot even though the project was completed 2 years ago (see man setboot for 11i).

As for the original question, a simple pvcreate -f should wipe out the VxVM headers.