Maximum disk size for NFSmount under Multinet 4.4A

[Fairfield, Ken :CO IR]

The following question (with somewhat less Multinet info) was also
posted to comp.os.vms ...

Vital Statistics:
-----------------
Process Software MultiNet V4.4 Rev A-X, hp AlphaServer GS1280 7/1300,
OpenVMS AXP V7.3-2

VERSION V4.4
REVISION A-X
DATE 30-Jan-2002
PRODUCER PSC
INSTALLED_IPAPPS YES
INSTALLED_NFSCLIENT YES
INSTALLED_NFSSERVER YES
.
*No NFS patches have been installed.
------------------------------------

In trying to take a copy of our database (Oracle/RMAN) to some a
"special" storage device, we're NFSmount'ing the external storage
and getting some very strange results, as follows:

A Linux box serves a 6.5TB mount point to VMS/Multinet. The
Multinet NFSMOUNT succeeds and we can copy files at will.

However, a SHOW DEVICE/FULL on the NFSx device gives
bogus number for various fields. For example, Total Blocks
reads 1453299520, or about 692GB (not 6500GB). The Logical
Volume size shows 51511616 (24GB), and the Free Blocks at
mount time is somewhere around 670000000 (315GB).

We tried ignoring the obviously incorrect data and forging ahead,
but in fact, the Free Blocks decrements as we write data and
the RMAN eventually bombed out when we got ~310GB
written to the device.

Given the age of the software, I'm not surprised. It looks
like Linux handed Multinet an integer size too big to fit in
a longword (or similar) and so it picks up and uses the
least significant 32bits (just speculation, but...).

So my question is, what is the maximum size of an ODS-2
disk prior to the recent enhancements for Alpha (not just
ODS-5, but the volume expansion stuff related to changes
in BITMAP.SYS, etc.)? The Help just talks about cluster
size (due to old BITMAP constraints).

I've opened a call with Process Software, but don't have
an answer yet. I'm hoping we can partition down the 6.5TB
into several manageable size pieces, but I'd like an idea
of what that size should be.

Thanks! -Ken
--
Ken & Ann Fairfield
What: Ken dot And dot Ann
Where: Gmail dot Com

IMPORTANT NOTICE:
This communication, including any attachment, contains information that may be confidential or privileged, and is intended solely for the entity or individual to whom it is addressed. If you are not the intended recipient, you should contact the sender and delete the message. Any unauthorized disclosure, copying, or distribution of this message is strictly prohibited. Nothing in this email, including any attachment, is intended to be a legally binding signature.

[Jeremy Begg]

Hi Ken,

There is a VMS Files-11 limitation of 1TB for any disk volume. MultiNet
tries hard to make NFS mountpoints look like native Files-11 volumes, hence
larger ones are almost certain to present problems.

Regards,

Jeremy Begg

Post by Fairfield, Ken :CO IR
The following question (with somewhat less Multinet info) was also
posted to comp.os.vms ...
-----------------
Process Software MultiNet V4.4 Rev A-X, hp AlphaServer GS1280 7/1300,
OpenVMS AXP V7.3-2
VERSION V4.4
REVISION A-X
DATE 30-Jan-2002
PRODUCER PSC
INSTALLED_IPAPPS YES
INSTALLED_NFSCLIENT YES
INSTALLED_NFSSERVER YES
..
*No NFS patches have been installed.
------------------------------------
In trying to take a copy of our database (Oracle/RMAN) to some a
"special" storage device, we're NFSmount'ing the external storage
A Linux box serves a 6.5TB mount point to VMS/Multinet. The
Multinet NFSMOUNT succeeds and we can copy files at will.
However, a SHOW DEVICE/FULL on the NFSx device gives
bogus number for various fields. For example, Total Blocks
reads 1453299520, or about 692GB (not 6500GB). The Logical
Volume size shows 51511616 (24GB), and the Free Blocks at
mount time is somewhere around 670000000 (315GB).
We tried ignoring the obviously incorrect data and forging ahead,
but in fact, the Free Blocks decrements as we write data and
the RMAN eventually bombed out when we got ~310GB
written to the device.
Given the age of the software, I'm not surprised. It looks
like Linux handed Multinet an integer size too big to fit in
a longword (or similar) and so it picks up and uses the
least significant 32bits (just speculation, but...).
So my question is, what is the maximum size of an ODS-2
disk prior to the recent enhancements for Alpha (not just
ODS-5, but the volume expansion stuff related to changes
in BITMAP.SYS, etc.)? The Help just talks about cluster
size (due to old BITMAP constraints).
I've opened a call with Process Software, but don't have
an answer yet. I'm hoping we can partition down the 6.5TB
into several manageable size pieces, but I'd like an idea
of what that size should be.
Thanks! -Ken
--
Ken & Ann Fairfield
What: Ken dot And dot Ann
Where: Gmail dot Com
This communication, including any attachment, contains information that may be confidential or privileged, and is intended solely for the entity or individual to whom it is addressed. If you are not the intended recipient, you should contact the sender and delete the message. Any unauthorized disclosure, copying, or distribution of this message is strictly prohibited. Nothing in this email, including any attachment, is intended to be a legally binding signature.