mount-f-exquota

[steve esson]

What are the likely resources that cause this error when trying to mount a device?

[Karl Rohwedder]

Steve,

pls. tell us the exact command/message you entered/received.

regards Kalle

[Jan van den Ende]

Steve,

Welcome to the VMS Forum!

In addition to Kalle's request: What VMS version/patch level?

Proost.

Have one on me.

jpe

[Richard W Hunt]

Answer also depends on whether this is a private or /SYSTEM (or /CLUSTER) mount, I think. Hits different quotas.

The items being charged are probably VCBs for the volume and FCBs for the file structure files like INDEXF.SYS, at least as a first blush. Where ever that stuff goes is the quota in question. Non-paged pool?

[John Travell]

Can I suggest looking at the process quotas immediately prior to attempting the mount ?
in another (privileged) process, do:
$ analyse/system
SDA> set process /id= (pid of process about to do the mount)
SDA> show process
You will get some of the current process quotas displayed in the bottom right corner of the display, e.g.

# open files remaining 97/100
Direct I/O count/limit 150/150
Buffered I/O count/limit 150/150
BUFIO byte count/limit 198976/198976
ASTs remaining 247/250
Timer entries remaining 10/10
Active page table count 0
Process WS page count 407

Look for any quotas where you already have a large difference between the first and second numbers, e.g.
Direct I/O count/limit 50/150
and in any case, post what you find.
JT:

[steve esson]

The VMS version is 5.5-2
The hardware is a vaxstation 3100.
It's still running after all this time with the original disks as well. Now thats what you call a reliable machine.

Unfortunately this is a real system from which I have to extract data. The data is stored on optical disks.

The drive required a custom device driver so the mount command goes thus:-
$ odmount
The error is thus:-
MOUNT-F-EXQUOTA, exceeded quota.

It's been so long since I touched VMS I had forgotten about the analyse/system command.

It shows nothing of any significance that I can see. There appear to be no exceeded process level quotas and the process virtual pages are restricted by neither virtualpagecnt nor pgflquo.

The pool areas appear to be largely empty.
Any chance it could be related to the ACP caching for dirs, headers, map areas etc ?

The optical disks themselves are standard ODS format but apparently without the volume image info (used in backup/image )
Any further suggestions would be most welcome.

Just for interest, the data I need to recover is the bench test radar descent data for the Huygens probe which landed on Titan 3 years ago. Why they chose to put the data on optical disks on a 15 year old workstation is another mystery of the universe.

[John Travell]

Steve,
You will not have any quotas already exceeded before mounting the device.
In the example I gave you,
Direct I/O count/limit 50/150
you started with 150 and have counted down so that you have 50 left. The absolute values you see here do not necessarily reflect the UAF quotas for the process, as there are circumstances where VMS makes permanent deductions from the UAF quota that are only returned when you logout...
Are there any of the 'xx' values in 'xx/yy' that are a LOT less than 'yy', or any where 'xx' is close to zero ?
JT:

[Hein van den Heuvel]

Steve,

Glad to see the detail information. Why not provide that right away?! 'mount' works all day every day for 1000's of VMS systems, without exquota. So there is something special about the setup. It now turns out that everything is special! A vax, 5.5-2 (amber: a fine and stable early 90's vintage), and not a Plain Old Disk with plain old mount either.

Anyway, my money for the mount problem is on SYSGEM BYTLM and/or AUTORIZE /BYTLM settings
$MCR AUTHORIZE SHOW xyz/full
or login and
$SHOW PROC/QUO: Buffered I/O byte count quota...).
$MCR SYSGEN SHOW PQL_DBY
and
$MCR SYSGEN SHOW PQL_MBY
or
$MCR SYSGEN SHOW /PQL

And uh... get ready to dive into the driver sources for odmount or call up support (ha!) for it.

hth,
Hein.

[Hoff]

This might be the osms/osds and the RWZ-series magneto-optical disk widgets.

Brute force: double your process quotas (all of them), and log in anew, and repeat until it works, or until the mount blows up differently.

Alternatively, find the operator account on this or the original box, and clone it.

osms/osds and the RWZ-class gear all moved from DEC over to US Design, IIRC, and moved over many years ago.

Or you could get somebody in to deal with this case and this investigation and the migration for you, if you're disinclined to start reverse-engineering an old MO environment and old (now) third-party code.