[mod.computers.vax] key words... MAIL and DISK QUOTA UTILITY

gregory@BRL-LVAX.ARPA.UUCP (03/05/87)

Hello out there,

I have a couple of separate questions regarding mail and the disk quota utility:

RE: Mail
While in mail and reading a mail message, occasionally the text at the
top of the screen will scroll out of sight after pressing RETURN to read more
of the message.  Does anyone know how to fix this?  I have found that
if before you enter mail you do a $ set term/page=n, where n is somewhere around
20, this will stop the scrolling of the screen but I consider this a bandaid
and not a fix.  Remember to $ set term/page=24 when you get out of mail.

RE: Disk Quota Utility
We are considering turning disk quotas on our user disks and I would
like to get some info from you regarding the pros and cons of the utility
such as I/O performance considerations, how/when does overdraft work,
how fast does the quota data base get updated, etc.  As far as overdraft
and how it works, I do not understand the following statement in the System 
Manager's Reference Manual (V4.4):
"If the rejected operation is an extension of a file opened for write, a user
 with overdraft can perform the operation by retrying it."
Is this "retrying it" automatically handled by RMS for say a FORTRAN program
that has a file opened for write?

RE: Our Site Configuration
We are running VMS V4.4 and MicroVMS V4.4.  We have a 780 and an 8600 clustered
together and DECneted to two Micro VAX II's and one 730.

Any and all info is greatly appreciated.  As always, if you can point me towards
the proper manual(s) that actually answer these question that will be just fine.


Thanks,

Dan Gregory
ARPA: GREGORY@BRL
Phone: 301-278-3485
------

LEICHTER-JERRY@YALE.ARPA.UUCP (03/06/87)

    RE: Mail
    While in mail and reading a mail message, occasionally the text at the
    top of the screen will scroll out of sight after pressing RETURN to read
    more of the message.
    
MAIL has a rather simple-minded way of doing paging:  It counts "lines", where
a "line" is terminated by a CR/LF.  (Actually, to be more accurate, it counts
records.)  If a line in a message is longer than can fit on one physical
line on the terminal - 80 or 132 characters on most terminals, almost anything
on a VAXStation - VMS or the terminal will wrap (depending on whether the
VMS /WRAP characteristic is set, or the corresponding SETUP option on the
terminal), inserting a newline where MAIL is unaware one exists.  So, of
course, MAIL's count is off and it writes too many lines.

Short of the work-around you mention - artificially setting the page size a
couple of lines smaller, leaving you with some extra space - all I can recom-
mend is that you convince your correspondents not to put really long lines in
mail messages.

    RE: Disk Quota Utility
    We are considering turning disk quotas on our user disks and I would
    like to get some info from you regarding the pros and cons of the utility
    such as I/O performance considerations,

As far as I've been able to determine, the overhead is quite small.  The only
time the quota database has to be updated is when you actually allocate or
free disk blocks, and hence are doing disk I/O anyway; the extra I/O to
maintain the quota database is lost in the noise.  Besides, the quota stuff
is cached.

					   how/when does overdraft work,

If an RMS operation requires you to allocate additional disk blocks, and you
are at your quota limit but have an overdraft available, the operation will
fail with a special status value.  If you re-try the operation, it will
succeed.  (I'm being vague here because the documentation is vague and I've
never played with this enough to REALLY understand it.  If, for example, you
have two jobs running, and one runs into the quota limit and stops, will
the other's next attempt to use the overdraft succeed without an error?  I'm
not sure.  It's not even clear whether the ability to go into the overdraft
region is inherent in the file system, or an RMS "added value".)

    how fast does the quota data base get updated?

Essentially immediately.

    ... I do not understand the following statement in the System Manager's
    Reference Manual (V4.4):  "If the rejected operation is an extension of
    a file opened for write, a user with overdraft can perform the operation
    by retrying it."
    Is this "retrying it" automatically handled by RMS for say a FORTRAN
    program that has a file opened for write?
    
No, RMS will not retry the operation automatically, nor as far as I know will
the FORTRAN I/O system; it will simply pass the error back to the calling
program to deal with.  Very few programs deal with this error intelligently.
I believe BACKUP does.  I've seen complaints that indicate that EDT does not -
most annoying, as it means you can lose your edit.  I'm not sure if TPU is any
more clever.

    As always, if you can point me towards the proper manual(s) that actually
    answer these question that will be just fine.
    
Unfortunately, the documentation of this whole area is pretty thin.  Then
again, so is the implementation.  In my experience, the overdraft feature is
pretty useless as currently implemented - it requires special-case handling
by every program that wants to write files, and hardly any bother.

							-- Jerry
-------