[daip] AIPS data disks limit

Eric Greisen egreisen at nrao.edu
Mon Mar 6 10:37:31 EST 2017


On 03/06/2017 06:53 AM, Marc Verheijen wrote:
> Dear Eric,
>
> the new Apertif Phased-Array Feed on the Westerbork telescope is close
> to starting its early-science projects.  The advantage of Apertif is
> that it can simultaneously observe up to 42 primary beams on the sky,
> greatly enhancing its field-of-view.
>
> In anticipation of the Medium-Deep Survey, I'm trying to set up a
> re-processing pipeline using AIPS and ParselTongue on my 8-node,
> 320-core compute cluster.  For an effective and 'clean' data management,
> I would like to assign a different AIPS data disk for each of the 42
> so-called compound beams on the sky.  It seems, however, that AIPS
> 'only' supports 35 data areas.
>
> Would it be possible to increase the number of data areas to more than
> 35? Alternatively, could you think of a transparant work-around for this
> problem?
>
> Many thanks in advance for any help or suggestion you can provide.
>

AIPS uses "extended hex" 0-9,A-Z as its basic numbering system - all 
file names on disk, "user numbers" in those disk file names, etc. are 
created using that system.  It would be impossible to change that now 
for the multi-ehex-digit aspects of this.  That said, it might not be 
impossible to change it solely for disk number.  There is one low-level 
routine that changes disk number to EHEX plus a messy set of .SH and .PL 
routines that read the dadevs files.

But the obvious extension is to add lower-case letters and that brings 
up an issue.  On Macs $DA0A and $DA0a may not always be treated as 
different.  Having written this however why not just allow the 0 to also 
be a 1.  I will try to see if there are consequences of changing the 
limit to 71.

Cheers

Eric


PS the other thing I was think about was MOVE to a variety of user 
numbers but this would be better.





More information about the Daip mailing list