[daip] frustrations trying to get some evla data to look at

Eric Greisen egreisen at nrao.edu
Mon Apr 11 11:10:34 EDT 2011


Susan Neff wrote:
> Hi Eric,
> 
> I'm having trouble getting EVLA data from the archive
> into AIPS.  So far I'm mostly having trouble getting the
> data from the archive to someplace I can see it.  Frazer
> worked with me this morning, and we got the data through
> a very circuitous route involving his UNIX machine, but
> that won't be available to me if I'm not here and/or he's
> not hanging around.  I'm writing this all down for you
> because some of it affects the cookbook writeup - and
> because I'm hoping you have some ideas how to make
> this easier.

The writing from the archive is done by a process that is not you and so 
your data area must be world writable for it to go directly to your place.

> 
> 
> I can get the archive to write files to the e2earchive 
> disk.
> It writes both UVFITS and SDMBDV files for me.  I see how
> to get it to write a tarball of the SDMBDV files.  I 
> don't see how
> to get it to write a tarball of UVFITS files (not sure 
> it's even
> possible to tar a single file).  I was not able to get 
> it to
> write directly to a zaurak disk - it said it did not have
> the requisite permissions to do so.
> 
See above.  You do not want to tar a uvfits file since you gain little 
by gzipping it.  The tarball of the ASDM file is what you want if you 
are taking it to some other institution besides NRAO Socorro - non tar 
is fine at nrao.
> 
> I am not able to run the obit scripts on the BDF/ASDM 
> files -
> zaurak tells me it doesn't know what AIPS_ROOT is 
> (although
> I am able to run AIPS on zaurak).   Do you know what I 
> need
> to do to be able to get the obit scripts to work in my 
> NRAO
> account here at AOC?

I plane to make a verb to do BDF2AIPS which will simplify this.  You 
need to source /home/AIPS/LOGIN.CSH on zaurak in order to make AIPS 
really know what is going on (you do this on your laptop).
> 
> I do not know how to get directly to (i.e. login to) 
> the e2earchive
> area from an AOC machine ... or from any other machine I
> guess.  Your Cookbook appendix says that the e2earchive
> area is in    /home/acorn2/ftp/pub/e2earchive
> However, zaurak says that there is no such directory.
> Not sure if this is a zaurak problem, or if the directory
> has moved.
> I can ssh from zaurak to a machine called acorn, but
> acorn does not seem to have a disk area anything like 
> the ftp
> area described above.  I did find an area called
> /e2earchive  on acorn, but it appears to be empty.
> 
They moved it - I will have to find it and update the documentation.
> 
> When I try to ftp from zaurak to ftp.aoc.nrao.edu, it 
> tells me that is
> an unknown host.   When I try to sftp, it asks me for 
> a password.
> Nothing I give it as "anonymous" seems to work.  When 
> I sftp as
> sgneff, it asks for my password, and then dies 
> ("Couldn't read packet")
> 
Bizarre

> So far, the only thing I've managed to do is to ftp 
> from my laptop
> to ftp.aoc.nrao.edu, which works.  (I have to login as 
> anonymous -
> when I come in as sgneff it puts me into some strange 
> directory).
> Then I was able to get the uvfits test file - it took 
> a half hour to
> transfer a 1.4GB file,  nrao to (laptop on network at) 
> nrao.  The
> data file I'll want to look at in May will be much 
> larger - probably
> a factor of 10ish larger, and will probably have to 
> come in through
> the goddard network or my phone line at home.
> 
> I'm hoping I'm just missing something obvious here ... but
> would like to find a way to get the data to move a bit 
> more
> quickly.
> 
> I was then able to copy the uvfits file from my laptop 
> to zaurak,
> (in another half hour).
> 
> I'm hoping there is perhaps something wrong with my 
> account
> here - this seems very cumbersome if I were, say, to 
> be doing
> a lot of testing.
> 
> 
> Frazer mentioned that you were considering building the
> obit scripts into AIPS, so that AIPS could read in the 
> SDMBDF
> files directly.  Is that anything that's likely to 
> happen in the
> near future, or is it a longer term thing?
Near future after I fix a CALIB issue.

Eric




More information about the Daip mailing list