[daip] AIPS: solaris vs. redhat
Skinner Steve
skinners at mambo.Colorado.EDU
Mon Nov 1 18:49:22 EST 2004
Eric-
Thanks for the info. Its encouraging to know that
my calibrated UVSRT and ICLN files should transfer
over to Linux.
I tried writing out a UVSRT file in FITS
format with 'wrtdisk' and am getting an error. Apparently
AIPS cannot find the correct Unix directory to write to.
This looks like an environment variable error. AIPS
cannot translate my environment variable name
MYAREA into the correct lower case Unix subdirectory
name for some reason.
Any idea what is causing this? I will also run this
by our Unix people here.
Maybe I am missing some crucial punctuation somewhere
(e.g. a backslash or colon), or possibly I have an
input parameter set incorrectly. I am able to write
into the output directory, but not sure if that means
AIPS has write privilege.
Tx,
Steve Skinner
-------------
*Running AIPS / wrtdisk *
Before starting AIPS,
I must ssh onto our Unix server:
ssh -l skinners origins.colorado.edu
Then, I set up the Unix/Solaris environment variable to define
the Unix subdirectory that I want the AIPS files
to go into:
ORIGINS:~>pwd
/home/origins/skinners
ORIGINS:~>setenv MYAREA '/home/origins/skinners'
ORIGINS:~>echo $MYAREA
/home/origins/skinners
Then, start AIPS , RUN WRTPROCS, and try to write out a
AIPS UVSRT file on indisk 6/cat no. 42 using wrtdisk:
>outfile ='MYAREA'
>type outfile
AIPS 1: 'MYAREA '
>wrtdisk
AIPS 1: Got(1) disk= 6 user=1554 type=UV WR110_C.UVSRT.1
FITTP1: Task FITTP (release of 31DEC02) begins
FITTP1: Writing to disk file: MYAREA:DISK06C0042
FITTP1: ZFULLN: NO TRANSLATION FOR LNAME='MYAREA'
FITTP1: ZTPOPN: NO SUCH LOGICAL DEVICE = MYAREA:DISK06C0042
FITTP1: TAPIO: ERROR 3 OPENING FILE
FITTP1: COULD NOT OPEN OUTPUT FILE. IER= 8
FITTP1: Purports to die of UNNATURAL causes
FITTP1: origins 31DEC02 TST: Cpu= 0.0 Real= 0
Note: IEEE floating-point exception traps enabled:
** Here are the inputs I get, after the above abort:
>inp
AIPS 1: FITTP: Task to write maps or uv data on a FITS-tape.
AIPS 1: Adverbs Values Comments
AIPS 1: ----------------------------------------------------------------
AIPS 1: DOALL -1 1 (T) do all, -1 (F) do 1st.
AIPS 1: INNAME 'WR110_C ' Image name(name). Standard.
AIPS 1: INCLASS 'UVSRT ' Image name(class). Standard.
AIPS 1: INSEQ 1 Image name(seq. #).
AIPS 1: INDISK 6 Disk drive #. 0 => any
AIPS 1: INTYPE 'UV' File type. Blank => any
AIPS 1: OUTTAPE 1 Tape drive #. 0 => 1
AIPS 1: OUTFILE 'MYAREA:DISK06C0042 '
AIPS 1: Disk file name
AIPS 1: DOEOT 1 1 (T) go to end tape first.
AIPS 1: DOSTOKES 0 > 0 => Convert to true
AIPS 1: Stokes (Formal I)
AIPS 1: ******** Changed 12/2/93 ***********************
AIPS 1: DONEWTAB 0 >0 standard FITS binary
AIPS 1: tables (15JAN94 or later)
AIPS 1: <= 0 AIPS FITS binary tables
AIPS 1: *AIPS 1: FORMAT 3 tape
format: 0 => 3.
AIPS 1: ** press RETURN for more, enter Q or next line to quit print **
#
*******************************************************
* *
* Steve Skinner *
* *
* Center for Astrophysics and Space Astronomy (CASA) *
* 389 UCB *
* University of Colorado *
* Boulder, CO 80309-0389 USA *
* *
* Tel. 303-492-4202 *
* Fax. 303-492-4052 *
* E-mail skinners at casa.colorado.edu *
* Web: http://casa.colorado.edu *
* *
*******************************************************
On Mon, 1 Nov 2004, Eric Greisen wrote:
> Skinner Steve writes:
>
> > I will soon be replacing my Sun workstation
> > with a Dell Precision 370n PC running RedHat
> > Linux (RedHat Enterprise WS 3.0).
> > We will attempt to install AIPS under
> > Linux. The actual AIPS install will likely
> > be on our Linux server (RedHat Enterprise WS 3.0)
> > and I will attempt to run AIPS from my
> > Dell 370n client (.... we dont know yet if this
> > will work - hopefully it will).
>
> We have a single copy of the aips source code here and run it from
> about 100 computers. There are two - soon to be three - copies of the
> binaries as part of this tree. In other words, it should run fine if
> the clients can mount the server's disks. We recommend having the
> data areas on the machine that is doing the computing since nfs can be
> slow. This includes the (nfs has file locking foibles) the control
> files. In other words, in $AIPS_ROOT/DA00 we have files named for
> each host that are in fact link files to a disk area on that host. In
> other words we ln -s /home/primate/AIPS/DA00 PRIMATE where primate is
> my computer.
>
> >
> > Assuming we can get AIPS running under RedHat,
> > will it be possible to transfer my Sun Solaris
> > AIPS files over to the RedHat system, and read
> > these files into the RedHat version of AIPS
> > and then manipulate them with AIPS without
> > recalibrating from the raw data?
>
> You may transfer all files but you must do it by writing them out as
> FITS files and then reading them back in. AIPS files in the data
> areas are binary and so are affected by the opposite byte order. FITS
> files are also binary but have a defined byte order and so can carry
> the data between architectures. We provide procedures to help with
> this
>
> RUN WRTPROCS
> help wrtdisk
> help readisk
>
> for info
>
> >
> > Specifically:
> >
> > 1. Will I be able to transfer clean map
> > files (.ICLN) that were generated under
> > SunSolaris onto the RedHat LInux machine,
> > read these into AIPS on the Linux machine,
> > and make flux and/or position measurements
> > off of the .ICLN files using the RedHat
> > version of AIPS?
> >
> > 2. Will I be able to copy sorted UV files
> > (.UVSRT) from SunSolaris onto the RedHat
> > Linux machine, read these into AIPS under
> > RedHat, and generate new cleaned images from
> > them using AIPS under RedHat?
> >
> > .... or (worst case scenario)
> >
> > 3. Are the existing .UVSRT and .ICLN files
> > that were generated using AIPS on the
> > SunSolaris machine going to be incompatible
> > with the AIPS RedHat version?
>
> The only files that cannot be transferred are plot and slice files.
> All of the uv data, calibration tables, images, clean component tables
> etc will transfer niucely.
>
> Eric Greisen
>
>
More information about the Daip
mailing list