ellen@calgary.gsfc.nasa.gov (Ellen Salmon) (11/16/90)
Fellow Cray Station (ab)users,
I would like to have a shell script "capture" (in a file or variable)
the informational messages that come back from the front-end station after a
fetch/acquire/dispose, especially if the fetch/acquire/dispose was unsuccess-
ful. I'm referring to messages like those starting with "CRAY TO VAX" and
the one starting with "13:25:46" below:
+ dispose datefile -mvx -tcartan"xrems xxxxxxxx"::datefile.txt
CRAY TO VAX: %RMS-S-NORMAL, normal successful completion
CRAY TO VAX: FILE=CARTAN"XREMS password"::DISK$USER01:[XREMS]DATEFILE.TXT;7
CRAY TO VAX: 29 BYTES TRANSFERRED
+ fetch vmfile.notes1 -mvm -tfn=vmfile,ft=note1,addr=304,tid=xrems,pw=xxxxxxx
13:25:46 CRIDSK004E CMS file VMFILE NOTE1 not found
fetch: transfer cancelled
I'm wondering if these messages are even capturable. They apparently
do not come back on stdout or stderr of the subshell doing the
fetch/acquire/dispose command (i.e., redirecting the fetch/acquire/dispose's
stderr and stdout does not capture the messages). Also, if an NQS batch job
doing this kind of transfer is not submitted from a front-end station (or via
uscpqsub from a job submitted from a front-end station), these messages never
appear in the batch job's stderr/stdout. This leads me to believe that
the messages may not be capturable because they might be (1) asynchronous,
(2) require something special in the station slot authorization file for
NQS jobs, or/and (3) are part of the magic world of the UNICOS Station Call
Processor and not a mere-mortal UN*X application. But I thought I'd consult
with the network, because I expect better brains than mine have wrestled
with this.
We're running UNICOS 5.1.9 with these front-end station software
releases: VMS 4.02, VM 7.00, MVS 3.02.
Thanks for any enlightenment,
Ellen Salmon, ST Systems Corp. <xrems@scfvm.gsfc.nasa.gov>
c/o NASA Ctr for Computational Sciences, NASA/GSFC Code 932, Greenbelt,
MD 20771