On all computer systems things go wrong due to user error, program error, or hardware failure. Unfortunately, is not immune to this. The section below reviews several general problem areas and their generalized solutions. Refer to §Z.1.5 for the details appropriate to NRAO’s computer systems. Some well-known possibilities follow.
If your workstation window is alive, but AIPS has “disappeared” you may have “suspended” it by typing CTRL Z. The AIPS can be left in a suspended state, placed into the “background” with bg, or returned to the “foreground” again with fg after which it will resume accepting terminal input. If your appears to be “suspended”, try typing jobs to see which jobs are attached to your window and then use fg %n to bring back job n where n is the job number of the suspended AIPS. If no job is suspended from the current window, check all other windows you have running on the workstation for the missing simian before starting a new AIPS. Otherwise, you may run out of allowed AIPSes and/or encounter mysterious file locking problems.
If your workstation window (or terminal on obsolete systems) is “dead”, i.e., refuses to show signs of talking to your computer, you have a problem. There are numerous possible causes. If typed characters are shown on the screen, but not executed, then
If typed characters do not appear, then
If you encounter the message CATOPN: ACCESS DENIED TO DISK n FOR USER mmm, it means that user mmm has not been given access to write (or read) on limited-access disk n. The access rights for all disks can be checked by typing FREESPAC in the AIPS session. In the list of mounted disks, the Access column can say Alluser, Scratch (scratch files only), Resrved (limited access including you), and Not you (limited access not including you). If you feel that you should have access to that particular disk, resume using your correct user number or see your Manager about enabling your user number.
If your data set seems to have disappeared, consider
The message write failed, file system is full will appear when the search for scratch space encounters a disk or disks without enough space. ( usually emits messages at this time as well.) This is only a problem when none of the disks available for scratch files has enough space, at which point the task will ”die of unnatural causes.” Run the verb FREESPAC to see how much disk is available and then review the inputs to the task to make sure that OUTDISK and BADDISK are set properly. Change them to include disks with space. Check the other adverbs to make sure that you have not requested something silly, such as a 2000-channel cube 8096 on a side. Then try again.
If there simply is not enough space, try some of the things suggested in §3.6, such as SCRDEST to delete orphan scratch files, DISKU to find the disk hogs, and, if all else fails, ZAP to delete some of your own files. DISKU may be run with DOALL = n to list catalog entries that occupy more than n Megabytes. This will help identify those files which will yield the most new space when deleted. Your Manager may help you by removing non- files from the data disks. Do not do this yourself unless they are your files.
All print operations now function by writing the output to a disk text file, then queuing the file to a printer, and then sometime later, deleting the file. After the job is queued, the task or verb will display information about the state of the queue. Read this carefully to be sure that the operation was successful and to find out the job number assigned to your print out. If you are concerned that your print job may be lengthy, or expect that you will only need a few numbers from the job, please consider using the DOCRT option to look at the display on your terminal or the OUTPRINT option to send the display to a file of your choosing without the automatic printing. See §Z.1.5.3 for information about printing such files later.
To find out what jobs are in the spooling queue for the relevant printer, type, at the monitor level:
$ lpq -Pppp C R
to show printer ppp.
$ lpstat ppp C R
to show printer ppp under Solaris, HP, SGI (Sys V systems).
where ppp is the name of the printer assigned to you when you began AIPS. If the file is still in the queue as job number nn, you can type simply
$ lprm -Pppp nn C R
to remove the job.
$ cancel nn C R
to remove the job under Sys V systems.
lprm and cancel will announce the names of any files that they remove and are silent if there are no jobs in the queue which match the request.
Since modern printers are capable of swallowing large amounts of input, your job may still be printing even though it is no longer visible in the queue. If you turn off the printer at this stage, you are likely to kill the remainder of your print job and quite possibly one or more other print jobs that followed yours. Use discretion. Do not turn the printer back on if the job is still in the queue. Most systems will start the print job over again after you turn the power back on without doing a lprm or cancel.
If your printout fails to appear
When AIPS does a software MOUNT of a magnetic tape, it actually reads the device on most systems. An error messages along the lines of ZMOUN2: Couldn’t open tape device … usually means that you have attempted the MOUNT before the device was ready. Wait for all whirring noises and blinking lights to subside and try again. Remote tape mounts are more fragile. If you get a message such as ZVTPO2 connect (INET): Connection refused, then the tape dæmon TPMON is probably not running on the remote host. EXIT and restart AIPS, specifying the remote host in the tp= option (see §2.2.3). If you are told AMOUNT: TAPE IS ALREADY MOUNTED BY TPMON, then there is a chance that you are trying to mount the wrong tape or that someone left the tape device in a mounted state. See §Z.1.5.7 for advice on curing this stand-off between AIPS, which knows that the tape is not mounted, and TPMON which knows that it is.
If you are having problems reading and writing a tape, consider