I have a situation on a productions system where our production jobs are failing with the 'Out of spool space on disk' error since yesterday. This process has been running for over a year and the data volumes being processed are consistent with the usual. The runs are currently failing whenever the total spool used is 1x although I have 6x space free in the database and the user is assigned >6x spool space limit for the user. There does not seem to be any change in the skewness of the data in the system either.
Usually when we run out of space, I see a 'no more spool space' error but this time it is 'out of spool space on disk' error. Teradata documentation does not provide much help in diagnosing and resolving the issue. There is nothing else running on the box and there is sufficient perm space. Any pointers on what else should be investigated will be appreciated.
There is no disk space to provide spool space for the process. The spool space available is provided by the free cylinders in the filesystem. This usually means that you are running out of space. Notice that this can happen on a single AMP which can be full (its filesystem) due to skew.
"No more spool space" is a 'logical limit' (imposed as a rule) whereas your error is a 'physical limit' (lack of resources).
Forgot to mention:
If the system activity is high when you execute your process, other processes very spool- demanding could be using the free space for their own spools, thus leaving too few space for your needs.