SPU swap partition – Disk temporary work space is full
SPU swap partition : Disk temporary work space is full
This error does not mean that we have a space issue on any of our dta slices but ratehr this error means that the maximum temporary workspace is full. This workspace is used by all the temporary tables, sorting and/or data skew. This can also occur if the query have cartesian joins (i.e. unresolved M::M relationships), causing the temporary workspace to grow out of control.
Based on the above causes you cna try one or all of the below option.
1) Verify if the tables involved in the query have high skew. If yes then try to remove skew by re-distrubiting the table with appropiate distribution key.
2) Try to drop the temporary tables once they are not needed in the session.
3) Try to run lesser number of concurent jobs with thise heavy query which is eating up all the temporary disk space.
4) Try to rewrite the query more efficiently by avoiding cartesion joins if possible.
Run nz_spu_swap_space support script to find the current usage of SPU swap space.