dbimport filling up /tmp?

Hi Family! I am trying to run a dbimport and have made sure there is plenty of room for the whole database - over 1.8 gig. I logged in as user informix, cd to /tmp and ran the following command: nohup dbimport -i /usr/local/dbexport -d dvgarpac mydatabase \\ >jakes_import 2>&1 & It hums along for a while, some tables taking more time than others, as I monitor the size of the file jakes_import and onstat -d to watch space left. Then I left for the night. The import aborted after 30 minutes after I left. The error message was: create index "informix".zzostari1 on "informix".zzostari (customer,store,division,season,style,color_code,lbl_code,dimension,size_bk,rec_type); *** execute sqlobj 212 - Cannot add index. 28 - No space left on device Looking at the chunks of the target dbspace: size free 102400 13 204800 34066 629760 629757 I see there was still plenty of room to pack the data into. What happened? According to the system admin person, the /tmp file system filled up about the same time the export aborted. I don't believe this is coincidence. The environment variable DBTEMP=/var/tmp, meant to bypass /tmp. Unfortunately, the /tmp files were not left over after the abort so we could not look at them. Q1. Does dbimport fill up /tmp with all kindsa stuff? Q2. Does dbimport fill the current directory with stuff? I will be testing for Q2 while I await help from y'all. Thanks. -- -- Jake (Never threw a glass house while living in a stone) +------------------------------------------------------------+ | The expedient performance of a task with excessive concern | | regarding its duration-to-completion engenders a virtual | | certainty of diminished benefit therefrom. | | -- Benjamin Franklin (but he said it in 3 words) | +------------------------------------------------------------+