Out of extents. But ...

IBM Informix Dynamic Server Version 11.50.FC8X3 on HP-UX 11.31 We have a table that has run out of extents. And indeed it has; the last 11 added have been absolutely tiny. Until those last few the extent size kept doubling until the last 11. Then they were tiny. We would like to know why. Our first thought was maybe that there wasn’t enough space in the dbspace. But there is: PREPROD 2: [informix@brsux128] # onstat -d | grep appdb c000000043465990 7 0x60001 7 5 2048 N B informix appdbs c000000043468830 7 7 0 10240000 97372 PO-B- /opt/informix/dbspaces21/appdbs_1 c000000043468a30 8 7 0 13107200 241303 PO-B- /opt/informix/dbspaces21/appdbs_2 c000000043468c30 9 7 0 10000000 1060631 PO-B- /opt/informix/dbspaces21/appdbs_3 c000000043468e30 10 7 0 15000000 1535609 PO-B- /opt/informix/dbspaces21/appdbs_4 c000000043471230 12 7 0 62840832 4220060 PO-B- /opt/informix/dbspaces21/appdbs_5 And yes, some of the free space in chunk appdbs_5 for example is in large contiguous chunks - much larger than has been added recently, and much larger than the NEXT EXTENT SIZE. When we run ‘oncheck –pe toverify that there are large extents, that dbspace showed this: DBspace Usage Report: appdbs Owner: informix Created: 01/24/2008 ERROR: Failed to get header page for partnum 0x706f52 (buffer may be locked). Please limit DDL/DML activity when running this command. ... which I think may be related, I've never seen it before. The IBM technician said it's to be expected on a busy system but I can't EVER recall seeing it before. Any ideas why whatever it is that searches the free page list would eschew large contiguous free space and allocate multiple tiny extents instead? Thanks Neil (This is the subject of PMR 49555,019,866 btw).