onstat -T flag 0x40000000

We are seeing this bit value in the third "flgs" column of "onstat -t" or "onstat -T" output for a minority of tables/indexes on IDS 11.70, which is not documented in the Knowledge Center or anywhere else. For example, on two entirely unrelated FC4/FC7 systems, but not some others, the following objects in the sysadmin database have this flag: ph_task ph_run#ix_ph_run_01 ph_run#ix_ph_run_02 ph_run#ix_ph_run_03 ph_alert#ix_ph_alert_01 ph_alert#ix_ph_alert_02 Try this on your system: onstat -T | grep ' 400' | partn -k5 -x You can download source for "partn" from the IIUG repository, which adds an object name column decoded from the part number. Has anyone else seen this flag and worked out what it means? Does anyone have an extended list for the documentation team? The highest currently described is 10000. Regards, Doug Lawry

Original post: We are seeing this bit value in the third "flgs" column of "onstat -t" or "onstat -T" output for a minority of tables/indexes on IDS 11.70, which is not documented in the Knowledge Center or anywhere else. For example, on two entirely unrelated FC4/FC7 systems, but not some others, the following objects in the sysadmin database have this flag: ph_task ph_run#ix_ph_run_01 ph_run#ix_ph_run_02 ph_run#ix_ph_run_03 ph_alert#ix_ph_alert_01 ph_alert#ix_ph_alert_02 Try this on your system: onstat -T | grep ' 400' | partn -k5 -x You can download source for "partn" from the IIUG repository, which adds an object name column decoded from the part number. Has anyone else seen this flag and worked out what it means? Does anyone have an extended list for the documentation team? The highest currently described is 10000. Regards, Doug Lawry Response: It looks like it's an internal flag which has some bearing on when during a checkpoint the partition structure in memory would get flushed to disk. So that might be why it's not really documented. A quick look at the doc for the onstat -T/-t and the flags we list, it would seem to me that there is aproblem with the documentation for 0x8000 and 0x10000. The flags might have been changed but I don't believe those flags mean what the documentation says they mean. There's a lot of flags that don't appear to have documentation when you compare the source with the doc. Jacques Renaut IBM Informix Advanced Support APD Team

Thanks, Jacques. I'll ask Hai-Nhu if this could be fully researched and the documentation updated. Regards, Doug