Stuart,

The error you are seeing can be broke down as follows:

-3 Error Manual says this is "Exceptional Error".


From the Manual:

-3         MESSAGE    TURBOIMAGE FAILURE
                      (FREADDIR FAILURE ON ROOT FILE AT xxxxx)

           MEANING    Root file was not readable.  Problem could be in
                      hardware.

           ACTION     Notify the database administrator of the error.


TurboImage return status -3:

This is usually the result of a call to DBGET or DBPUT when a broken chain
exists in the database. IMAGE passes to FREADDIR an MPE record number
derived
from the bad pointer, and FREADDIR determines that the desired record number
lies outside of the file (data set).

Lost Free Space or high water mark messages:
(1) The High Water Mark (HWM)*: If there is no delete chain, then we will
add one to the HWM and that is the record location we want to DBPUT. In this
instance, the record number we want to use should have the "in-use" bit off.
However, if that bit is on, then IMAGE doesn''t know what is going on,
hence, the Lost Free Space message.

(2) There is a problem with the Delete Chain head: If the Delete Chain head
is not 0, when a DBPUT is to be done, we get the record number from the
Delete Chain head,  rather than using the HWM. The "in-use" bit of the
record should be off and the data portion of the record should be binary
zeros. If *both* of these conditions are not met, a Lost Free Space message
will be issued.

To Fix it:
Customer must use database maintenance program such as adager, dbgeneral or
unload/load of the dataset that reports the error.

Recommendations:

1) Run diogenes against the dataset referenced in the error message.
   Provide report to HP Response Center.

2) Enable dumping using DBUTIL.PUB.SYS on all product databases.
   If image fails it will create dump files.  These dump files can contain
   additional information needed to determine the reason for the failure.
   The dump files are reviewed by the HP Response Center.

3) Have Hardware checked. System logs,  Disc drive logs, predictive reports
   etc.  The problem with the database may have been caused by a hardware
   problem.   If the disc drives appear to be fine, request that the
   following be checked: SCSI DA, bus cables, terminator, power supply
   and controller for all volumes configured on the system.

Hope this helps.
Cathlene Mc Rae
Sr Response Center Engineer.

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *