p_manage_27 stalls / gets "bad size ... Internal heap ERROR"

• Article Type: General
• Product: Aleph
• Product Version: 16.02

Description:
A first run of p_manage_27 stalled with only 10 cycles done. The cycles file and the p_manage_27_a logs were not been written to for more than a day. The only $data_scratch file being written to is p_manage_27_c.log -- with "sleeping" being written there constantly.

Subsequent runs of p_manage_27 consistently stopped with a HEAP dump, like this:

39039 READING DOC 000289538 AT 17:29:01
kghalo bad size 0x04000020
********** Internal heap ERROR KGHALO2 addr=0 ********

Other jobs -- especially multi-process jobs -- were also getting this error.

Resolution:
p_manage_27 ran successfully after the following:

1. Export data from xxxxx (all libraries) using ALEPH export procedure:

csh -f $aleph_proc/exp_all_libraries > & $TMPDIR/exp_all_libraries.log &

(Ensure there is enough disk space.)

2. Remove existing DB ORA xxxxxx and create a new one with the same size

3. Reinstall Oracle: copy /exlibris/app/oracle/product/b.920.tar.gz from Ex Libris or from server where it is working OK to target server.

4. Import data to new xxxxxx using ALEPH import procedure:

csh -f $aleph_proc/imp_all_libraries yes > & $TMPDIR/imp_all_libraries &

[Make sure that there are no errors -- such as errors due to the fact that the .seq files are > 2G.]