Hi all,

IDS 11.50.FC5 and IDS 11.50.FC5W2

I tried to start a new replicate with synchronisation. The table holds about
2 Million rows on the primary. On the secondary grp_db5 the table was just
created and empty.

Here are the commands and the results:

cdr define repl -c grp_sun1 -i -A -R -C ignore -S row foo_repl \
"P p3_neu at (no spam) grp_sun1:informix.$i" "select * from foo" \
"R p3_neu at (no spam) grp_db5:informix.$i" "select * from foo" ;

cdr start repl foo_repl --syncdatasource=grp_sun1

The replicate was created and a cdr sychronisation job was running.

With cdr list repl there were listed the replicate foo_repl and a child
which was the synchronisation job.

With onstat -g sql | grep cdr you could see 3 sessions, one with sql error
255.

On the SECONDARY one could see some activity though no rows where inserted
in the replicated table. Online.log were growing rapidly with entries like
that:
11:56:39  CDR CDRD_5: transaction aborted (All rows in a transaction defined
with row scope were rejected) with sql error 0 isam error 0.
11:56:39  CDR CDRD_4: transaction aborted (All rows in a transaction defined
with row scope were rejected) with sql error 0 isam error 0.
11:56:39  CDR CDRD_4: failed rows spooled to file
/home/informix/ris/ris.grp_db5.grp_sun1.D_4.091103_11:56:39.1
11:56:39  CDR CDRD_6: transaction aborted (All rows in a transaction defined
with row scope were rejected) with sql error 0 isam error 0.
11:56:39  CDR CDRD_4: failed transaction spooled to file
/home/informix/ats/ats.grp_db5.grp_sun1.D_4.091103_11:56:39.2
11:56:39  CDR CDRD_5: failed rows spooled to file
/home/informix/ris/ris.grp_db5.grp_sun1.D_5.091103_11:56:39.1
11:56:39  CDR CDRD_4: transaction aborted (All rows in a transaction defined
with row scope were rejected) with sql error 0 isam error 0.
11:56:39  CDR CDRD_5: failed transaction spooled to file
/home/informix/ats/ats.grp_db5.grp_sun1.D_5.091103_11:56:39.2
11:56:40  CDR CDRD_5: transaction aborted (All rows in a transaction defined
with row scope were rejected) with sql error 0 isam error 0.
11:56:40  CDR CDRD_4: failed rows spooled to file
/home/informix/ris/ris.grp_db5.grp_sun1.D_4.091103_11:56:39.3
11:56:40  CDR CDRD_7: transaction aborted (All rows in a transaction defined
with row scope were rejected) with sql error 0 isam error 0.
11:56:40  CDR CDRD_4: failed transaction spooled to file
/home/informix/ats/ats.grp_db5.grp_sun1.D_4.091103_11:56:39.4
11:56:40  CDR CDRD_5: failed rows spooled to file
/home/informix/ris/ris.grp_db5.grp_sun1.D_5.091103_11:56:39.3
11:56:40  CDR CDRD_4: transaction aborted (All rows in a transaction defined
with row scope were rejected) with sql error 0 isam error 0.
11:56:40  CDR CDRD_5: failed transaction spooled to file
/home/informix/ats/ats.grp_db5.grp_sun1.D_5.091103_11:56:39.4
11:56:40  CDR CDRD_7: failed rows spooled to file
/home/informix/ris/ris.grp_db5.grp_sun1.D_7.091103_11:56:39.1
11:56:40  CDR CDRD_5: transaction aborted (All rows in a transaction defined
with row scope were rejected) with sql error 0 isam error 0.
11:56:40  CDR CDRD_4: failed rows spooled to file
/home/informix/ris/ris.grp_db5.grp_sun1.D_4.091103_11:56:39.5
11:56:40  CDR CDRD_4: failed transaction spooled to file
/home/informix/ats/ats.grp_db5.grp_sun1.D_4.091103_11:56:39.6
11:56:40  CDR CDRD_7: failed transaction spooled to file
/home/informix/ats/ats.grp_db5.grp_sun1.D_7.091103_11:56:39.2
11:56:40  CDR CDRD_5: failed rows spooled to file
/home/informix/ris/ris.grp_db5.grp_sun1.D_5.091103_11:56:39.5
11:56:40  CDR CDRD_4: transaction aborted (All rows in a transaction defined
with row scope were rejected) with sql error 0 isam error 0.
11:56:40  CDR CDRD_7: transaction aborted (All rows in a transaction defined
with row scope were rejected) with sql error 0 isam error 0.
11:56:40  CDR CDRD_5: failed transaction spooled to file
/home/informix/ats/ats.grp_db5.grp_sun1.D_5.091103_11:56:39.6
11:56:40  CDR CDRD_4: failed rows spooled to file
/home/informix/ris/ris.grp_db5.grp_sun1.D_4.091103_11:56:39.7
...
The ris-directory filled with thousands of files with the following content:

TXH Source ID:5 / Name:grp_sun1 / CommitTime:09-11-03 12:00:21
TXH Target ID:67 / Name:grp_db5 / ReceiveTime:09-11-03 12:17:14
----------
RRH Row:1 / Replicate Id: 327793 / Table: p3_... at (no spam) informix.foo / DbOp:Update
RRH CDR:6 (Error: Update aborted, row does not exist in target table) /
SQL:0 / ISAM:0
RRD 504604925|M3RGZ|-1|0|200909|303506713|||||
=========> TXH Transaction aborted
TXH ATS file:/foo/ats/ats.grp_db5.grp_sun1.D_9.091103_12:17:14.27428 has
also been created for this transaction

The ats-directory filled with thousands of files with the following content:
TXH RIS
file:/foo/informix/ris/ris.grp_db5.grp_sun1.D_9.091103_12:17:14.27427 has
also been created for this transaction
=========> TXH Source ID:5 / Name:grp_sun1 / CommitTime:09-11-03 12:00:21
TXH Target ID:67 / Name:grp_db5 / ReceiveTime:09-11-03 12:17:14
TXH Number of rows processed when transaction was aborted:1
TXH All rows in a transaction defined with row scope were rejected
TXH CDR:6 (Error: Update aborted, row does not exist in target table) /
SQL:0 / ISAM:0
----------
RRH Row:1 / Replicate Id: 327793 / Table: p3_... at (no spam) informix.foo / DbOp:Update
RRD 504604925|M3RGZ|-1|0|200909|303506713|||||

Those files are reporting UPDATE errors. Sure, the receiving part of the
synchronisation was empty. I expected that INSERTS were done without any
error.

On the primary (sending) server the sbspace (er_sbspace) which is configured
for ER buffering filled up. The whole process stalled. I enlarged the
dbspace by adding a chunk of 2 GB but he filled up in minutes again.

I tried to stop the synchronisation with:
cdr stop repair foo_repl_job_5_22_1257245799 but nothing happens
Then I tried:
cdr delete repair foo_repl_job_5_22_1257245799 -S grp_sun1
The delete was acknowledged but the job was running obliviously.

Now I made a real mistake: I deleted the replicate with cdr delete repl
foo_repl. The synch.-job was still running but the server wasn't happy:
12:40:26  Assert Failed: No Exception Handler
12:40:26   Who: Session(1057795, root at (no spam) A40DBS0S, 0, 25f4187e0)
                Thread(1066673, CDRScn_327793, 261a251a8, 14)
                File: mtex.c Line: 491
12:40:26   Results: Exception Caught. Type: MT_EX_OS, Context: mem

I killed the cdr synch.-job with kill -9 what I better had done ealier,
anyhow the server crashed. Sh.., it was a production server.

I had tried the same with smaller tables up to 500.000 rows before and it
_had_ worked. What has gone wrong in this case? Why are UPDATE errors
reported? Can anybody help?

TIA,
Reinhard.