This error is a symptom of a corrupt transaction log. We can get you to synchronize again, but we might need to skip uploading the data between offset 71898135 and the current log offset.

1) Can you run dbtran against the transaction log that contains offset 71898135, possibly using the -ir switch to scan up to the bad offset and then a second run of dbtran to translate operations after the bad offset? If you can see the operations in the log after the bad offset, you will need to manually add these operations to the consolidated database (if you want). If you can't run dbtran at all, and the operations in the transaction log after offset 71898135 are critical, you will need to contact technical support to see if your corrupted transaction log can be repaired.

2) Rename and restart the active transaction log

3) Drop and re-add the synchronization subscriptions in your remote database. You can dbunload -n on the remote database to generate a reload.sql file if you're not sure what the exact CREATE SYNCHRONIZATION SUBSCRIPTION command is.

4) The next time you run dbmlsync it will start scanning at whatever the current log offset was when you dropped and recreated the synchronization subscriptions, which will be past offset 71898135.