Windoze SQL Anywhere 9.something.3951
So I have a large distributed database system...
It's kinda legacy at this point and every few years or so I have a large purge script that will delete multiple years worth of transactions.
We apply the script to the consolidated database and then the delete statements replicate out via dbremote running on a scheduled task.
A month later in production I find that although replication seems to be in synch it's as if the deletes never occurred on the remote database. So then I go back to the log and all I see is a roll back the end of a bajillion deletes.
Any ideas? I realize I'm out of luck being that the Syabse is EOL, and we are trying to move to 16 here shortly... But just kind of curious if anyone has seen this before?
Translating the log with a -g shows me:
I have seen this happen before due to foreign keys. For whatever reason, a remote had a record that didn't exist in the consolidated, so when replication tried to commit the deletions, a foreign key error occurred on the remote. Might be something to look for.
answered 16 May '14, 11:39