We are currently experiencing issues with missing data from Oracle over to a CDC server. We can see in the back end CDC tables (_CT), that some data appears to be missing, but for no obvious reason.

The Oracle DBAs think that CDC might be reading the online redo logs instead of the archive, meaning that it sometimes misses data when the logs switch over. Hopefully someone has some ideas. Here is an extract of our data to show the issues, just the ID and date are shown, the row is about 70 columns wide in total. Its currently around 20-30% of the data missing.


Oracle Server CDC Server
ID Date ID Date
16 10/13/2015 09:03:11 16 13/10/2015 09:03
15 10/13/2015 09:00:05 15 13/10/2015 09:00
14 10/13/2015 08:58:42 14 13/10/2015 08:58
13 10/13/2015 08:55:52 13 13/10/2015 08:55
12 10/13/2015 08:53:49 Missing from CDC server
11 10/13/2015 08:53:44 11 13/10/2015 08:53
10 10/13/2015 08:52:40 10 13/10/2015 08:52
9 10/13/2015 08:52:40 9 13/10/2015 08:52
8 10/13/2015 08:46:08 8 13/10/2015 08:46
7 10/13/2015 08:25:56 Missing from CDC server
6 10/13/2015 08:21:35 6 13/10/2015 08:21
5 10/13/2015 08:19:19 Missing from CDC server
4 10/13/2015 08:10:53 Missing from CDC server
3 10/13/2015 08:10:27 3 13/10/2015 08:10
2 10/13/2015 07:47:06 2 13/10/2015 07:47
1 10/13/2015 07:45:57 1 13/10/2015 07:45