At a client I work for, we are using Attunity CDC for SSIS to replicate data from a Oracle source to a MS SQL server destination. We are currently testing the solution and while testing, an error occurred: Attunity got the status Suspended, substatus LOGGER.
This implicates there is an temporary error on de Oracle redo-log, but Attunity never got past this error. Our interpretation is that Attunity was looking for the “Transaction log head change number”, but could not find it. We temporarily solved this by changing the “Transaction log head change number” manually (without looking for a CSN in the alert-log), from that point on Attunity was working again. We are not sure if there was any data loss. So we are trying to force an error in Attunity to test its reliability and stability, and to reproduce this error. One of the goals is to see if we can get Attunity back on track with a “transaction log head change number” (that we created based on the CSN from the Oracle Alert-log).
The reason we think Attunity could not find the “Transaction log head change number”, is because the Oracle-server Attunity mines was restarted, but before it restarted it killed all of its sessions. We think Attunity already processed some new records that Oracle has rolled back afterwards.
Based on this story I have three questions:
- How can we let Attunity crash and get is into Status Suspended, substatus LOGGER?
- Do you know a way to get Attunity back on track without, or with minimum, data loss?
- Is there a lists of know errors that force Attunity to crash?