Task Crashing with EXCEPTION_ACCESS_VIOLATION During Full Load (MySQL ODBC 8.0) and Recurring Bulk Apply Failures
We are experiencing two severe issues while performing Full Load and CDC replication using Qlik Replicate (Version: V2025.5.0.247) from MySQL (Source) to SQL Server (Target). We would appreciate any insight into the potential causes and solutions for these problems.
1. Critical Crash During Full Load
The replication task is repeatedly crashing during the Full Load phase with a critical Windows exception.
Error Type: EXCEPTION_ACCESS_VIOLATION
Involved Thread: SOURCE_UNLOAD thread
Analysis: Backtraces strongly suggest the involvement of the MySQL ODBC 8.0 Unicode Driver (myodbc8w.dll). This crash occurred multiple times after resuming the task.
【Question】 Are there any known issues or conflicts regarding the MySQL ODBC 8.0 (Version 08.00.0032) when used with Qlik Replicate (V2025.5.0.247)? What are the recommended stable ODBC driver versions, or would switching to a native connector be a better solution for stability?
2. Recurring Bulk Apply Failures in CDC (Performance Issue)
After the Full Load completes, during the Change Data Capture (CDC) phase, the task frequently fails to perform bulk updates for a specific table, resulting in significant performance degradation.
Affected Table: iims'.'tigi_damage_base (Id=84)
Error: Bulk apply operation failed and repeatedly reverted to the slower one-by-one mode for updates.
Related Warning: During the Full Load of another table (iims.tigi_damage), an invalid timestamp value ('1900-01-00 00:00:00.000000000') was detected and converted to NULL.
【Question】
What are the likely root causes (e.g., target side constraints, indexing conflicts, or Replicate internal settings) that could lead to the bulk apply failure and forced fallback to one-by-one mode for the iims'.'tigi_damage_base table?
Could the warning about the invalid timestamp value in a related table, iims.tigi_damage, indicate an underlying data quality issue that might be contributing to the Bulk Apply failures?
Replicate 2025.5 uses MySQL ODBC 8.0.32, which is the certified and fully supported version, so there shouldn’t be a compatibility problem. However, I don’t yet see a clear cause for the crash. To help isolate the issue, could you temporarily switch the target endpoint to NULL Device and check whether the crash still occurs? This will let us determine whether the problem originates on the source side or the target side.
2. Recurring Bulk Apply Failures in CDC (Performance Issue)
MySQL permits storing zero dates (and zero months), which can trigger bulk apply errors or cause performance degradation in CDC. Please refer to my article on handling this scenario:
For the task crash, we may need you to open a support case and attach verbose logs and a crash dump file as the logs from Replicate may not collect detailed enough data to get to the root cause. Please refer to this link:
