DMS error - Total storage used by swap files exceeded the limit 0 bytes

Question

Hi there,
I'm having issues with my DMS Full Load + CDC migration task where after the full load completes, the source latency continues to grow indefinitely. 
In my Cloudwatch logs for the task, I'm seeing an odd warning message around the time it happens:
```
2023-12-17T09:45:05 [SORTER          ]W:  Reading from source is paused. Total storage used by swap files exceeded the limit 0 bytes  (sorter_transaction.c:110)
```
1. I'm not sure how it's possible to have the limit for swap files be set to 0 bytes.
2. I'm not sure if this is related to the source latency issue I'm getting. Restarting the migration task does not help.

Any ideas?

Thank you,
DT

Answer

Have you modified BatchApplyMemoryLimit value from default to higher value?

When the swap files exceeds the size limit, sorter pauses reading from source and task logs shows as 0 bytes instead of actual size, when BatchApplyMemoryLimit value is set to higher value (higher than 4095mb).

If the BatchApplyMemoryLimit value is set to higher value, reduce it to less than 4095mb, so the actual size will be displayed in the log instead of 0 bytes.

To avoid swap file from growing too large follow this doc - https://repost.aws/knowledge-center/dms-swap-files-consuming-space

DMS error - Total storage used by swap files exceeded the limit 0 bytes

Relevant content