- Mais recentes
- Mais votos
- Mais comentários
In your AWS DMS setup, using the resume-processing
task type with EventBridge for scheduled tasks is intended for ongoing replication and not for initiating new full loads. This explains why your DMS task did not transfer any data when triggered; it was expecting to find new changes for replication but found none. To conduct weekly full data loads, adjust the EventBridge Scheduler to use the start-replication
task type, which initiates a new full load each time it is triggered.
ℹ️ The
resume-processing
option isn't applicable for a full-load task, because you can't resume partially loaded tables during the full load phase.
🔗 For more detailed, check to API_StartReplicationTask_RequestParameters
Hi,
When you resume a task, the task doesn't restart, instead it resumes from the checkpoint it last stopped at.
This is done depending on the source, for example, Postgres will read from the replication slot but if the task is restarted, then the replication slot is recreated. Remember, resume won't cause another full load, it will just continue from the maintained checkpoint. [1]
If you resume the task, then only changes that were captured after the last stop point are applied to the database. If the migration task stops during the CDC phase, then AWS DMS maintains the checkpoint information for future use. You can view the task checkpoint in the Overview details tab of the AWS DMS console.
Depending on the CDCLatencyTarget [2] values, we can determine if the changes are still being applied to the target via swap files (to catch up with the source) the target may be trying to catch up with the source or reading from the swap files to do the same. We recommend checking the target latency here, if there is none and the latency matches or is similar to the CDCLatencySource values., then the delay is at the source, not the target.
FullLoadThroughputBandwidthTarget - Outgoing data transmitted from a full load for the target in KB per second.
In our case you are resuming the task "resume-processing" for CDC only, so the full load will not be performed, you can use the CDC metrics instead to verify if the changes are occurring.
If you notice no changes being made at CDC, we recommend increasing the logging severity to verify if any changes are being applied at the log level, should no changes be applied, restarting the task will perform another full load. You can also reach out to AWS support at this point to check the logs and latency.
References:
[1] Restarting and resuming a DMS task: https://repost.aws/knowledge-center/dms-restart-resume-failed-task
[2] Metrics in DMS: https://docs.aws.amazon.com/dms/latest/userguide/CHAP_Monitoring.html
[3] How sources behave depending on resume: https://docs.aws.amazon.com/dms/latest/userguide/CHAP_Source.html
Conteúdo relevante
- AWS OFICIALAtualizada há 2 anos
- AWS OFICIALAtualizada há um ano
In my EventBridge Schedule, I changed the param for task type to 'start-replication' in my settings like so: "StartReplicationTaskType": "start-replication" It was supposed to run last night at 7 pm US Central. The DMS task looks like there was an attempt at a run at that time according to the 'CloudWatch metrics' tab. But the data point is at the bottom of the graph and the pop-up message when I hover over the data point states: "FullLoadThroughputBandwidthTarget 0". And there are no CloudWatch logs tied to this attempt.
Please advise.