Hi,
I observed longer time for Greengrass Core to resume operations on deployment request after updating to GGC v1.7.1 (via OTA).
Before, with GGC v1.6.0, from the runtime log stream (/aws/greengrass/GreengrassSystem/runtime), normally I saw ggc system components restarted once after a group New deployment and normally it took around 4-6 seconds for GGC to resume operations.
I use following messages for tracking from CloudWatch logs:
- Received deploymentId
- Finished executing the group downloading step
- Greengrass Root: /greengrass
- Started all system components
However, after updating to V1.7.1, for new deployment, I often see message 4 appears twice. And it takes longer for GGC to resume operations (~24 seconds). From what I observed, the major time portion falls into shutdown stage, i.e. between message-2 and message-3.
Is it normal behavior for GGC v1.7.1 to have longer time for deployment process?
What is the change in deployment workflow between v1.6.0 and v1.7.1?
Note: GGC runs on EC2 instance.
Thank you
Log: Total 24 seconds from deployment request to resume operations
[2019-03-20T23:43:57.707-07:00][INFO]-Start to process new message from the cloud.
[2019-03-20T23:43:57.707-07:00][INFO]-Received deploymentId ...
...
[2019-03-20T23:44:09.946-07:00][INFO]-Setting up worker process factory
[2019-03-20T23:44:09.946-07:00][INFO]-Found cgroup: pids
...
[2019-03-20T23:44:10.716-07:00][INFO]-Started all system components
[2019-03-20T23:44:10.768-07:00][INFO]-Daemon restarting itself after deployment
[2019-03-20T23:44:21.366-07:00][INFO]-Greengrass Root: /greengrass
...
[2019-03-20T23:44:21.369-07:00][INFO]-Setting up worker process factory
[2019-03-20T23:44:21.369-07:00][INFO]-Found cgroup: pids
...
[2019-03-20T23:44:22.146-07:00][INFO]-Started all system components