Tried this with Ubuntu 22.04 and using both v2.0.1 and v1.36.0 of efs-utils.
The symptoms are that mounts succeds and works for ~30 seconds until it is terminated by killing stunnel. Here is an excerpt of mount-watchdog.log
root@ip-10-0-1-248:/var/log/amazon/efs# cat mount-watchdog.log
2024-05-20 12:37:54 UTC - INFO - amazon-efs-mount-watchdog, version 1.36.0, is enabled and started
...
2024-05-20 12:38:30 UTC - DEBUG - Current local NFS mounts: [Mount(server='127.0.0.1:/', mountpoint='/.efs', type='nfs4', options='rw,relatime,vers=4.1,rsize=1048576,wsize=1048576,namlen=255,hard,noresvport,proto=tcp,port=20407,timeo=600,retrans=2,sec=sys,clientaddr=127.0.0.1,local_lock=none,addr=127.0.0.1', freq='0', passno='0')]
2024-05-20 12:38:30 UTC - DEBUG - Translating "fs-09d2f41ca8574c530.efs.20407" into mount point and port "efs.20407"
2024-05-20 12:38:30 UTC - DEBUG - Current state files in "/var/run/efs": ['fs-09d2f41ca8574c530.efs.20407']
2024-05-20 12:38:30 UTC - INFO - No mount found for "fs-09d2f41ca8574c530.efs.20407"
2024-05-20 12:38:30 UTC - DEBUG - Marking fs-09d2f41ca8574c530.efs.20407 as unmounted at 1716208710
Tried this with Ubuntu 22.04 and using both v2.0.1 and v1.36.0 of efs-utils.
The symptoms are that mounts succeds and works for ~30 seconds until it is terminated by killing stunnel. Here is an excerpt of mount-watchdog.log
30 seconds later SIGTERM is sent to stunnel
If using /efs as the mount path, it works fine