Open bsutton opened 3 years ago
When using NamedLock in docker, if a dcli app crashes it can leave a stale namedlock.
The named lock logic uses the processes pid to determine if a lock is not stale.
In docker an app will always get the same pid id (1) so it will always look like it is use.
Consider using the processes start time. to help identify the pid is the one we thought it is.
When using NamedLock in docker, if a dcli app crashes it can leave a stale namedlock.
The named lock logic uses the processes pid to determine if a lock is not stale.
In docker an app will always get the same pid id (1) so it will always look like it is use.
Consider using the processes start time. to help identify the pid is the one we thought it is.