It is now possible to create a backup of the contents of memfs before
the core is shutdown. Provide the path where to write the files and
a list of glob patterns for files to be written to disk. Use the
pattern "**" for all files. During start of the core the backuped
files will be written back to memfs.
By providing CORE_SESSIONS_SESSION_LOG_PATH_PATTERN (e.g. "/log/%Y-%m-%d.log") all
finished sessions will be logged to a file according to the provided strftime-pattern. The
actual value is calculated from when the session closed. CORE_SESSIONS_PERSIST must
be set. Default: not set.
Set CORE_SESSIONS_SESSION_LOG_BUFFER_SEC to the number of seconds the log should be
buffered in memory before persisted to disk. Default 15 seconds.
Replace CORE_CLUSTER_JOIN_ADDRESS with CORE_CLUSTER_PEERS. This is
a comma separated list of cluster members with their IDs of the form
ID@host:port
On startup the node tries to connect to all the peers. In case of
sudden deaths of a node this will allow to find back into the
cluster. The list in CLUSTER_PEERS is a starting point of known
peers. Other node that are not in that list can still join the
cluster.
File and stream proxy has been moved to the Proxy type.
