Applications use client sessions to interface with server sessions.
Server sessions are available for replica sets and sharded clusters only.
Starting in 3.6, MongoDB drivers associate all operations with a server session, with the exception of unacknowledged writes. The following options are available for all commands to support association with a server session:
mongosh and the drivers assign these options
to the commands in the session.
The document that specifies the unique ID of the session associated with the command. If the
A strictly increasing non-negative number that uniquely identifies the command in the command's session.
If specified, the command must also include the
Do not manually set
stmtIds. MongoDB sets the
to be strictly increasing non-negative numbers.
Array of 32-bit integers
Array of numbers that uniquely identify their respective write operations within the write command.
The following commands can be used to list, manage, and kill server sessions throughout MongoDB clusters:
Expires specified server sessions.
Kills all server sessions.
Kills all server sessions that match the specified pattern.
Kills specified server sessions.
Refreshes idle server sessions.
Starts a new server session.
If the deployment enforces authentication/authorization, the user must be authenticated to start a session, and only that user can use the session.
To use Client Sessions and Causal Consistency Guarantees with
$external authentication users
(Kerberos, LDAP, or x.509 users), usernames cannot be greater
than 10k bytes.
If the deployment does not enforce authentication/authorization, a created session has no owner and can be used by any user on any connection. If a user authenticates and creates a session for a deployment that does not enforce authentication/authorization, that user owns the session. However, any user on any connection may use the session.
If the deployment transitions to authentication without any downtime, any sessions without an owner cannot be used.