Create and start a new thread for handling slot advance queries. +

Reset state of the daemon. +

Thread main loop entrypoint. + Note

  Thread will wait until a sync is scheduled from outside, normally triggered during the HA loop or a wakeup call.
+

Trigger a synchronisation of slots. + This is the main entrypoint for Patroni HA loop wakeup call. +

Execute a pg_replication_slot_advance query and store success for scheduled synchronisation task. +

Synchronise slots for all scheduled databases. +

Synchronise slots for a single database. +

+

Create an instance with storage attributes for replication slots and schedule the first synchronisation. +

Select values from src dictionary to update in dst dictionary for optional supplied keys. +

Compare required slots and configured as permanent slots with those found, dropping extraneous ones. + Note

Slots that are not contained in _slots_ will be dropped. Slots can be filtered out with `+ignore_slots+` configuration.

  Slots that have matching names but do not match attributes in _slots_ will also be dropped.
  +
  Parameters:::
+

Create any missing logical replication slots on the primary. + If the logical slot already exists, copy state information into the replication slots structure stored in the class instance. +

Update logical slots on replicas. + If the logical slot already exists, copy state information into the replication slots structure stored in the class instance. Slots that exist are also advanced if their confirmed_flush_lsn is greater than the stored state of the slot. + As logical slots can only be created when the primary is available, pass the list of slots that need to be copied back to the caller. They will be created on replicas with SlotsHandler.copy_logical_slots(). +

Create or advance physical replication slots. + Any failures are logged and do not interrupt creation of all slots. +

Create a new database connection to the leader. + Note

  Uses rewind user credentials because it has enough permissions to read files from PGDATA. Sets the options `+connect_timeout+` to `+3+` and `+statement_timeout+` to `+2000+`.
  +
  Parameters:::
    *leader* – object with information on the leader
  Yields:::
    connection cursor object, note implementation varies depending on version of `+psycopg+`.
+

Helper method for Postgresql.query(). +

Ready logical slots by comparing primary physical slot catalog_xmin to logical catalog_xmin. + The logical slot on a replica is safe to use when the physical replica slot on the primary: + __\\__

Store pending logical slot information for catalog_xmin on the primary. + Remember catalog_xmin of logical slots on the primary when catalog_xmin of the physical slot became valid. Logical slots on replica will be safe to use after promote when catalog_xmin of the physical slot overtakes these values. +

Determine whether all known logical slots are synchronised from the leader. +

Create logical replication slots on standby nodes. +

Drop a named slot from Postgres. +

Create a new database connection to local server. + Create a non-blocking connection cursor to avoid the situation where an execution of the query of pg_replication_slot_advance takes longer than the timeout on a HA loop, which could cause a false failure state. +

Check if slot name should not be managed by Patroni. +

Query replication slot information from the database and store it for processing by other tasks. + Note

  Only supported from PostgreSQL version 9.4 onwards.
  +
  Store replication slot `+name+`, `+type+`, `+plugin+`, `+database+` and `+datoid+`. If PostgreSQL version is 10 or newer also store `+catalog_xmin+` and `+confirmed_flush_lsn+`.
  +
  When using logical slots, store information separately for slot synchronisation on replica nodes.
+

Entry point from HA cycle used when a standby node is to be promoted to primary. + Note

  If logical replication slot synchronisation is enabled then slot advancement will be triggered. If any logical slots that were copied are yet to be confirmed as ready a warning message will be logged.
+

Process replication slot information from the host and prepare information used in subsequent cluster tasks. + Note

\\__\\__
This methods solves three problems.

The `+cluster_info_query+` from :class:`+Postgresql+` is executed every HA loop and returns information about all replication slots that exists on the current host.

Based on this information perform the following actions:

Schedule the loading of slot information from the database. +

Wrapper to ensure slots advance daemon thread is started if not already. +

During the HA loop read, check and alter replication slots found in the cluster. + Read physical and logical slots from pg_replication_slots, then compare to those configured in the DCS. Drop any slots that do not match those required by configuration and are not configured as permanent. Create any missing physical slots, or advance their position according to feedback stored in DCS. If we are the primary then create logical slots, otherwise if logical slots are known and active create them on replica nodes by copying slot files from the primary. +

If the first argument is a physical replication slot then only the type of the second slot is compared. If the first argument is another type (e.g. logical) then dbid and plugin are compared. +

True if the slot type of s1 and s2 is matches, and the type of s1 is physical, OR the types match AND the dbid and plugin attributes are equal.