def _run_remote_retention(self, now: datetime):
push_replication_tasks = list(
filter(self._is_push_replication_task,
select_by_class(ReplicationTask, self.tasks)))
local_snapshots_grouped = group_snapshots_by_datasets(
multilist_snapshots(
self.local_shell,
replication_tasks_source_datasets_queries(
push_replication_tasks)))
for transport, replication_tasks in self._transport_for_replication_tasks(
push_replication_tasks):
shell = self._get_shell(transport)
remote_snapshots = multilist_snapshots(
shell,
[(replication_task.target_dataset, replication_task.recursive)
for replication_task in replication_tasks])
remote_snapshots_grouped = group_snapshots_by_datasets(
remote_snapshots)
owners = [
ExecutedReplicationTaskSnapshotOwner(now, replication_task,
local_snapshots_grouped,
remote_snapshots_grouped)
for replication_task in replication_tasks
]
snapshots_to_destroy = calculate_snapshots_to_remove(
owners, remote_snapshots)
logger.info("Retention on transport %r destroying snapshots: %r",
transport, snapshots_to_destroy)
destroy_snapshots(shell, snapshots_to_destroy)
def _run_remote_retention(self, now: datetime):
push_replication_tasks = list(
filter(self._is_push_replication_task,
select_by_class(ReplicationTask, self.tasks)))
local_snapshots_grouped = group_snapshots_by_datasets(
multilist_snapshots(
self.local_shell,
replication_tasks_source_datasets_queries(
push_replication_tasks)))
for transport, replication_tasks in self._transport_for_replication_tasks(
push_replication_tasks):
shell = self._get_retention_shell(transport)
remote_snapshots_queries = [
(replication_task.target_dataset, replication_task.recursive)
for replication_task in replication_tasks
]
try:
# Prevent hanging remote from breaking all the replications
with ShellTimeoutContext(3600):
remote_snapshots = multilist_snapshots(
shell, remote_snapshots_queries)
except Exception as e:
logger.warning(
"Remote retention failed on %r: error listing snapshots: %r",
transport, e)
continue
remote_snapshots_grouped = group_snapshots_by_datasets(
remote_snapshots)
owners = [
ExecutedReplicationTaskSnapshotOwner(now, replication_task,
local_snapshots_grouped,
remote_snapshots_grouped)
for replication_task in replication_tasks
]
snapshots_to_destroy = calculate_snapshots_to_remove(
owners, remote_snapshots)
logger.info("Retention on %r destroying snapshots: %r", transport,
snapshots_to_destroy)
try:
# Prevent hanging remote from breaking all the replications
with ShellTimeoutContext(3600):
destroy_snapshots(shell, snapshots_to_destroy)
except Exception as e:
logger.warning(
"Remote retention failed on %r: error destroying snapshots: %r",
transport, e)
continue
def set_tasks(self, tasks):
self.tasks = tasks
self.scheduler.set_tasks(list(filter(self._is_scheduler_task, tasks)))
def _run_local_retention(self, now: datetime):
periodic_snapshot_tasks = select_by_class(PeriodicSnapshotTask,
self.tasks)
replication_tasks = select_by_class(ReplicationTask, self.tasks)
push_replication_tasks_that_can_hold = [
replication_task for replication_task in replication_tasks
if replication_task.hold_pending_snapshots
]
pull_replications_tasks = list(
filter(self._is_pull_replication_task, replication_tasks))
local_snapshots_queries = []
local_snapshots_queries.extend([
(periodic_snapshot_task.dataset, periodic_snapshot_task.recursive)
for periodic_snapshot_task in periodic_snapshot_tasks
])
local_snapshots_queries.extend(
replication_tasks_source_datasets_queries(
push_replication_tasks_that_can_hold))
local_snapshots_queries.extend([
(replication_task.target_dataset, replication_task.recursive)
for replication_task in pull_replications_tasks
])
local_snapshots = multilist_snapshots(self.local_shell,
local_snapshots_queries)
local_snapshots_grouped = group_snapshots_by_datasets(local_snapshots)
owners = []
owners.extend([
PeriodicSnapshotTaskSnapshotOwner(now, periodic_snapshot_task)
for periodic_snapshot_task in periodic_snapshot_tasks
])
# These are always only PUSH replication tasks
for transport, replication_tasks in self._transport_for_replication_tasks(
push_replication_tasks_that_can_hold):
shell = self._get_retention_shell(transport)
owners.extend(
pending_push_replication_task_snapshot_owners(
local_snapshots_grouped, shell, replication_tasks))
for transport, replication_tasks in self._transport_for_replication_tasks(
pull_replications_tasks):
shell = self._get_retention_shell(transport)
remote_snapshots_queries = replication_tasks_source_datasets_queries(
replication_tasks)
try:
remote_snapshots = multilist_snapshots(
shell, remote_snapshots_queries)
except Exception as e:
logger.warning(
"Local retention failed: error listing snapshots on %r: %r",
transport, e)
return
remote_snapshots_grouped = group_snapshots_by_datasets(
remote_snapshots)
owners.extend([
executed_pull_replication_task_snapshot_owner(
now, replication_task, remote_snapshots_grouped,
local_snapshots_grouped)
for replication_task in replication_tasks
])
snapshots_to_destroy = calculate_snapshots_to_remove(
owners, local_snapshots)
logger.info("Retention destroying local snapshots: %r",
snapshots_to_destroy)
destroy_snapshots(self.local_shell, snapshots_to_destroy)
def _spawn_pending_tasks(self):
for pending_task in list(self.pending_tasks):
if self._can_spawn_replication_task(pending_task):
self._spawn_replication_task(pending_task)
self.pending_tasks.remove(pending_task)