class SmartService(SystemServiceService):
class Config:
service = "smartd"
service_model = "smart"
datastore_extend = "smart.smart_extend"
datastore_prefix = "smart_"
@private
async def smart_extend(self, smart):
smart["powermode"] = smart["powermode"].upper()
smart["email"] = smart["email"].split(",")
return smart
@accepts(
Dict('smart_update',
Int('interval'),
Str('powermode', enum=['NEVER', 'SLEEP', 'STANDBY', 'IDLE']),
Int('difference'),
Int('informational'),
Int('critical'),
List('email',
validators=[Unique()],
items=[Str('email', validators=[Email()])]),
update=True))
async def do_update(self, data):
"""
Update SMART Service Configuration.
`interval` is an integer value in minutes which defines how often smartd activates to check if any tests
are configured to run.
`critical`, `informational` and `difference` are integer values on which alerts for SMART are configured if
the disks temperature crosses the assigned threshold for each respective attribute. They default to 0 which
indicates they are disabled.
Email of log level LOG_CRIT is issued when disk temperature crosses `critical`.
Email of log level LOG_INFO is issued when disk temperature crosses `informational`.
If temperature of a disk changes by `difference` degree Celsius since the last report, SMART reports this.
`email` is a list of valid emails to receive SMART alerts.
"""
old = await self.config()
new = old.copy()
new.update(data)
new["powermode"] = new["powermode"].lower()
new["email"] = ",".join([email.strip() for email in new["email"]])
await self._update_service(old, new)
if new["powermode"] != old["powermode"]:
await self.middleware.call("service.restart", "collectd",
{"onetime": False})
await self.smart_extend(new)
return new
class SmartService(SystemServiceService):
class Config:
service = "smartd"
service_model = "smart"
datastore_extend = "smart.smart_extend"
datastore_prefix = "smart_"
@private
async def smart_extend(self, smart):
smart["powermode"] = smart["powermode"].upper()
smart["email"] = smart["email"].split(",")
return smart
@accepts(
Dict('smart_update',
Int('interval'),
Str('powermode', enum=['NEVER', 'SLEEP', 'STANDBY', 'IDLE']),
Int('difference'),
Int('informational'),
Int('critical'),
List('email',
validators=[Unique()],
items=[Str('email', validators=[Email()])]),
update=True))
async def do_update(self, data):
old = await self.config()
new = old.copy()
new.update(data)
new["powermode"] = new["powermode"].lower()
new["email"] = ",".join([email.strip() for email in new["email"]])
await self._update_service(old, new)
if new["powermode"] != old["powermode"]:
await self.middleware.call("service.restart", "collectd",
{"onetime": False})
await self.smart_extend(new)
return new
class ReplicationService(CRUDService):
class Config:
datastore = "storage.replication"
datastore_prefix = "repl_"
datastore_extend = "replication.extend"
datastore_extend_context = "replication.extend_context"
cli_namespace = "task.replication"
@private
async def extend_context(self, rows, extra):
return {
"state": await self.middleware.call("zettarepl.get_state"),
}
@private
async def extend(self, data, context):
data["periodic_snapshot_tasks"] = [{
k.replace("task_", ""): v
for k, v in task.items()
} for task in data["periodic_snapshot_tasks"]]
for task in data["periodic_snapshot_tasks"]:
Cron.convert_db_format_to_schedule(task, begin_end=True)
if data["direction"] == "PUSH":
data["also_include_naming_schema"] = data["naming_schema"]
data["naming_schema"] = []
if data["direction"] == "PULL":
data["also_include_naming_schema"] = []
Cron.convert_db_format_to_schedule(data,
"schedule",
key_prefix="schedule_",
begin_end=True)
Cron.convert_db_format_to_schedule(data,
"restrict_schedule",
key_prefix="restrict_schedule_",
begin_end=True)
if "error" in context["state"]:
data["state"] = context["state"]["error"]
else:
data["state"] = context["state"]["tasks"].get(
f"replication_task_{data['id']}", {
"state": "PENDING",
})
data["job"] = data["state"].pop("job", None)
return data
@private
async def compress(self, data):
if data["direction"] == "PUSH":
data["naming_schema"] = data["also_include_naming_schema"]
del data["also_include_naming_schema"]
Cron.convert_schedule_to_db_format(data,
"schedule",
key_prefix="schedule_",
begin_end=True)
Cron.convert_schedule_to_db_format(data,
"restrict_schedule",
key_prefix="restrict_schedule_",
begin_end=True)
del data["periodic_snapshot_tasks"]
return data
@accepts(
Dict(
"replication_create",
Str("name", required=True),
Str("direction", enum=["PUSH", "PULL"], required=True),
Str("transport",
enum=["SSH", "SSH+NETCAT", "LOCAL"],
required=True),
Int("ssh_credentials", null=True, default=None),
Str("netcat_active_side",
enum=["LOCAL", "REMOTE"],
null=True,
default=None),
Str("netcat_active_side_listen_address", null=True, default=None),
Int("netcat_active_side_port_min",
null=True,
default=None,
validators=[Port()]),
Int("netcat_active_side_port_max",
null=True,
default=None,
validators=[Port()]),
Str("netcat_passive_side_connect_address", null=True,
default=None),
List("source_datasets", items=[Dataset("dataset")], empty=False),
Dataset("target_dataset", required=True),
Bool("recursive", required=True),
List("exclude", items=[Dataset("dataset")]),
Bool("properties", default=True),
List("properties_exclude", items=[Str("property", empty=False)]),
Dict("properties_override", additional_attrs=True),
Bool("replicate", default=False),
Bool("encryption", default=False),
Str("encryption_key", null=True, default=None),
Str("encryption_key_format",
enum=["HEX", "PASSPHRASE"],
null=True,
default=None),
Str("encryption_key_location", null=True, default=None),
List("periodic_snapshot_tasks",
items=[Int("periodic_snapshot_task")],
validators=[Unique()]),
List("naming_schema",
items=[
Str("naming_schema",
validators=[ReplicationSnapshotNamingSchema()])
]),
List("also_include_naming_schema",
items=[
Str("naming_schema",
validators=[ReplicationSnapshotNamingSchema()])
]),
Bool("auto", required=True),
Cron("schedule",
defaults={"minute": "00"},
begin_end=True,
null=True,
default=None),
Cron("restrict_schedule",
defaults={"minute": "00"},
begin_end=True,
null=True,
default=None),
Bool("only_matching_schedule", default=False),
Bool("allow_from_scratch", default=False),
Str("readonly", enum=["SET", "REQUIRE", "IGNORE"], default="SET"),
Bool("hold_pending_snapshots", default=False),
Str("retention_policy",
enum=["SOURCE", "CUSTOM", "NONE"],
required=True),
Int("lifetime_value",
null=True,
default=None,
validators=[Range(min=1)]),
Str("lifetime_unit",
null=True,
default=None,
enum=["HOUR", "DAY", "WEEK", "MONTH", "YEAR"]),
Str("compression",
enum=["LZ4", "PIGZ", "PLZIP"],
null=True,
default=None),
Int("speed_limit",
null=True,
default=None,
validators=[Range(min=1)]),
Bool("large_block", default=True),
Bool("embed", default=False),
Bool("compressed", default=True),
Int("retries", default=5, validators=[Range(min=1)]),
Str("logging_level",
enum=["DEBUG", "INFO", "WARNING", "ERROR"],
null=True,
default=None),
Bool("enabled", default=True),
register=True,
strict=True,
))
async def do_create(self, data):
"""
Create a Replication Task
Create a Replication Task that will push or pull ZFS snapshots to or from remote host..
* `name` specifies a name for replication task
* `direction` specifies whether task will `PUSH` or `PULL` snapshots
* `transport` is a method of snapshots transfer:
* `SSH` transfers snapshots via SSH connection. This method is supported everywhere but does not achieve
great performance
`ssh_credentials` is a required field for this transport (Keychain Credential ID of type `SSH_CREDENTIALS`)
* `SSH+NETCAT` uses unencrypted connection for data transfer. This can only be used in trusted networks
and requires a port (specified by range from `netcat_active_side_port_min` to `netcat_active_side_port_max`)
to be open on `netcat_active_side`
`ssh_credentials` is also required for control connection
* `LOCAL` replicates to or from localhost
* `source_datasets` is a non-empty list of datasets to replicate snapshots from
* `target_dataset` is a dataset to put snapshots into. It must exist on target side
* `recursive` and `exclude` have the same meaning as for Periodic Snapshot Task
* `properties` control whether we should send dataset properties along with snapshots
* `periodic_snapshot_tasks` is a list of periodic snapshot task IDs that are sources of snapshots for this
replication task. Only push replication tasks can be bound to periodic snapshot tasks.
* `naming_schema` is a list of naming schemas for pull replication
* `also_include_naming_schema` is a list of naming schemas for push replication
* `auto` allows replication to run automatically on schedule or after bound periodic snapshot task
* `schedule` is a schedule to run replication task. Only `auto` replication tasks without bound periodic
snapshot tasks can have a schedule
* `restrict_schedule` restricts when replication task with bound periodic snapshot tasks runs. For example,
you can have periodic snapshot tasks that run every 15 minutes, but only run replication task every hour.
* Enabling `only_matching_schedule` will only replicate snapshots that match `schedule` or
`restrict_schedule`
* `allow_from_scratch` will destroy all snapshots on target side and replicate everything from scratch if none
of the snapshots on target side matches source snapshots
* `readonly` controls destination datasets readonly property:
* `SET` will set all destination datasets to readonly=on after finishing the replication
* `REQUIRE` will require all existing destination datasets to have readonly=on property
* `IGNORE` will avoid this kind of behavior
* `hold_pending_snapshots` will prevent source snapshots from being deleted by retention of replication fails
for some reason
* `retention_policy` specifies how to delete old snapshots on target side:
* `SOURCE` deletes snapshots that are absent on source side
* `CUSTOM` deletes snapshots that are older than `lifetime_value` and `lifetime_unit`
* `NONE` does not delete any snapshots
* `compression` compresses SSH stream. Available only for SSH transport
* `speed_limit` limits speed of SSH stream. Available only for SSH transport
* `large_block`, `embed` and `compressed` are various ZFS stream flag documented in `man zfs send`
* `retries` specifies number of retries before considering replication failed
.. examples(websocket)::
:::javascript
{
"id": "6841f242-840a-11e6-a437-00e04d680384",
"msg": "method",
"method": "replication.create",
"params": [{
"name": "Work Backup",
"direction": "PUSH",
"transport": "SSH",
"ssh_credentials": [12],
"source_datasets", ["data/work"],
"target_dataset": "repl/work",
"recursive": true,
"periodic_snapshot_tasks": [5],
"auto": true,
"restrict_schedule": {
"minute": "0",
"hour": "*/2",
"dom": "*",
"month": "*",
"dow": "1,2,3,4,5",
"begin": "09:00",
"end": "18:00"
},
"only_matching_schedule": true,
"retention_policy": "CUSTOM",
"lifetime_value": 1,
"lifetime_unit": "WEEK",
}]
}
"""
verrors = ValidationErrors()
verrors.add_child("replication_create", await self._validate(data))
if verrors:
raise verrors
periodic_snapshot_tasks = data["periodic_snapshot_tasks"]
await self.compress(data)
id = await self.middleware.call(
"datastore.insert", self._config.datastore, data,
{"prefix": self._config.datastore_prefix})
await self._set_periodic_snapshot_tasks(id, periodic_snapshot_tasks)
await self.middleware.call("zettarepl.update_tasks")
return await self._get_instance(id)
@accepts(Int("id"),
Patch(
"replication_create",
"replication_update",
("attr", {
"update": True
}),
))
async def do_update(self, id, data):
"""
Update a Replication Task with specific `id`
See the documentation for `create` method for information on payload contents
.. examples(websocket)::
:::javascript
{
"id": "6841f242-840a-11e6-a437-00e04d680384",
"msg": "method",
"method": "replication.update",
"params": [
7,
{
"name": "Work Backup",
"direction": "PUSH",
"transport": "SSH",
"ssh_credentials": [12],
"source_datasets", ["data/work"],
"target_dataset": "repl/work",
"recursive": true,
"periodic_snapshot_tasks": [5],
"auto": true,
"restrict_schedule": {
"minute": "0",
"hour": "*/2",
"dom": "*",
"month": "*",
"dow": "1,2,3,4,5",
"begin": "09:00",
"end": "18:00"
},
"only_matching_schedule": true,
"retention_policy": "CUSTOM",
"lifetime_value": 1,
"lifetime_unit": "WEEK",
}
]
}
"""
old = await self._get_instance(id)
new = old.copy()
if new["ssh_credentials"]:
new["ssh_credentials"] = new["ssh_credentials"]["id"]
new["periodic_snapshot_tasks"] = [
task["id"] for task in new["periodic_snapshot_tasks"]
]
new.update(data)
verrors = ValidationErrors()
verrors.add_child("replication_update", await self._validate(new, id))
if verrors:
raise verrors
periodic_snapshot_tasks = new["periodic_snapshot_tasks"]
await self.compress(new)
new.pop("state", None)
new.pop("job", None)
await self.middleware.call("datastore.update", self._config.datastore,
id, new,
{'prefix': self._config.datastore_prefix})
await self._set_periodic_snapshot_tasks(id, periodic_snapshot_tasks)
await self.middleware.call("zettarepl.update_tasks")
return await self._get_instance(id)
@accepts(Int("id"))
async def do_delete(self, id):
"""
Delete a Replication Task with specific `id`
.. examples(websocket)::
:::javascript
{
"id": "6841f242-840a-11e6-a437-00e04d680384",
"msg": "method",
"method": "replication.delete",
"params": [
1
]
}
"""
response = await self.middleware.call("datastore.delete",
self._config.datastore, id)
await self.middleware.call("zettarepl.update_tasks")
return response
@item_method
@accepts(Int("id"), Bool("really_run", default=True, hidden=True))
@job(logs=True)
async def run(self, job, id, really_run):
"""
Run Replication Task of `id`.
"""
if really_run:
task = await self._get_instance(id)
if not task["enabled"]:
raise CallError("Task is not enabled")
if task["state"]["state"] == "RUNNING":
raise CallError("Task is already running")
if task["state"]["state"] == "HOLD":
raise CallError("Task is on hold")
await self.middleware.call("zettarepl.run_replication_task", id,
really_run, job)
@accepts(
Patch(
"replication_create",
"replication_run_onetime",
("rm", {
"name": "name"
}),
("rm", {
"name": "auto"
}),
("rm", {
"name": "schedule"
}),
("rm", {
"name": "only_matching_schedule"
}),
("rm", {
"name": "enabled"
}),
), )
@job(logs=True)
async def run_onetime(self, job, data):
"""
Run replication task without creating it.
"""
data["name"] = f"Temporary replication task for job {job.id}"
data["schedule"] = None
data["only_matching_schedule"] = False
data["auto"] = False
data["enabled"] = True
verrors = ValidationErrors()
verrors.add_child("replication_run_onetime", await
self._validate(data))
if verrors:
raise verrors
await self.middleware.call("zettarepl.run_onetime_replication_task",
job, data)
async def _validate(self, data, id=None):
verrors = ValidationErrors()
await self._ensure_unique(verrors, "", "name", data["name"], id)
# Direction
snapshot_tasks = []
if data["direction"] == "PUSH":
e, snapshot_tasks = await self._query_periodic_snapshot_tasks(
data["periodic_snapshot_tasks"])
verrors.add_child("periodic_snapshot_tasks", e)
if data["naming_schema"]:
verrors.add("naming_schema",
"This field has no sense for push replication")
if not snapshot_tasks and not data["also_include_naming_schema"]:
verrors.add(
"periodic_snapshot_tasks",
"You must at least either bind a periodic snapshot task or provide "
"\"Also Include Naming Schema\" for push replication task")
if data["schedule"] is None and data[
"auto"] and not data["periodic_snapshot_tasks"]:
verrors.add(
"auto",
"Push replication that runs automatically must be either "
"bound to a periodic snapshot task or have a schedule")
if data["direction"] == "PULL":
if data["schedule"] is None and data["auto"]:
verrors.add(
"auto",
"Pull replication that runs automatically must have a schedule"
)
if data["periodic_snapshot_tasks"]:
verrors.add(
"periodic_snapshot_tasks",
"Pull replication can't be bound to a periodic snapshot task"
)
if not data["naming_schema"]:
verrors.add("naming_schema",
"Naming schema is required for pull replication")
if data["also_include_naming_schema"]:
verrors.add("also_include_naming_schema",
"This field has no sense for pull replication")
if data["hold_pending_snapshots"]:
verrors.add(
"hold_pending_snapshots",
"Pull replication tasks can't hold pending snapshots because "
"they don't do source retention")
# Transport
if data["transport"] == "SSH+NETCAT":
if data["netcat_active_side"] is None:
verrors.add(
"netcat_active_side",
"You must choose active side for SSH+netcat replication")
if data["netcat_active_side_port_min"] is not None and data[
"netcat_active_side_port_max"] is not None:
if data["netcat_active_side_port_min"] > data[
"netcat_active_side_port_max"]:
verrors.add(
"netcat_active_side_port_max",
"Please specify value greater or equal than netcat_active_side_port_min"
)
if data["compression"] is not None:
verrors.add(
"compression",
"Compression is not supported for SSH+netcat replication")
if data["speed_limit"] is not None:
verrors.add(
"speed_limit",
"Speed limit is not supported for SSH+netcat replication")
else:
if data["netcat_active_side"] is not None:
verrors.add(
"netcat_active_side",
"This field only has sense for SSH+netcat replication")
for k in [
"netcat_active_side_listen_address",
"netcat_active_side_port_min",
"netcat_active_side_port_max",
"netcat_passive_side_connect_address"
]:
if data[k] is not None:
verrors.add(
k,
"This field only has sense for SSH+netcat replication")
if data["transport"] == "LOCAL":
if data["ssh_credentials"] is not None:
verrors.add(
"ssh_credentials",
"Remote credentials have no sense for local replication")
if data["compression"] is not None:
verrors.add("compression",
"Compression has no sense for local replication")
if data["speed_limit"] is not None:
verrors.add("speed_limit",
"Speed limit has no sense for local replication")
else:
if data["ssh_credentials"] is None:
verrors.add(
"ssh_credentials",
"SSH Credentials are required for non-local replication")
else:
try:
await self.middleware.call(
"keychaincredential.get_of_type",
data["ssh_credentials"], "SSH_CREDENTIALS")
except CallError as e:
verrors.add("ssh_credentials", str(e))
# Common for all directions and transports
for i, source_dataset in enumerate(data["source_datasets"]):
for snapshot_task in snapshot_tasks:
if is_child(source_dataset, snapshot_task["dataset"]):
if data["recursive"]:
for exclude in snapshot_task["exclude"]:
if is_child(exclude, source_dataset
) and exclude not in data["exclude"]:
verrors.add(
"exclude",
f"You should exclude {exclude!r} as bound periodic snapshot "
f"task dataset {snapshot_task['dataset']!r} does"
)
else:
if source_dataset in snapshot_task["exclude"]:
verrors.add(
f"source_datasets.{i}",
f"Dataset {source_dataset!r} is excluded by bound "
f"periodic snapshot task for dataset "
f"{snapshot_task['dataset']!r}")
if not data["recursive"] and data["exclude"]:
verrors.add(
"exclude",
"Excluding child datasets is only supported for recursive replication"
)
for i, v in enumerate(data["exclude"]):
if not any(
v.startswith(ds + "/") for ds in data["source_datasets"]):
verrors.add(
f"exclude.{i}",
"This dataset is not a child of any of source datasets")
if data["replicate"]:
if not data["recursive"]:
verrors.add(
"recursive",
"This option is required for full filesystem replication")
if data["exclude"]:
verrors.add(
"exclude",
"This option is not supported for full filesystem replication"
)
if not data["properties"]:
verrors.add(
"properties",
"This option is required for full filesystem replication")
if data["encryption"]:
for k in [
"encryption_key", "encryption_key_format",
"encryption_key_location"
]:
if data[k] is None:
verrors.add(
k,
"This property is required when remote dataset encryption is enabled"
)
if data["schedule"]:
if not data["auto"]:
verrors.add(
"schedule",
"You can't have schedule for replication that does not run automatically"
)
else:
if data["only_matching_schedule"]:
verrors.add(
"only_matching_schedule",
"You can't have only-matching-schedule without schedule")
if data["retention_policy"] == "CUSTOM":
if data["lifetime_value"] is None:
verrors.add(
"lifetime_value",
"This field is required for custom retention policy")
if data["lifetime_unit"] is None:
verrors.add(
"lifetime_value",
"This field is required for custom retention policy")
else:
if data["lifetime_value"] is not None:
verrors.add(
"lifetime_value",
"This field has no sense for specified retention policy")
if data["lifetime_unit"] is not None:
verrors.add(
"lifetime_unit",
"This field has no sense for specified retention policy")
if data["enabled"]:
for i, snapshot_task in enumerate(snapshot_tasks):
if not snapshot_task["enabled"]:
verrors.add(
f"periodic_snapshot_tasks.{i}",
"You can't bind disabled periodic snapshot task to enabled replication task"
)
return verrors
async def _set_periodic_snapshot_tasks(self, replication_task_id,
periodic_snapshot_tasks_ids):
await self.middleware.call(
"datastore.delete",
"storage.replication_repl_periodic_snapshot_tasks",
[["replication_id", "=", replication_task_id]])
for periodic_snapshot_task_id in periodic_snapshot_tasks_ids:
await self.middleware.call(
"datastore.insert",
"storage.replication_repl_periodic_snapshot_tasks",
{
"replication_id": replication_task_id,
"task_id": periodic_snapshot_task_id,
},
)
async def _query_periodic_snapshot_tasks(self, ids):
verrors = ValidationErrors()
query_result = await self.middleware.call("pool.snapshottask.query",
[["id", "in", ids]])
snapshot_tasks = []
for i, task_id in enumerate(ids):
for task in query_result:
if task["id"] == task_id:
snapshot_tasks.append(task)
break
else:
verrors.add(str(i), "This snapshot task does not exist")
return verrors, snapshot_tasks
@accepts(
Str("transport", enum=["SSH", "SSH+NETCAT", "LOCAL"], required=True),
Int("ssh_credentials", null=True, default=None))
async def list_datasets(self, transport, ssh_credentials):
"""
List datasets on remote side
Accepts `transport` and SSH credentials ID (for non-local transport)
.. examples(websocket)::
:::javascript
{
"id": "6841f242-840a-11e6-a437-00e04d680384",
"msg": "method",
"method": "replication.list_datasets",
"params": [
"SSH",
7
]
}
"""
return await self.middleware.call("zettarepl.list_datasets", transport,
ssh_credentials)
@accepts(Str("dataset", required=True),
Str("transport",
enum=["SSH", "SSH+NETCAT", "LOCAL"],
required=True), Int("ssh_credentials",
null=True,
default=None))
async def create_dataset(self, dataset, transport, ssh_credentials):
"""
Creates dataset on remote side
Accepts `dataset` name, `transport` and SSH credentials ID (for non-local transport)
.. examples(websocket)::
:::javascript
{
"id": "6841f242-840a-11e6-a437-00e04d680384",
"msg": "method",
"method": "replication.create_dataset",
"params": [
"repl/work",
"SSH",
7
]
}
"""
return await self.middleware.call("zettarepl.create_dataset", dataset,
transport, ssh_credentials)
@accepts()
async def list_naming_schemas(self):
"""
List all naming schemas used in periodic snapshot and replication tasks.
"""
naming_schemas = []
for snapshottask in await self.middleware.call(
"pool.snapshottask.query"):
naming_schemas.append(snapshottask["naming_schema"])
for replication in await self.middleware.call("replication.query"):
naming_schemas.extend(replication["naming_schema"])
naming_schemas.extend(replication["also_include_naming_schema"])
return sorted(set(naming_schemas))
@accepts(
List("datasets", empty=False, items=[Dataset("dataset")]),
List("naming_schema",
empty=False,
items=[
Str("naming_schema",
validators=[ReplicationSnapshotNamingSchema()])
]),
Str("transport", enum=["SSH", "SSH+NETCAT", "LOCAL"], required=True),
Int("ssh_credentials", null=True, default=None),
)
async def count_eligible_manual_snapshots(self, datasets, naming_schema,
transport, ssh_credentials):
"""
Count how many existing snapshots of `dataset` match `naming_schema`.
.. examples(websocket)::
:::javascript
{
"id": "6841f242-840a-11e6-a437-00e04d680384",
"msg": "method",
"method": "replication.count_eligible_manual_snapshots",
"params": [
"repl/work",
["auto-%Y-%m-%d_%H-%M"],
"SSH",
4
]
}
"""
return await self.middleware.call(
"zettarepl.count_eligible_manual_snapshots", datasets,
naming_schema, transport, ssh_credentials)
@accepts(
Str("direction", enum=["PUSH", "PULL"], required=True),
List("source_datasets",
items=[Dataset("dataset")],
required=True,
empty=False),
Dataset("target_dataset", required=True),
Str("transport",
enum=["SSH", "SSH+NETCAT", "LOCAL", "LEGACY"],
required=True),
Int("ssh_credentials", null=True, default=None),
)
async def target_unmatched_snapshots(self, direction, source_datasets,
target_dataset, transport,
ssh_credentials):
"""
Check if target has any snapshots that do not exist on source.
.. examples(websocket)::
:::javascript
{
"id": "6841f242-840a-11e6-a437-00e04d680384",
"msg": "method",
"method": "replication.target_unmatched_snapshots",
"params": [
"PUSH",
["repl/work", "repl/games"],
"backup",
"SSH",
4
]
}
Returns
{
"backup/work": ["auto-2019-10-15_13-00", "auto-2019-10-15_09-00"],
"backup/games": ["auto-2019-10-15_13-00"],
}
"""
return await self.middleware.call(
"zettarepl.target_unmatched_snapshots", direction, source_datasets,
target_dataset, transport, ssh_credentials)
@private
def new_snapshot_name(self, naming_schema):
return datetime.now().strftime(naming_schema)
# Legacy pair support
@private
@accepts(
Dict(
"replication-pair-data",
Str("hostname", required=True),
Str("public-key", required=True),
Str("user", null=True),
))
async def pair(self, data):
result = await self.middleware.call(
"keychaincredential.ssh_pair", {
"remote_hostname": data["hostname"],
"username": data["user"] or "root",
"public_key": data["public-key"],
})
return {
"ssh_port": result["port"],
"ssh_hostkey": result["host_key"],
}
class ReplicationService(CRUDService):
class Config:
datastore = "storage.replication"
datastore_prefix = "repl_"
datastore_extend = "replication.extend"
datastore_extend_context = "replication.extend_context"
@private
async def extend_context(self):
legacy_result, legacy_result_datetime = await self.middleware.run_in_thread(
self._legacy_extend_context)
return {
"state": await self.middleware.call("zettarepl.get_state"),
"legacy_result": legacy_result,
"legacy_result_datetime": legacy_result_datetime,
}
def _legacy_extend_context(self):
try:
with open("/tmp/.repl-result", "rb") as f:
data = f.read()
legacy_result = pickle.loads(data)
legacy_result_datetime = datetime.fromtimestamp(
os.stat("/tmp/.repl-result").st_mtime)
except Exception:
legacy_result = defaultdict(dict)
legacy_result_datetime = None
return legacy_result, legacy_result_datetime
@private
async def extend(self, data, context):
data["periodic_snapshot_tasks"] = [{
k.replace("task_", ""): v
for k, v in task.items()
} for task in data["periodic_snapshot_tasks"]]
for task in data["periodic_snapshot_tasks"]:
Cron.convert_db_format_to_schedule(task, begin_end=True)
if data["direction"] == "PUSH":
data["also_include_naming_schema"] = data["naming_schema"]
data["naming_schema"] = []
if data["direction"] == "PULL":
data["also_include_naming_schema"] = []
Cron.convert_db_format_to_schedule(data,
"schedule",
key_prefix="schedule_",
begin_end=True)
Cron.convert_db_format_to_schedule(data,
"restrict_schedule",
key_prefix="restrict_schedule_",
begin_end=True)
if data["transport"] == 'LEGACY':
if data["id"] in context["legacy_result"]:
legacy_result = context["legacy_result"][data["id"]]
msg = legacy_result.get("msg")
if msg == "Running":
state = "RUNNING"
elif msg in ["Succeeded", "Up to date"]:
state = "FINISHED"
else:
state = "ERROR"
data["state"] = {
"datetime": context["legacy_result_datetime"],
"state": state,
"last_snapshot": legacy_result.get("last_snapshot"),
}
if state == "ERROR":
data["state"]["error"] = msg
else:
data["state"] = {
"state": "PENDING",
}
else:
data["state"] = context["state"].get(
f"replication_task_{data['id']}", {
"state": "PENDING",
})
return data
@private
async def compress(self, data):
if data["direction"] == "PUSH":
data["naming_schema"] = data["also_include_naming_schema"]
del data["also_include_naming_schema"]
Cron.convert_schedule_to_db_format(data,
"schedule",
key_prefix="schedule_",
begin_end=True)
Cron.convert_schedule_to_db_format(data,
"restrict_schedule",
key_prefix="restrict_schedule_",
begin_end=True)
del data["periodic_snapshot_tasks"]
return data
@accepts(
Dict(
"replication_create",
Str("name", required=True),
Str("direction", enum=["PUSH", "PULL"], required=True),
Str("transport",
enum=["SSH", "SSH+NETCAT", "LOCAL", "LEGACY"],
required=True),
Int("ssh_credentials", null=True, default=None),
Str("netcat_active_side",
enum=["LOCAL", "REMOTE"],
null=True,
default=None),
Str("netcat_active_side_listen_address", null=True, default=None),
Int("netcat_active_side_port_min",
null=True,
default=None,
validators=[Port()]),
Int("netcat_active_side_port_max",
null=True,
default=None,
validators=[Port()]),
Str("netcat_passive_side_connect_address", null=True,
default=None),
List("source_datasets",
items=[Path("dataset", empty=False)],
required=True,
empty=False),
Path("target_dataset", required=True, empty=False),
Bool("recursive", required=True),
List("exclude", items=[Path("dataset", empty=False)], default=[]),
List("periodic_snapshot_tasks",
items=[Int("periodic_snapshot_task")],
default=[],
validators=[Unique()]),
List("naming_schema",
items=[
Str("naming_schema",
validators=[ReplicationSnapshotNamingSchema()])
],
default=[]),
List("also_include_naming_schema",
items=[
Str("naming_schema",
validators=[ReplicationSnapshotNamingSchema()])
],
default=[]),
Bool("auto", required=True),
Cron("schedule",
defaults={"minute": "00"},
begin_end=True,
null=True,
default=None),
Cron("restrict_schedule",
defaults={"minute": "00"},
begin_end=True,
null=True,
default=None),
Bool("only_matching_schedule", default=False),
Bool("allow_from_scratch", default=False),
Bool("hold_pending_snapshots", default=False),
Str("retention_policy",
enum=["SOURCE", "CUSTOM", "NONE"],
required=True),
Int("lifetime_value",
null=True,
default=None,
validators=[Range(min=1)]),
Str("lifetime_unit",
null=True,
default=None,
enum=["HOUR", "DAY", "WEEK", "MONTH", "YEAR"]),
Str("compression",
enum=["LZ4", "PIGZ", "PLZIP"],
null=True,
default=None),
Int("speed_limit",
null=True,
default=None,
validators=[Range(min=1)]),
Bool("dedup", default=True),
Bool("large_block", default=True),
Bool("embed", default=True),
Bool("compressed", default=True),
Int("retries", default=5, validators=[Range(min=1)]),
Str("logging_level",
enum=["DEBUG", "INFO", "WARNING", "ERROR"],
null=True,
default=None),
Bool("enabled", default=True),
register=True,
strict=True,
))
async def do_create(self, data):
"""
Create a Replication Task
Create a Replication Task that will push or pull ZFS snapshots to or from remote host..
* `name` specifies a name for replication task
* `direction` specifies whether task will `PUSH` or `PULL` snapshots
* `transport` is a method of snapshots transfer:
* `SSH` transfers snapshots via SSH connection. This method is supported everywhere but does not achieve
great performance
`ssh_credentials` is a required field for this transport (Keychain Credential ID of type `SSH_CREDENTIALS`)
* `SSH+NETCAT` uses unencrypted connection for data transfer. This can only be used in trusted networks
and requires a port (specified by range from `netcat_active_side_port_min` to `netcat_active_side_port_max`)
to be open on `netcat_active_side`
`ssh_credentials` is also required for control connection
* `LOCAL` replicates to or from localhost
* `LEGACY` uses legacy replication engine prior to FreeNAS 11.3
* `source_datasets` is a non-empty list of datasets to replicate snapshots from
* `target_dataset` is a dataset to put snapshots into. It must exist on target side
* `recursive` and `exclude` have the same meaning as for Periodic Snapshot Task
* `periodic_snapshot_tasks` is a list of periodic snapshot task IDs that are sources of snapshots for this
replication task. Only push replication tasks can be bound to periodic snapshot tasks.
* `naming_schema` is a list of naming schemas for pull replication
* `also_include_naming_schema` is a list of naming schemas for push replication
* `auto` allows replication to run automatically on schedule or after bound periodic snapshot task
* `schedule` is a schedule to run replication task. Only `auto` replication tasks without bound periodic
snapshot tasks can have a schedule
* `restrict_schedule` restricts when replication task with bound periodic snapshot tasks runs. For example,
you can have periodic snapshot tasks that run every 15 minutes, but only run replication task every hour.
* Enabling `only_matching_schedule` will only replicate snapshots that match `schedule` or
`restrict_schedule`
* `allow_from_scratch` will destroy all snapshots on target side and replicate everything from scratch if none
of the snapshots on target side matches source snapshots
* `hold_pending_snapshots` will prevent source snapshots from being deleted by retention of replication fails
for some reason
* `retention_policy` specifies how to delete old snapshots on target side:
* `SOURCE` deletes snapshots that are absent on source side
* `CUSTOM` deletes snapshots that are older than `lifetime_value` and `lifetime_unit`
* `NONE` does not delete any snapshots
* `compression` compresses SSH stream. Available only for SSH transport
* `speed_limit` limits speed of SSH stream. Available only for SSH transport
* `dedup`, `large_block`, `embed` and `compressed` are various ZFS stream flag documented in `man zfs send`
* `retries` specifies number of retries before considering replication failed
.. examples(websocket)::
:::javascript
{
"id": "6841f242-840a-11e6-a437-00e04d680384",
"msg": "method",
"method": "replication.create",
"params": [{
"name": "Work Backup",
"direction": "PUSH",
"transport": "SSH",
"ssh_credentials": [12],
"source_datasets", ["data/work"],
"target_dataset": "repl/work",
"recursive": true,
"periodic_snapshot_tasks": [5],
"auto": true,
"restrict_schedule": {
"minute": "0",
"hour": "*/2",
"dom": "*",
"month": "*",
"dow": "1,2,3,4,5",
"begin": "09:00",
"end": "18:00"
},
"only_matching_schedule": true,
"retention_policy": "CUSTOM",
"lifetime_value": 1,
"lifetime_unit": "WEEK",
}]
}
"""
verrors = ValidationErrors()
verrors.add_child("replication_create", await self._validate(data))
if verrors:
raise verrors
periodic_snapshot_tasks = data["periodic_snapshot_tasks"]
await self.compress(data)
id = await self.middleware.call(
"datastore.insert", self._config.datastore, data,
{"prefix": self._config.datastore_prefix})
await self._set_periodic_snapshot_tasks(id, periodic_snapshot_tasks)
await self.middleware.call("service.restart", "cron")
await self.middleware.call("zettarepl.update_tasks")
return await self._get_instance(id)
@accepts(Int("id"),
Patch(
"replication_create",
"replication_update",
("attr", {
"update": True
}),
))
async def do_update(self, id, data):
"""
Update a Replication Task with specific `id`
See the documentation for `create` method for information on payload contents
.. examples(websocket)::
:::javascript
{
"id": "6841f242-840a-11e6-a437-00e04d680384",
"msg": "method",
"method": "replication.update",
"params": [
7,
{
"name": "Work Backup",
"direction": "PUSH",
"transport": "SSH",
"ssh_credentials": [12],
"source_datasets", ["data/work"],
"target_dataset": "repl/work",
"recursive": true,
"periodic_snapshot_tasks": [5],
"auto": true,
"restrict_schedule": {
"minute": "0",
"hour": "*/2",
"dom": "*",
"month": "*",
"dow": "1,2,3,4,5",
"begin": "09:00",
"end": "18:00"
},
"only_matching_schedule": true,
"retention_policy": "CUSTOM",
"lifetime_value": 1,
"lifetime_unit": "WEEK",
}
]
}
"""
old = await self._get_instance(id)
new = old.copy()
if new["ssh_credentials"]:
new["ssh_credentials"] = new["ssh_credentials"]["id"]
new["periodic_snapshot_tasks"] = [
task["id"] for task in new["periodic_snapshot_tasks"]
]
new.update(data)
verrors = ValidationErrors()
verrors.add_child("replication_update", await self._validate(new))
if verrors:
raise verrors
periodic_snapshot_tasks = new["periodic_snapshot_tasks"]
await self.compress(new)
new.pop('state', None)
await self.middleware.call("datastore.update", self._config.datastore,
id, new,
{'prefix': self._config.datastore_prefix})
await self._set_periodic_snapshot_tasks(id, periodic_snapshot_tasks)
await self.middleware.call("service.restart", "cron")
await self.middleware.call("zettarepl.update_tasks")
return await self._get_instance(id)
@accepts(Int("id"))
async def do_delete(self, id):
"""
Delete a Replication Task with specific `id`
.. examples(websocket)::
:::javascript
{
"id": "6841f242-840a-11e6-a437-00e04d680384",
"msg": "method",
"method": "replication.delete",
"params": [
1
]
}
"""
response = await self.middleware.call("datastore.delete",
self._config.datastore, id)
await self.middleware.call("service.restart", "cron")
await self.middleware.call("zettarepl.update_tasks")
return response
@item_method
@accepts(Int("id"))
async def run(self, id):
"""
Run Replication Task of `id`.
"""
task = await self._get_instance(id)
if not task["enabled"]:
raise CallError("Task is not enabled")
await self.middleware.call("zettarepl.run_replication_task",
task["id"])
async def _validate(self, data):
verrors = ValidationErrors()
# Direction
snapshot_tasks = []
if data["direction"] == "PUSH":
e, snapshot_tasks = await self._query_periodic_snapshot_tasks(
data["periodic_snapshot_tasks"])
verrors.add_child("periodic_snapshot_tasks", e)
if data["naming_schema"]:
verrors.add("naming_schema",
"This field has no sense for push replication")
if data["schedule"]:
if data["periodic_snapshot_tasks"]:
verrors.add(
"schedule",
"Push replication can't be bound to periodic snapshot task and have "
"schedule at the same time")
else:
if data["auto"] and not data[
"periodic_snapshot_tasks"] and data[
"transport"] != "LEGACY":
verrors.add(
"auto",
"Push replication that runs automatically must be either "
"bound to periodic snapshot task or have schedule")
if data["direction"] == "PULL":
if data["schedule"]:
pass
else:
if data["auto"]:
verrors.add(
"auto",
"Pull replication that runs automatically must have schedule"
)
if data["periodic_snapshot_tasks"]:
verrors.add(
"periodic_snapshot_tasks",
"Pull replication can't be bound to periodic snapshot task"
)
if not data["naming_schema"]:
verrors.add("naming_schema",
"Naming schema is required for pull replication")
if data["also_include_naming_schema"]:
verrors.add("also_include_naming_schema",
"This field has no sense for pull replication")
if data["hold_pending_snapshots"]:
verrors.add(
"hold_pending_snapshots",
"Pull replication tasks can't hold pending snapshots because "
"they don't do source retention")
# Transport
if data["transport"] == "SSH+NETCAT":
if data["netcat_active_side"] is None:
verrors.add(
"netcat_active_side",
"You must choose active side for SSH+netcat replication")
if data["netcat_active_side_port_min"] is not None and data[
"netcat_active_side_port_max"] is not None:
if data["netcat_active_side_port_min"] > data[
"netcat_active_side_port_max"]:
verrors.add(
"netcat_active_side_port_max",
"Please specify value greater or equal than netcat_active_side_port_min"
)
if data["compression"] is not None:
verrors.add(
"compression",
"Compression is not supported for SSH+netcat replication")
if data["speed_limit"] is not None:
verrors.add(
"speed_limit",
"Speed limit is not supported for SSH+netcat replication")
else:
if data["netcat_active_side"] is not None:
verrors.add(
"netcat_active_side",
"This field only has sense for SSH+netcat replication")
for k in [
"netcat_active_side_listen_address",
"netcat_active_side_port_min",
"netcat_active_side_port_max",
"netcat_passive_side_connect_address"
]:
if data[k] is not None:
verrors.add(
k,
"This field only has sense for SSH+netcat replication")
if data["transport"] == "LOCAL":
if data["ssh_credentials"] is not None:
verrors.add(
"ssh_credentials",
"Remote credentials have no sense for local replication")
if data["compression"] is not None:
verrors.add("compression",
"Compression has no sense for local replication")
if data["speed_limit"] is not None:
verrors.add("speed_limit",
"Speed limit has no sense for local replication")
else:
if data["ssh_credentials"] is None:
verrors.add(
"ssh_credentials",
"SSH Credentials are required for non-local replication")
else:
try:
await self.middleware.call(
"keychaincredential.get_of_type",
data["ssh_credentials"], "SSH_CREDENTIALS")
except CallError as e:
verrors.add("ssh_credentials", str(e))
if data["transport"] == "LEGACY":
for should_be_true in ["auto", "allow_from_scratch"]:
if not data[should_be_true]:
verrors.add(
should_be_true,
"Legacy replication does not support disabling this option"
)
for should_be_false in [
"exclude", "periodic_snapshot_tasks", "naming_schema",
"also_include_naming_schema", "only_matching_schedule",
"dedup", "large_block", "embed", "compressed"
]:
if data[should_be_false]:
verrors.add(
should_be_false,
"Legacy replication does not support this option")
if data["direction"] != "PUSH":
verrors.add(
"direction",
"Only push application is allowed for Legacy transport")
if len(data["source_datasets"]) != 1:
verrors.add(
"source_datasets",
"You can only have one source dataset for legacy replication"
)
if data["retries"] != 1:
verrors.add("retries",
"This value should be 1 for legacy replication")
# Common for all directions and transports
for i, source_dataset in enumerate(data["source_datasets"]):
for snapshot_task in snapshot_tasks:
if is_child(source_dataset, snapshot_task["dataset"]):
if data["recursive"]:
for exclude in snapshot_task["exclude"]:
if exclude not in data["exclude"]:
verrors.add(
"exclude",
f"You should exclude {exclude!r} as bound periodic snapshot "
f"task dataset {snapshot_task['dataset']!r} does"
)
else:
if source_dataset in snapshot_task["exclude"]:
verrors.add(
f"source_datasets.{i}",
f"Dataset {source_dataset!r} is excluded by bound "
f"periodic snapshot task for dataset "
f"{snapshot_task['dataset']!r}")
if not data["recursive"] and data["exclude"]:
verrors.add(
"exclude",
"Excluding child datasets is only supported for recursive replication"
)
for i, v in enumerate(data["exclude"]):
if not any(
v.startswith(ds + "/") for ds in data["source_datasets"]):
verrors.add(
f"exclude.{i}",
"This dataset is not a child of any of source datasets")
if data["schedule"]:
if not data["auto"]:
verrors.add(
"schedule",
"You can't have schedule for replication that does not run automatically"
)
else:
if data["only_matching_schedule"]:
verrors.add(
"only_matching_schedule",
"You can't have only-matching-schedule without schedule")
if data["retention_policy"] == "CUSTOM":
if data["lifetime_value"] is None:
verrors.add(
"lifetime_value",
"This field is required for custom retention policy")
if data["lifetime_unit"] is None:
verrors.add(
"lifetime_value",
"This field is required for custom retention policy")
else:
if data["lifetime_value"] is not None:
verrors.add(
"lifetime_value",
"This field has no sense for specified retention policy")
if data["lifetime_unit"] is not None:
verrors.add(
"lifetime_unit",
"This field has no sense for specified retention policy")
if data["enabled"]:
for i, snapshot_task in enumerate(snapshot_tasks):
if not snapshot_task["enabled"]:
verrors.add(
f"periodic_snapshot_tasks.{i}",
"You can't bind disabled periodic snapshot task to enabled replication task"
)
return verrors
async def _set_periodic_snapshot_tasks(self, replication_task_id,
periodic_snapshot_tasks_ids):
await self.middleware.call(
"datastore.delete",
"storage.replication_repl_periodic_snapshot_tasks",
[["replication_id", "=", replication_task_id]])
for periodic_snapshot_task_id in periodic_snapshot_tasks_ids:
await self.middleware.call(
"datastore.insert",
"storage.replication_repl_periodic_snapshot_tasks",
{
"replication_id": replication_task_id,
"task_id": periodic_snapshot_task_id,
},
)
async def _query_periodic_snapshot_tasks(self, ids):
verrors = ValidationErrors()
query_result = await self.middleware.call("pool.snapshottask.query",
[["id", "in", ids]])
snapshot_tasks = []
for i, task_id in enumerate(ids):
for task in query_result:
if task["id"] == task_id:
snapshot_tasks.append(task)
break
else:
verrors.add(str(i), "This snapshot task does not exist")
return verrors, snapshot_tasks
@accepts(
Str("transport",
enum=["SSH", "SSH+NETCAT", "LOCAL", "LEGACY"],
required=True), Int("ssh_credentials", null=True, default=None))
async def list_datasets(self, transport, ssh_credentials=None):
"""
List datasets on remote side
Accepts `transport` and SSH credentials ID (for non-local transport)
.. examples(websocket)::
:::javascript
{
"id": "6841f242-840a-11e6-a437-00e04d680384",
"msg": "method",
"method": "replication.list_datasets",
"params": [
"SSH",
7
]
}
"""
return await self.middleware.call("zettarepl.list_datasets", transport,
ssh_credentials)
@accepts(Str("dataset", required=True),
Str("transport",
enum=["SSH", "SSH+NETCAT", "LOCAL", "LEGACY"],
required=True), Int("ssh_credentials",
null=True,
default=None))
async def create_dataset(self, dataset, transport, ssh_credentials=None):
"""
Creates dataset on remote side
Accepts `dataset` name, `transport` and SSH credentials ID (for non-local transport)
.. examples(websocket)::
:::javascript
{
"id": "6841f242-840a-11e6-a437-00e04d680384",
"msg": "method",
"method": "replication.create_dataset",
"params": [
"repl/work",
"SSH",
7
]
}
"""
return await self.middleware.call("zettarepl.create_dataset", dataset,
transport, ssh_credentials)
# Legacy pair support
@private
@accepts(
Dict(
"replication-pair-data",
Str("hostname", required=True),
Str("public-key", required=True),
Str("user"),
))
async def pair(self, data):
result = await self.middleware.call(
"keychaincredential.ssh_pair", {
"remote_hostname": data["hostname"],
"username": data["user"],
"public_key": data["public-key"],
})
return {
"ssh_port": result["port"],
"ssh_hostkey": result["host_key"],
}
class NetDataService(SystemServiceService):
class Config:
service = 'netdata'
service_model = 'netdataglobalsettings'
service_verb = 'restart'
datastore_extend = 'netdata.netdata_extend'
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self._alarms = {}
self._initialize_alarms()
@private
async def netdata_extend(self, data):
# We get data alarms as a dict e.g
# {"alarms": {"alarm1": {"enabled": True}, "alarm2": {"enabled": True}}}
alarms = copy.deepcopy(self._alarms)
alarms.update(data['alarms'])
data['alarms'] = alarms
for alarm in data['alarms']:
# Remove conf file paths
data['alarms'][alarm].pop('path', None)
return data
@private
async def list_alarms(self):
alarms = copy.deepcopy(self._alarms)
config = await self.config()
for alarm in config['alarms']:
if alarm not in alarms:
# An unlikely case when a previously configured alarm does not exist in conf files anymore
alarms[alarm] = {}
alarms[alarm]['enabled'] = config['alarms'][alarm]['enabled']
return alarms
@private
def _initialize_alarms(self):
path = '/usr/local/etc/netdata/health.d/'
pattern = re.compile(r'alarm: +(.*)(?:[\s\S]*?os: +(.*)\n)?')
for file in [f for f in os.listdir(path) if 'sample' not in f]:
with open(path + file, 'r') as f:
for alarm in re.findall(pattern, f.read()):
# By default all alarms are enabled in netdata
# When we list alarms, alarms which have been configured by user to be disabled
# will show up as disabled only
if 'freebsd' in alarm[1] or not alarm[1]:
self._alarms[alarm[0].strip()] = {
'path': path + file,
'enabled': True
}
@private
async def validate_attrs(self, data):
verrors = ValidationErrors()
additional_params = data.get('additional_params')
if additional_params:
# Let's be very generic here and introduce very basic validation
# Expected format is as following
# [ipv6.icmpneighbor]
# history = 86400
# enabled = yes
#
# While we are here, we will also introduce basic formatting to the file to ensure
# that we can make it as compliable as possible
param_str = ''
for i in additional_params.split('\n'):
i = i.strip()
if not i:
continue
if i.startswith('#'):
# Let's not validate this
if i.replace('#', '').startswith('['):
param_str += f'\n\n{i}'
else:
param_str += f'\n\t{i}'
continue
if i.startswith('[') and not i.endswith(']'):
verrors.add(
'netdata_update.additional_params',
f'Please correct format for {i}. i.e [system.intr]')
elif not i.startswith('[') and '=' not in i:
verrors.add(
'netdata_update.additional_params',
f'Please correct format for {i}. i.e enabled = yes')
if i.startswith('['):
param_str += f'\n\n{i}'
else:
param_str += f'\n\t{i}'
data['additional_params'] = param_str + '\n'
bind_to_ips = data.get('bind')
if bind_to_ips:
valid_ips = [
ip['address']
for ip in await self.middleware.call('interfaces.ip_in_use')
]
valid_ips.extend(['127.0.0.1', '::1', '0.0.0.0', '::'])
for bind_ip in bind_to_ips:
if bind_ip not in valid_ips:
verrors.add('netdata_update.bind',
f'Invalid {bind_ip} bind IP')
else:
verrors.add('netdata_update.bind', 'This field is required')
update_alarms = data.pop('update_alarms', {})
valid_alarms = self._alarms
if update_alarms:
for alarm in update_alarms:
if alarm not in valid_alarms:
verrors.add('netdata_update.alarms',
f'{alarm} not a valid alarm')
verrors.extend(
validate_attributes([
Dict(key, Bool('enabled', required=True))
for key in update_alarms
], {'attributes': update_alarms}))
# Validating streaming metrics now
stream_mode = data.get('stream_mode')
if stream_mode == 'SLAVE':
for key in ('api_key', 'destination'):
if not data.get(key):
verrors.add(
f'netdata_update.{key}',
f'{key} is required with stream mode as SLAVE')
destinations = data.get('destination')
if destinations:
ip_addr = IpAddress()
port = Port()
for dest in destinations:
ip = dest.split(':')[0]
try:
ip_addr(ip)
except ValueError as e:
verrors.add('netdata_update.destination', str(e))
else:
if ':' in dest:
try:
port(dest.split(':')[1])
except ValueError as e:
verrors.add('netdata_update.destination',
f'Not a valid port: {e}')
elif stream_mode == 'MASTER':
for key in ('allow_from', 'api_key'):
if not data.get(key):
verrors.add(
f'netdata_update.{key}',
f'{key} is required with stream mode as MASTER')
verrors.check()
data['alarms'].update(update_alarms)
return data
@accepts(
Dict('netdata_update',
Str('additional_params'),
Dict('alarms', additional_attrs=True),
List('allow_from', items=[Str('pattern')]),
Str('api_key', validators=[UUID()]),
List('bind', validators=[Unique()], items=[Str('bind_ip')]),
Int('port', validators=[Port()]),
List('destination',
validators=[Unique()],
items=[Str('destination')]),
Int('history'),
Int('http_port_listen_backlog'),
Str('stream_mode', enum=['NONE', 'MASTER', 'SLAVE']),
Int('update_every'),
update=True))
async def do_update(self, data):
"""
Update Netdata Service Configuration
`alarms` is a dictionary where user specifies a key,value pair with key being alarm name and value is a
dictionary which is of the schema "{'enabled': True}" indicating should the alarm be enabled or not.
Middlewared supports interacting (changing) alarms in /usr/local/etc/netdata/health.d/ directory.
`allow_from` is used when netdata service is expected to be used as a master. It defaults to "['*']". This field
expects a list of Netdata patterns which Netdata will use to set restrictions on incoming connections from slave
accordingly.
`api_key` is a valid UUID which can be generated in command line by typing uuidgen.
`destination` is used when netdata service is expected to be used as a slave. Destination is a list of potential
destinations to which netdata should stream metrics. We expect the format to be IP:PORT ( port is optional ).
The first working destination is used by Netdata service.
`history` is the number of entries the netdata daemon will by default keep in memory for each chart dimension.
It defaults to 86400.
.. examples(websocket)::
Update Netdata Service Configuration
:::javascript
{
"id": "6841f242-840a-11e6-a437-00e04d680384",
"msg": "method",
"method": "netdata.update",
"params": [{
"history": 80000,
"alarms": {
"used_swap": {"enabled": true},
"ram_in_swap": {"enabled": true}
}
}]
}
"""
old = await self.config()
new = old.copy()
# We separate alarms we have in db and the ones user supplies
new['update_alarms'] = data.pop('alarms', {})
new.update(data)
new = await self.validate_attrs(new)
# If port value is updated, we don't restart nginx, that has to be done manually
await self._update_service(old, new)
return await self.config()
