Commit graph

2046 commits

Author SHA1 Message Date
Evgenii Stratonikov
4baf00aa21 [#1884] services/object: Fallback to GET in GET_RANGE
Current spec allows denying GET_RANGE requests from other storage nodes.
However, GET should always be allowed and it is enough to perform
GET_RANGE locally

Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-12 17:05:51 +03:00
Leonard Lyubich
dde4d4df2a [#1878] services/object: Fix child check in GET
In previous implementation `ObjectService.Get` RPC handler failed with
`parent address in child object differs` while assembling the "big"
object. This was caused by the child check which required parent
reference to be set in all child objects. The check was impracticable
because not all elements of the split-chain have a link to the parent.

Make `execCtx.isChild` to return `true` if parameterized object has no
parent header in its own header.

Signed-off-by: Leonard Lyubich <ctulhurider@gmail.com>
2022-10-12 16:56:37 +03:00
Evgenii Stratonikov
49eab6318c [#1867] control: Fix degraded-read-only mode parsing
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-12 11:20:48 +03:00
Evgenii Stratonikov
c0199dee93 [#1867] services/control: Interpret empty list of IDs as all shards
In neofs-cli the flag is still required, but `all` can be used to
process all shards.

Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-12 11:20:48 +03:00
Evgenii Stratonikov
19c0a74e94 [#1867] services/control: Allow to provide multiple shard IDs to some commands
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-12 11:20:48 +03:00
74d2f2c8d3 [#1854] cli: Unify help messages
Signed-off-by: Anton Nikiforov <an.nikiforov@yadro.com>
2022-10-12 09:53:11 +03:00
Leonard Lyubich
050ad2762c [#1680] replicator: Consider NODE_UNDER_MAINTENANCE as OK
Node response with `NODE_UNDER_MAINTENANCE` status signals that the node
was switched to maintenance mode. There is a delay between the actual
switch and the reflection in the network map of up to one epoch. To
speed up the reaction to the maintenance, it is required to recognize
such node responses in the Policer.

Make `Policer.processNodes` to exclude elements with shortage decreasing
on `NODE_UNDER_MAINTENANCE` status response.

Signed-off-by: Leonard Lyubich <ctulhurider@gmail.com>
2022-10-11 12:54:27 +03:00
Leonard Lyubich
e99e25b52f [#1680] replicator: Consider nodes under maintenance as OK
Nodes under maintenance SHOULD not respond to object requests. Based on
this, storage node's Policer SHOULD consider such nodes as problem ones.
However, to prevent spam with the new replicas, on the contrary, Policer
should consider them normal.

Make `Policer.processNodes` to exclude elements if `IsMaintenance()`
with shortage decreasing.

Signed-off-by: Leonard Lyubich <ctulhurider@gmail.com>
2022-10-11 12:54:27 +03:00
Leonard Lyubich
df5d7bf729 [#1680] replicator: Work with netmap.NodeInfo in TaskResult
Make `replicator.TaskResult` to accept `netmap.NodeInfo` type instead of
uint64 in order to clarify the meaning and prevent passing the random
numbers.

Signed-off-by: Leonard Lyubich <ctulhurider@gmail.com>
2022-10-11 12:54:27 +03:00
Leonard Lyubich
e6f8904040 [#1680] policer: Refactor tracking the processed nodes
Add clear methods with docs. Use the methods instead of direct map
and bool instructions.

Signed-off-by: Leonard Lyubich <ctulhurider@gmail.com>
2022-10-11 12:54:27 +03:00
feaa9eace7 [#1768] Add healthcheck method to the Tree service
Signed-off-by: Anton Nikiforov <an.nikiforov@yadro.com>
2022-10-11 12:53:45 +03:00
Leonard Lyubich
807c0a1321 [#1859] services/object: Do not session check relation in PUT
It doesn't make sense to check object relation in session check of
`ObjectService.Put` RPC which has been spawned by `ObjectService.Delete`
with session. Session issuer can't predict identifier of the tombstone
object to be created.

Signed-off-by: Leonard Lyubich <ctulhurider@gmail.com>
2022-10-10 20:09:47 +03:00
Evgenii Stratonikov
2d43892fc9 [#1840] neofs-node: Use blobstor paths to identify shard
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-10 11:14:55 +03:00
Evgenii Stratonikov
4b005d3178 [#1840] blobstor: Return info about all components
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-10 11:14:55 +03:00
Evgenii Stratonikov
6d7ffefec5 [#1840] blobstor/test: Add tests for auxiliary functions
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-10 11:14:55 +03:00
Evgenii Stratonikov
328691c94f [#1840] blobstor: Return Path from components
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-10 11:14:55 +03:00
Evgenii Stratonikov
9b241e4a17 [#1840] neofs-node: Allow to use mode: disabled in config
Currently, when removing shard special care must be taken with respect
to shard numbering. `mode: disabled` allows to leave shard configuration
in place while also ignoring it during initialization. This makes
disk replacement much more convenient.

Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-10 11:14:55 +03:00
Evgenii Stratonikov
b2aa9947c2 [#1829] engine: Delete split objects properly
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-07 16:35:46 +03:00
Evgenii Stratonikov
6557f5d249 [#1839] engine: Handle Inhume errors properly
If shard is in read-only or degraded mode, there is no need to increase
error counter.

Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-07 14:25:52 +03:00
Evgenii Stratonikov
7356ee91ff [#1837] services/object: Optimize uniqueIDWriter
Avoid encoding object ID to string.

Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-07 13:29:15 +03:00
Evgenii Stratonikov
2ac42b70ce [#1826] services/object: Parallelize object search
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-07 13:29:15 +03:00
Evgenii Stratonikov
0cb8e7f6f1 [#1837] services/object: Fix log messages
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-07 13:29:15 +03:00
Leonard Lyubich
e54b52ec03 [#1420] object/acl: Fix correlation of object session to request
In previous implementation of `neofs-node` app object session was not
checked for substitution of the object related to it. Also, for access
checks, the session object was substituted instead of the one from the
request. This, on the one hand, made it possible to inherit the session
from the parent object for authorization for certain actions. On the
other hand, it covered the mentioned object substitution, which is a
critical vulnerability.

Next changes are applied to processing of all Object service requests:
 - check if object session relates to the requested object
 - use requested object in access checks.

Disclosed problem of object context inheritance will be solved within

Signed-off-by: Leonard Lyubich <ctulhurider@gmail.com>
2022-10-07 10:34:38 +03:00
Evgenii Stratonikov
5834f9807e [#1847] services/policer: Provide container ID in logs
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-07 09:58:16 +03:00
Evgenii Stratonikov
c20c2e3b39 [#1845] morph/client: Use 0 as OfflineState in update peer
As the documentation prescribes.

Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-06 12:26:25 +03:00
Leonard Lyubich
713aea06fa [#1681] node: Block only Object service ops under maintenance
In previous implementation node blocked any operation of local object
storage in maintenance mode. There is a need to perform some storage
operations like data evacuation or restoration.

Do not call block storage engine in maintenance mode. Make all Object
service operations to return `apistatus.NodeUnderMaintenance` error from
each local op.

Signed-off-by: Leonard Lyubich <ctulhurider@gmail.com>
2022-10-05 11:41:49 +03:00
Leonard Lyubich
eb1fba5182 [#1680] morph/netmap: Adopt to recent contract changes
After recent Netmap contract changes all read methods which return
network map (either candidates or snapshots) encode node descriptors
into same structure.

Decode `netmap.Node` contract-side structure from the call results.
Replace node state with the value from the `netmap.Node.State` field.

Signed-off-by: Leonard Lyubich <ctulhurider@gmail.com>
2022-10-05 11:41:49 +03:00
Leonard Lyubich
6a2ec21641 [#1681] ir/netmap: Check MAINTENANCE state in UpdateState
Storage node can be requested to be switched into `MAINTENANCE` state.
Inner Ring should accept such requests only if network configuration
allows it.

Make `Processor` of Netmap contract's notifications to depend on
`state.NetworkSettings`. Make `Processor.processUpdatePeer` to call
`MaintenanceModeAllowed` if notification event relates to `MAINTENANCE`
mode`. Share singe `state.NetworkSettings` provider in Inner Ring
application.

Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
2022-10-05 11:41:49 +03:00
Leonard Lyubich
bdb8243a5a [#1681] morph/netmap: Support MAINTENANCE state notification
After recent changes Netmap contract can send `UpdateState` notification
event with `MAINTENANCE` node's state. There is a need to provide
functionality to work with the status.

Provide `UpdatePeer.Maintenance` method. Support new state in
`ParseUpdatePeer` and `ParseUpdatePeerNotary` functions.

Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
2022-10-05 11:41:49 +03:00
Leonard Lyubich
0903b2af93 [#1680] morph/netmap: Allow to set MAINTENANCE state
After recent changes in NeoFS protocol storage nodes can be in
`MAINTENANCE` state. There is a need to support this state in
`UpdateState` operation.

Add `UpdatePeerPrm.SetMaintenance` method which makes node to be
switched into `MAINTENANCE` mode after the `UpdatePeerState` operation.

New functionality is going to be used in Storage node application for
Control API serving.

Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
2022-10-05 11:41:49 +03:00
Leonard Lyubich
6b99f2a0df [#1680] ir/netmap: Correlate node's state with network settings
Inner Ring should allow registering of storage nodes with `MAINTENANCE`
state in the NeoFS network only if its configuration allows this status.

Make `networkSettings.MaintenanceModeAllowed` to call
`MaintenanceModeAllowed` method of underlying Netmap contract's client
in order to assert state allowance.

From now nodes will be accepted to the network with `MAINTENANCE` state
only with the appropriate network configuration.

Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
2022-10-05 11:41:49 +03:00
Leonard Lyubich
0826310b2a [#1681] morph/netmap: Support MaintenanceModeAllowed config
After recent changes in the NeoFS API protocol network configuration
contains `MaintenanceModeAllowed` boolean flag. There is a need to
support the config value in all NeoFS applications.

Provide `Client.MaintenanceModeAllowed` method which read the config
from the Sidechain. Extend `NetworkConfiguration` structure with
`MaintenanceModeAllowed` field.

Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
2022-10-05 11:41:49 +03:00
Leonard Lyubich
d2d4191868 [#1680] morph/netmap: Support HomomorphicHashingDisabled config
`NetworkConfiguration` represents NeoFS network configuration stored in
the Sidechain. In previous implementation the configuration missed flag
of disabled homomorphic hashing.

Add `NetworkConfiguration.HomomorphicHashingDisabled` boolean field.
Decode the field in `Client.ReadNetworkConfiguration` method. Print this
value in `netmap netinfo` command of NeoFS CLI.

Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
2022-10-05 11:41:49 +03:00
Leonard Lyubich
e0194dbde5 [#1681] morph/netmap: Refactor reading the boolean configurations
`readBoolConfig` method is going to be reused for reading other
configuration values. All boolean settings are `false` by default, so it
makes sense to return default value on missing key directly from
`readBoolConfig`.

Handle `ErrConfigNotFound` case in `readBoolConfig` method. Change
`HomomorphicHashDisabled` method to call `readBoolConfig` directly.

Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
2022-10-05 11:41:49 +03:00
Leonard Lyubich
f64ae55806 [#1681] ir/netmap: Require MAINTENANCE mode to be allowed by network
There is a need to prevent limitless abuse of MAINTENANCE status of the
storage nodes. To do this, configuration of the NeoFS network is going
to be extended with the flag which allows the state. Until this is done,
it makes sense to prepare a site for this in the code.

Define `state.NetworkSettings` interface as an abstraction of global
network configuration within the `state` package. Make
`NetMapCandidateValidator` to depend on `NetworkSettings` and provide
corresponding field setter. Change `VerifyAndUpdate` method's behavior
to return an error for candidates with MAINTENANCE state if this state
is disallowed by the network configuration. Provide `NetworkSettings`
from the wrapper over Netmap contract's client on Inner Ring application
side. The provider is implemented to statically disallow MAINTENANCE
mode in order to save previous behavior.

Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Signed-off-by: Leonard Lyubich <ctulhurider@gmail.com>
2022-10-05 11:41:49 +03:00
Leonard Lyubich
42fb40e841 [#1680] morph/netmap: Pre-refactor processing of node states
New network status of storage nodes is going to be introduced. To
simplify the addition, it would be useful to prepare the code for this.

Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
2022-10-05 11:41:49 +03:00
Leonard Lyubich
8858840751 [#1796] ir/netmap: Allow to call AddPeer with ONLINE state only
In previous implementation Inner Ring allowed storage nodes with any
state to register in the network. According to the current design, only
nodes with ONLINE state are allowed to enter the network map.

Create new `state` sub-package of `nodevalidation` package of Inner Ring
application. Define `state.NetMapCandidateValidator` type and provide
`NodeValidator` interface required by the Inner Ring's processor of
`Netmap` contract's notification events. Embed new validator into the
one used by the Inner Ring application.

From now all `AddPeer` notifications with node state other than `ONLINE`
will be denied.

Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
2022-10-05 11:41:49 +03:00
Pavel Karpy
4eb0ed11f8 [#1809] node: Do not boot up if metabase is outdated
Signed-off-by: Pavel Karpy <carpawell@nspcc.ru>
2022-10-04 12:32:10 +03:00
Evgenii Stratonikov
8b3b16fe62 [#1825] writecache: Flush cache when moving to the DEGRADED mode
Degraded mode allows us to operate without an SSD,
thus writecache should be unavailable in this mode.

Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-04 12:13:09 +03:00
Evgenii Stratonikov
236414df49 [#1817] network: Allow to use network addresses from the iterator
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-04 12:12:18 +03:00
Evgenii Stratonikov
2e3ef817f4 [#1819] engine: Increase error counter for PUT errors
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-04 10:11:52 +03:00
Evgenii Stratonikov
b89e71fa78 [#1819] common: Add ErrNoSpace
Add a common error for this case because it is not an error
which should increase error counter. Single error simplifies checks on
the call-site.

Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-04 10:11:52 +03:00
Evgenii Stratonikov
af56574849 [#1819] engine: Fix error counter in Inhume
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-10-04 10:11:52 +03:00
Pavel Karpy
8ebe95747e [#1770] node: Do not lock on shard's Close call
Signed-off-by: Pavel Karpy <carpawell@nspcc.ru>
2022-10-04 10:08:55 +03:00
Pavel Karpy
887afeaddb [#1770] engine: Do not lock on shard init
Init can take a lot of time. Because the mutex is taken, all new operations
are blocked.

Signed-off-by: Pavel Karpy <carpawell@nspcc.ru>
2022-10-04 10:08:55 +03:00
Pavel Karpy
fbd5bc8c38 [#1770] engine: Support configuration reload
Currently, it only supports changing the compound of the shards.

Signed-off-by: Pavel Karpy <carpawell@nspcc.ru>
2022-10-04 10:08:55 +03:00
Pavel Karpy
5c69e19016 [#1770] fstree: Depth parameter int -> uint64
Negative values have no sense. On the other hand it differs from the
blobovnicza's configuration and prevents unification.

Signed-off-by: Pavel Karpy <carpawell@nspcc.ru>
2022-10-04 10:08:55 +03:00
Pavel Karpy
2d7166f8d0 [#1770] shard: Move NewEpoch event routing on SE level
It will allow dynamic shard management. Closing a shard does not allow
removing event handlers.

Signed-off-by: Pavel Karpy <carpawell@nspcc.ru>
2022-10-04 10:08:55 +03:00
Evgenii Stratonikov
1edc048870 [#1697] services/object: Return proper error if session token is missing
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-09-30 16:13:01 +03:00
Evgenii Stratonikov
0a411908ee [#1806] writecache: Allow to ignore read errors during flush
Signed-off-by: Evgenii Stratonikov <evgeniy@morphbits.ru>
2022-09-28 09:28:01 +03:00