The client needs of the Object service are limited and change not often.
Interface changes of the client library should not affect the operation of
various service packages, if they do not change their requirements for
the provided functionality. To localize the use of the base client and
facilitate further support, an auxiliary package is implemented that will
only be used by the Object service.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
There is no point to pass key storage in parameters because
it can be defined on the service level of application.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
`CommonPrm` structure has private key for remote operations.
It obtained in the beginning of request processing. However,
not every operation triggers remote calls. Therefore, key
might not be used. It is important to avoid early key fetching
because `TokenStore` now returns error if session token does not
exist. This is valid case when container nodes receive request with
session token (for ACL pass) and they should process request locally.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
This is invalid operation for storage nodes that receive part of split
object. While object is signed by session token, the message itself
should be signed by the node key.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
After storage engine started to limit number of PUT operations there is no
need to limited worker pool in Object Put service.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Session token can be present in both object header and
request meta header. They are the same during initial object
placement.
At the object replication, storage node puts object without
any session tokens attached to the request. If container's eACL
denies object.Put for USER role (use bearer to upload), then
replication might fail on objects with session tokens of the
signed by container owner. It is incorrect, so use session
token directly from request meta header.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
There is a need to check if public key in the RPC response matches the
public key of the related storage node declared in network map.
Define `ErrWrongPublicKey` error. Implement RPC response handler's
constructor `AssertKeyResponseCallback` which checks public key. Construct
handler and pass it to client's option `WithResponseInfoHandler`.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
There is a need to have the ability to expand the data needed for client
construction.
Replace `network.AddressGroup` parameter of client cache interfaces with
`client.NodeInfo`.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
In previous implementation Object PUT used single pool of workers for local
and remote ops, but these ops are heterogeneous.
Use remote/local pool for remote/local operations in PUT service. At first
the pools are configured with the same size.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
In previous implementation `distributedTarget` didn't check if next node is
local. This check was performed by the handlers (target initializer and
relay func).
Make `distributedTarget` to calculate node's locality. Pass locality flag to
the handlers.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Forwarding mechanism resends original request. During split object chain traversal,
storage node performs multiple `object.Head` requests on each child. If request
forwarding happens, then `object.Head` returns object ID of the original request.
This produces infinite assembly loop.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
Make `errIncompletePut` to be a structure which wraps single client error.
Wrap error of the last client into `errIncompletePut` during placement
execution.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
In previous implementation Object service's handler returned const error in
case of failure (full or partial) of PUT operation. This did not even allow
us to roughly guess what the reason is. Not as a complete solution, but to
alleviate some cases where all nodes in a container return the same error,
it is suggested to return the error of the last server that responded.
Return latest server error from placement loop of `iteratePlacement` method
of `distributedTarget` type.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
In previous implementation sticky bit could disrupt access of container
nodes to replication. According to NeoFS specification sticky bit should not
affect the requests sent by nodes from SYSTEM group.
Add role check to `stickyBitCheck`.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Some software components regulate the way of working with placement arrays
when a local node enters it. In the previous implementation, the locality
criterion was the correspondence between the announced network address
(group) and the address with which the node was configured. However, by
design, network addresses are not unique identifiers of storage nodes in the
system.
Change comparisons by network addresses to comparisons by keys in all
packages with the logic described above. Implement `netmap.AnnouncedKeys`
interface on `cfg` type in the storage node application.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
In previous implementation `placement.Traverser.Next` method returned slice
of `network.AddressGroup` elements. There is a need to process keys of
storage nodes besides network addresses for intra-container communication.
Wrap `network.AddressGroup` in a new type `placement.Node` that summarizes
the storage node information required for communication. Return slice of
`Node` instances from `Traverser.Next` method. Fix compilation breaks in
dependent packages.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Do not init caches for eACL, containers and netmap
if `disable_cache` config options is `true`, use
direct RPC calls instead.
Signed-off-by: Pavel Karpy <carpawell@nspcc.ru>
Make placement `Traverser.Next` method to return ``[]network.AddressGroup`
in order to support multiple addresses of the storeage nodes.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Make Object Head service to work with `AddressGroup` instead of `Address`
in order to support multiple addresses of the storage node.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Make Object Search service to work with `AddressGroup` instead of `Address`
in order to support multiple addresses of the storage node.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Make Object Get service to work with `AddressGroup` instead of `Address` in
order to support multiple addresses of the storage node.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Make Object Put service to work with `AddressGroup` instead of `Address` in
order to support multiple addresses of the storage node.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Make `LocalAddressSource.LocalAddress` method to return `AddressGroup`. Make
`IsLocalAddress` function to accept parameter of type `AddressGroup`. Adopt
the application code with temporary `GroupFromAddress` helper.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
There is a need to generalize single-address client to group-address client.
To do this, we can re-implement `Client` interface from NeoFS API Go library
and still use it in the application code. There is a problem with method
`Raw` which must return single-address raw client. So as not to make changes
to API library we need to overload Client interface in order to support
`Raw` method in group-address client implementation.
Define `Client` interface in new `pkg/core/client` package. Completely
inherit API `Client` interface. Add `RawForAddress` method to build raw
client for the single node address. Adopt the application code that used Raw
method to work with new `Client`.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
`network.Address` structure in most cases created once and used read-only.
Replace `AddressFromString` function with `Address.FromString` method with
the same purpose and implementation. Make all libraries to work with value.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Return tyype is clear from the method's signature and docs, there is no
point in reflecting it in the name.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Panic if internal `manet.DialArgs` call returns error since this is
unexpected according to `AddressFromString` implementation.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
NeoFS network dictates the limitation of the maximum size of the "finished"
objects. The payload size of an object is recorded in its header. When
writing finished objects, the nodes must check:
* satisfying the constraint;
* matching the value in the header to the number of payload bytes.
Provide value returned by `MaxSizeSource` component to `validatingTarget`.
Check max size value during the stream of the "finished" objects. Check
header value during the streaming and on-close. Check payload size in v2
relay scenario.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
As it explained in previous commit, session token also should
be presented in original meta header but can be omitted in higher
layers.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
Request meta headers are organized in a layers, where
upper layers re-sign down layers. Bearer token should be
a part of original meta header and it can be omitted in
upper layers. Therefore we need to traverse over linked list
of meta header to the original meta header to get bearer token.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
Request forwarding callback should be called only if set since it is an
optional parameter.
Call `forwarder` function only if it is non-nil.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Request forwarding callback should be called only if set since it is an
optional parameter. In GetRangeHash forwarder is never set.
Call `forwarder` function only if it is non-nil. Remove no longer needed
`hashOnly` option.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Wrap functions at `pkg/errors` return nil if error argument
was nil. fmt.Errorf always returns error so we need to add
missing error checks to the code.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
In current implementation `Object.GetRangeHash` RPC handler forwards range
requests for payload data. Missing request forwarder of the original request
caused NPE during execution.
Do not call request forwarder if payload range hash is requested.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
To enable TLS support we can't operate with IP addresses directly.
Certificates are issued with host names so it is required to
pass them into RPC client. DNS resolving should be done by transport
layer and not be a part of node. Therefore `IPAddrString` usage is
removed from code.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
Allocate capacity instead of length of the slice to write the object payload
range since each chunk is written through `append`.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
In previous implementation node's Object Get/Head/GetRange V2 service
handlers created a new request for each RPC. Now original requests are
re-signed according to API specification. Logical handler abstracts from
this version-dependent logic through `RequestForwarder` callback.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
In previous implementation node's Object Search V2 service handler created a
new request for each RPC. Now original requests are re-signed according to
API specification. Logical handler abstracts from this version-dependent
logic through `RequestForwarder` callback.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Replace usage of `cache.ClientCache` type with interface with similar
signature. This will further allow overloading clients without affecting the
logic of dependent packages.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
At object.Put operation node does not transfer any
data. It is done in the stream inside, so we should
register duration from `object.Put` invocation until
`CloseAndRecv` on the stream.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
We can't rely on object size in the header because it
might be omitted on initial put or it can be 0xFF.. on
streaming data.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
Because options are not used when client is already in cache
providing them to shared cache is misleading at best.
In the worst case `dial_timeout` is set randomly (because of race
condition) which can lead to one service having `dial_timeout` of
another. Thus we set default client creation options when cache is
created.
Signed-off-by: Evgenii Stratonikov <evgeniy@nspcc.ru>
Replace `eacl.Storage` interface implementation from eACL lib to neofs-node
app package. Remove `eacl.WithMorphClient` option in order to abstract eACL
validator from eACL storage implementation.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Make object delete service to use network information to calculate and set
expiration of the created tombstone.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
The lifetime of an object can be limited by specifying a correspondin
well-known attribute. Node should refuse to save expired objects.
Checking objects in FormatValidator is extended with an expiration attribute
parsing step.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Add SignWithRFC6979 option to signature verification function since eACL
table are signed by users with this option.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Since the contract started returning the table signature, it became
necessary to check its correctness.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Classifier looks at list of inner ring nodes and container
nodes from current and previous epoch to classify request.
Sometimes these checks might return error.
Consider there is a request from unknown key and container's
placement policy valid for current epoch and invalid for past
epoch. Classifier tries to find if key belongs to container
node from current epoch -- it is not. Then it tries to find if
key belongs to container node from past epoch and it throws
error, because placement policy is invalid for past epoch.
This is a legit case and classifier should ignore such errors
to provide best effort in matching. The only error classifier
should return is an error when request does not contain
public key to classify it.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
Support processing of NetmapEpoch and NetmapLookupDepth X-headers when
processing object read operations. Placement for operations
Get/Head/GetRange/GetRangeHash/Search is built for the epoch specified in
NetmapEpoch X-header (by default latest). Also the specified operations are
processed until success is achieved for network maps from the past up to
NetmapLookupDepth value. Behavior for default values (zero or missing) left
unchanged.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Add binary salt field to RangeHashPrm struct. Implement field setter. Set
salt from the request in v2 service.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
In previous implementation of eACL service v2 the response X-headers were
validated at the stage of re-checking eACL. This provoked a mismatch of
records in the eACL table with requests. Fix this behavior by checking the
headers from the request, not the response.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Forward request X-headers to client calls during internal processing of
Object operations on the node.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
In previous implementation eACL validator didn't take into account container
and object ID fields of request bodies.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
With the update of the local storage engine, the headers of virtual objects
are directly given. In this regard, the step with obtaining the the right
child header is removed.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
With updated specification of object related operation
we don't have this search attribute any more and we
should not use functions related to this attribute.
This commit breaks object service logic, however it will
be fixed later.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
ErrEACLNotFound error was defined in implementation package. EACL validator
checked this error after the call of eACL storage interface method. Replace
ErrEACLNotFound to core container library. in order to: on the one hand not
use an implementation error, on the other hand, to be able to reuse a
generic type error (404).
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Now morph library returns error if there is not eACL in
sidechain storage. However in this case eACL check should
be passed since it is the same as having empty eACL table.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
Sticky bit checks if object owner and request owner are the
same. Container owner should not used in this check.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
`owner` field may be misused as request sender owner, however
it is a owner of a container for that request. New naming
should be clear.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
Define ErrRelationNotFound error in searchsvc package. Return
ErrRelationNotFound from RelationSearcher.SearchRelation method if search
result is empty.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Refactored pkg/netmap package provides JSON converters for
NodeInfo and PlacementPolicy structures, that has been used
by client applications.
It also updates Node structure itself so it is a part of
grpc <-> v2 <-> pkg conversion chain.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
Replace meta Bucket with meta.DB instance in local storage implementation.
Adopt all dependent components to new local storage.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
There is a bit to allow or deny bearer token check for
each object service method. If this bit is not set then
ignore bearer token and use extended ACL table from
sidechain.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
If bearer token is presented in the request then check
if it is a valid one and then use it to process extended
ACL checks.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
Malicious user can stole public session key and use
it by sending request from it's own scope. To prevent
this each session token is signed and signature private
key must be corresponded with owner id in token. Therefore
malicious node cannot impersonate request without private
key to sign token.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
Define ErrNotFound error in headsvc package. Return ErrNotFound from Head
method if the header was not found in the container.
Signed-off-by: Leonard Lyubich <leonard@nspcc.ru>
Wrong boolean operation order made matcher return false
on `non-root` search query with non-regular objects. Instead
it should return true for `non-root` query and false for `root`
query.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
If object service produces new request, the should contain
session token. This is the only way for node to grant access
for a private container.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
Classifier fetches public key of the request owner
and owner itself. Extended ACL check should rely on
this public key, because it might be extracted from
session token.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
With one tombstone for split objects we can't simply
place it in container. We should inform all nodes that
store split objects of removed original object.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
Owner of the request is stored in session token most of the times.
Put request contains session token in the object body, so we have
to fetch it from there.
Signed-off-by: Alex Vanin <alexey@nspcc.ru>
