Initial prefetch size can be arbitrary an restricted only by VM/RPC
limits. For TraverseIterator() there is an explicit check on the
server-side, though.
Introduced in df055fead5.
Refs #931.
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
Proxy contract can now be used as an owner of NNS domains, thus we need
it not only to pay for the transaction but also to check domain
ownership. CalledByEntry is not enough, because we may register NNS
domains owned by proxy indirectly from the container contract.
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
ContainersOf() is better in almost every aspect, besides creating a
session when the containers number is between 1024 and 2048 (prefetch
script does limited unwrapping). Making List() private helps to ensure
it is no longer used and can be safely removed in future.
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
* Those methods that can access already existing containers and thus
can get container properties should read namespace from Zone
property. If Zone is not set, take a namespace for root.
* Otherwise, define namespaces by owner ID via frostfs-id contract.
* Improve unit-tests, consider more cases.
Signed-off-by: Airat Arifullin <a.arifullin@yadro.com>
Besides VM stack item limit we also have restrictions on the size of
JSON for a stackitem, which is 128k. This limit is much harder to
calculate, because JSON representation includes type and the encoding is
different for different items. Thus is makes no sense to invent our own
default, so use the one provided by neo-go. But for container listing we
know exactly what we process, so use big enough value, which is tested.
Introduced in be8607a1f6.
Refs #902
Refs https://github.com/nspcc-dev/neo-go/blob/v0.105.0/pkg/vm/stackitem/json.go#L353
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
* Introduce switch rpc guarded rpc actor in morph client
to prevent using invalidated rpc actor when RPC switch
happens.
* Initialize NewContractStorage with SwitchRPCGuardedActor.
Signed-off-by: Airat Arifullin <a.arifullin@yadro.com>
When the number of items to iterate over is less than 2048, there is no
need to create a session and consume resources.
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
* Introduce common method EverExisted
* Define DeletionInfo for struct that must implement Source
* Refactor tree srv
Signed-off-by: Airat Arifullin <a.arifullin@yadro.com>
It's unused and not needed, default fallback lifetime is set by Notary
actor.
Signed-off-by: Anna Shaleva <anna@nspcc.ru>
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
It's not really needed, closing the connection works fine when exiting and
normally the app doesn't need to unsubscribe at all.
Signed-off-by: Roman Khimov <roman@nspcc.ru>
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
After 75d7891ca1
`neo-go` does claim that an empty invocation script is the only way to
fill missing signature for unsigned notary requests. The new notary actor
does it that way and, therefore, breaks notary request parsing by the
Alphabet because of skipping any request that is not filled with a dummy (64
zeros) invocation script. Support both way. The "Dummy" approach will be
dropped later.
Signed-off-by: Pavel Karpy <p.karpy@yadro.com>
It can't be uint64 in fact, but this error is buried deeply in the NetworkInfo
API structure, so we're not touching MagicNumber() for now.
Signed-off-by: Roman Khimov <roman@nspcc.ru>
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
That's the reason #2230 and #2263 were not detected earlier, we actually had
Global scope being used before reconnection to RPC node.
Signed-off-by: Roman Khimov <roman@nspcc.ru>
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
Fixes#2230, fixes#2263. CustomGroups are nice while we're only calling NeoFS
contracts, but it doesn't work at all for standard ones like GAS or Notary.
Signed-off-by: Roman Khimov <roman@nspcc.ru>
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
With non-blocking pool restricted by 10 in capacity, the probability of
dropping events is unexpectedly big. Notifications are an essential part of the FrostFS,
we should not drop anything, especially new epochs.
```
Mar 31 07:07:03 vedi neofs-ir[19164]: 2023-03-31T07:07:03.901Z debug subscriber/subscriber.go:154 new notification event from sidechain {"name": "NewEpoch"}
Mar 31 07:07:03 vedi neofs-ir[19164]: 2023-03-31T07:07:03.901Z warn event/listener.go:248 listener worker pool drained {"chain": "morph", "capacity": 10}
```
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
It led to a neo-go dead-lock in the `subscriber` component. Subscribing to
notifications is the same RPC as any others, so it could also be blocked
forever if no async listening (reading the notification channel) routine
exists. If a number of subscriptions is big enough (or a caller is lucky
enough) subscribing loop might have not finished subscribing before the
first notification is received and then: subscribing RPC is blocked by
received notification (non)handling and listening notifications routine is
blocked by not finished subscription loop.
That commit starts listening notification channel _before_ any subscription
actions.
Signed-off-by: Pavel Karpy <p.karpy@yadro.com>
We have already had and solved plenty of deposit issues and notary balance
is a really important thing. Deserves to be INFO even before the huge logs
severity refactor, happens on an app start only.
Signed-off-by: Pavel Karpy <p.karpy@yadro.com>
It does not use deprecated methods anymore but also adds more code that
removes. Future refactor that will affect more components will optimize
usage of the updated API.
Signed-off-by: Pavel Karpy <p.karpy@yadro.com>
