Previously RPC server shutdown procedure listened to the execution
channel and stopped at the first element that arrived in the queue. This
could lead to the following problems:
* stopper could steal the execution result from subscriber
* stopper didn't wait for other subscription actions to complete
Add dedicated channel to `Server` for subscription routine. Close the
channel on `handleSubEvents` return and wait for signal in `Shutdown`.
Signed-off-by: Leonard Lyubich <leonard@morphbits.io>
Previously RPC server could never be shut down completely due to
some start precondition failure (in particular, inability to serve HTTP
on any configured endpoint). The problem was caused by next facts:
* start method ran subscription routine after HTTP init succeeded only
* stop method blocked waiting for the subscription routine to return
Run `handleSubEvents` routine on fresh `Start` unconditionally. With
this change, `Shutdown` method won't produce deadlock since
`handleSubEvents` closes wait channel.
Refs #2896.
Signed-off-by: Leonard Lyubich <leonard@morphbits.io>
There is an existing problem with RPC server shutdown freeze after start
failure due to some init actions (at least HTTP listen) described in
#2896.
Add dedicated unit test which checks that `Shutdown` returns within 5s
after `Start` method encounters internal problems.
Signed-off-by: Leonard Lyubich <leonard@morphbits.io>
Move them to the core/network packages, close#2950. The name of
mempool's unsorted transactions metrics has been changed along the
way to match the core's metrics naming convention.
Signed-off-by: Anna Shaleva <shaleva.ann@nspcc.ru>
I've carefully checked the way how new service can be added to the
Blockchain instance or to be removed from it. Current implemention
of SetNotary and SetOracle methods doesn't contain dangerous code,
and native contracts have atomic values everywhere where service
is stored.
Current implementation of Notary, Oracle and StateRoot services'
reload/disabling/enabling on SIGUSR1 is safe and doesn't require
any adjustment.
This commit closes#2944, it's not a bug in the code, it's just
stale documentation.
Signed-off-by: Anna Shaleva <shaleva.ann@nspcc.ru>
`cache` persisting operation is concurrent with the storage modifications
made by further state reset stages. We can't allow situation when data
from the next stage are leaking into the previous stage. State reset stages
must be atomic in turms of DB persisting, thus, use another `upperCache`
MemCached store to keep them apart.
Here are the results of mainnet's BoltDB reset from 1.1M to 6K:
This patch:
```
anna@kiwi:~/Documents/GitProjects/nspcc-dev/neo-go$ ./bin/neo-go db reset -m --debug --height 600000
2023-04-11T16:15:25.783+0300 INFO MaxBlockSize is not set or wrong, setting default value {"MaxBlockSize": 262144}
2023-04-11T16:15:25.783+0300 INFO MaxBlockSystemFee is not set or wrong, setting default value {"MaxBlockSystemFee": 900000000000}
2023-04-11T16:15:25.783+0300 INFO MaxTransactionsPerBlock is not set or wrong, using default value {"MaxTransactionsPerBlock": 512}
2023-04-11T16:15:25.783+0300 INFO MaxValidUntilBlockIncrement is not set or wrong, using default value {"MaxValidUntilBlockIncrement": 5760}
2023-04-11T16:15:25.787+0300 INFO restoring blockchain {"version": "0.2.8"}
2023-04-11T16:15:25.906+0300 INFO initializing state reset {"target height": 600000}
2023-04-11T16:15:25.906+0300 DEBUG trying to reset blocks, transactions and AERs
2023-04-11T16:15:57.031+0300 INFO intermediate batch of removed blocks, transactions and AERs is collected {"batch": 1, "took": "31.125057214s"}
2023-04-11T16:16:12.644+0300 DEBUG intermediate batch of removed blocks, transactions and AERs is persisted {"batch": 1, "took": "15.613156971s", "keys": 321895}
2023-04-11T16:16:13.895+0300 INFO intermediate batch of removed blocks, transactions and AERs is collected {"batch": 2, "took": "16.663444208s"}
2023-04-11T16:16:19.784+0300 INFO last batch of removed blocks, transactions and AERs is collected {"batch": 3, "took": "5.760308543s"}
2023-04-11T16:16:19.784+0300 INFO blocks, transactions ans AERs are reset {"took": "53.878632911s"}
2023-04-11T16:16:22.870+0300 DEBUG intermediate batch of removed blocks, transactions and AERs is persisted {"batch": 2, "took": "8.974838893s", "keys": 334823}
2023-04-11T16:16:22.870+0300 DEBUG trying to reset contract storage items
2023-04-11T16:16:27.272+0300 DEBUG last batch of removed blocks, transactions and AERs is persisted {"batch": 3, "took": "7.487357441s", "keys": 208913}
2023-04-11T16:17:23.678+0300 INFO intermediate batch of contract storage items and IDs is collected {"batch": 1, "took": "1m0.80711106s"}
2023-04-11T16:18:00.769+0300 INFO intermediate batch of contract storage items and IDs is collected {"batch": 2, "took": "36.967660061s"}
2023-04-11T16:18:20.478+0300 DEBUG intermediate batch of contract storage items is persisted {"batch": 1, "took": "56.796257788s", "keys": 200000}
2023-04-11T16:18:54.115+0300 INFO intermediate batch of contract storage items and IDs is collected {"batch": 3, "took": "33.637412437s"}
2023-04-11T16:19:18.844+0300 DEBUG intermediate batch of contract storage items is persisted {"batch": 2, "took": "1m18.0737668s", "keys": 200000}
2023-04-11T16:19:27.650+0300 INFO last batch of contract storage items is collected {"batch": 4, "took": "8.806264019s"}
2023-04-11T16:19:27.650+0300 INFO contract storage items are reset {"took": "3m4.780232077s", "keys": 656944}
2023-04-11T16:20:15.660+0300 DEBUG intermediate batch of contract storage items is persisted {"batch": 3, "took": "1m21.544386403s", "keys": 200000}
2023-04-11T16:20:15.660+0300 DEBUG trying to reset headers information
2023-04-11T16:20:16.385+0300 INFO headers information is reset {"took": "725.174932ms"}
2023-04-11T16:20:19.586+0300 DEBUG last batch of contract storage items and IDs is persisted {"batch": 4, "took": "51.936278608s", "keys": 56945}
2023-04-11T16:20:19.587+0300 DEBUG trying to reset state root information and NEP transfers
2023-04-11T16:20:35.845+0300 INFO state root information and NEP transfers are reset {"took": "16.25852114s"}
2023-04-11T16:21:10.000+0300 DEBUG headers information is persisted {"took": "53.613638429s", "keys": 528438}
2023-04-11T16:21:10.003+0300 DEBUG trying to remove stale storage items
2023-04-11T16:21:18.108+0300 INFO stale storage items are reset {"took": "8.105140658s", "keys": 1350176}
2023-04-11T16:21:18.108+0300 DEBUG trying to remove state reset point
2023-04-11T16:21:18.108+0300 INFO state reset point is removed {"took": "8.554µs"}
2023-04-11T16:21:20.151+0300 DEBUG state root information and NEP transfers are persisted {"took": "44.305707049s", "keys": 602578}
2023-04-11T16:21:20.212+0300 INFO state reset point information is persisted {"took": "2.103764633s", "keys": 2}
2023-04-11T16:21:20.213+0300 INFO reset finished successfully {"took": "5m54.306861367s"}
```
The previous commit:
```
anna@kiwi:~/Documents/GitProjects/nspcc-dev/neo-go$ ./bin/neo-go db reset -m --debug --height 600000
2023-04-11T16:24:04.256+0300 INFO MaxBlockSize is not set or wrong, setting default value {"MaxBlockSize": 262144}
2023-04-11T16:24:04.256+0300 INFO MaxBlockSystemFee is not set or wrong, setting default value {"MaxBlockSystemFee": 900000000000}
2023-04-11T16:24:04.256+0300 INFO MaxTransactionsPerBlock is not set or wrong, using default value {"MaxTransactionsPerBlock": 512}
2023-04-11T16:24:04.256+0300 INFO MaxValidUntilBlockIncrement is not set or wrong, using default value {"MaxValidUntilBlockIncrement": 5760}
2023-04-11T16:24:04.261+0300 INFO restoring blockchain {"version": "0.2.8"}
2023-04-11T16:24:04.368+0300 INFO initializing state reset {"target height": 600000}
2023-04-11T16:24:04.368+0300 DEBUG trying to reset blocks, transactions and AERs
2023-04-11T16:24:30.363+0300 INFO intermediate batch of removed blocks, transactions and AERs is collected {"batch": 1, "took": "25.995261037s"}
2023-04-11T16:24:44.947+0300 DEBUG intermediate batch of removed blocks, transactions and AERs is persisted {"batch": 1, "took": "14.584447338s", "keys": 321897}
2023-04-11T16:24:45.791+0300 INFO intermediate batch of removed blocks, transactions and AERs is collected {"batch": 2, "took": "15.428492824s"}
2023-04-11T16:24:51.252+0300 INFO last batch of removed blocks, transactions and AERs is collected {"batch": 3, "took": "5.460662766s"}
2023-04-11T16:24:51.252+0300 INFO blocks, transactions ans AERs are reset {"took": "46.884558096s"}
2023-04-11T16:24:55.399+0300 DEBUG intermediate batch of removed blocks, transactions and AERs is persisted {"batch": 2, "took": "9.607820004s", "keys": 334821}
2023-04-11T16:24:55.399+0300 DEBUG trying to reset contract storage items
2023-04-11T16:24:59.981+0300 DEBUG last batch of removed blocks, transactions and AERs is persisted {"batch": 3, "took": "8.728713255s", "keys": 208913}
2023-04-11T16:25:50.827+0300 INFO intermediate batch of contract storage items and IDs is collected {"batch": 1, "took": "55.426411416s"}
2023-04-11T16:26:28.734+0300 INFO intermediate batch of contract storage items and IDs is collected {"batch": 2, "took": "37.902647706s"}
2023-04-11T16:26:53.960+0300 DEBUG intermediate batch of contract storage items is persisted {"batch": 1, "took": "1m3.129453265s", "keys": 200001}
2023-04-11T16:27:27.645+0300 INFO intermediate batch of contract storage items and IDs is collected {"batch": 3, "took": "33.685283662s"}
2023-04-11T16:27:52.173+0300 DEBUG intermediate batch of contract storage items is persisted {"batch": 2, "took": "1m23.438465575s", "keys": 199999}
2023-04-11T16:28:00.995+0300 INFO last batch of contract storage items is collected {"batch": 4, "took": "8.821990443s"}
2023-04-11T16:28:00.995+0300 INFO contract storage items are reset {"took": "3m5.595950958s", "keys": 656944}
2023-04-11T16:28:49.164+0300 DEBUG intermediate batch of contract storage items is persisted {"batch": 3, "took": "1m21.518344712s", "keys": 200000}
2023-04-11T16:28:49.164+0300 DEBUG trying to reset headers information
2023-04-11T16:28:49.936+0300 INFO headers information is reset {"took": "772.36435ms"}
2023-04-11T16:28:53.122+0300 DEBUG last batch of contract storage items and IDs is persisted {"batch": 4, "took": "52.126928092s", "keys": 56945}
2023-04-11T16:28:53.122+0300 DEBUG trying to reset state root information and NEP transfers
2023-04-11T16:29:09.332+0300 INFO state root information and NEP transfers are reset {"took": "16.20921699s"}
2023-04-11T16:29:46.264+0300 DEBUG headers information is persisted {"took": "56.326715249s", "keys": 528438}
2023-04-11T16:29:46.267+0300 DEBUG trying to remove stale storage items
2023-04-11T16:29:53.986+0300 INFO stale storage items are reset {"took": "7.718950145s", "keys": 1350176}
2023-04-11T16:29:53.986+0300 DEBUG trying to remove state reset point
2023-04-11T16:29:53.986+0300 INFO state reset point is removed {"took": "6.013µs"}
2023-04-11T16:29:55.899+0300 DEBUG state root information and NEP transfers are persisted {"took": "46.567302762s", "keys": 602578}
2023-04-11T16:29:55.929+0300 INFO state reset point information is persisted {"took": "1.942392208s", "keys": 2}
2023-04-11T16:29:55.929+0300 INFO reset finished successfully {"took": "5m51.561573137s"}
```
Master:
```
anna@kiwi:~/Documents/GitProjects/nspcc-dev/neo-go$ ./bin/neo-go db reset -m --debug --height 600000
2023-04-11T16:34:12.410+0300 INFO MaxBlockSize is not set or wrong, setting default value {"MaxBlockSize": 262144}
2023-04-11T16:34:12.410+0300 INFO MaxBlockSystemFee is not set or wrong, setting default value {"MaxBlockSystemFee": 900000000000}
2023-04-11T16:34:12.410+0300 INFO MaxTransactionsPerBlock is not set or wrong, using default value {"MaxTransactionsPerBlock": 512}
2023-04-11T16:34:12.410+0300 INFO MaxValidUntilBlockIncrement is not set or wrong, using default value {"MaxValidUntilBlockIncrement": 5760}
2023-04-11T16:34:12.413+0300 INFO restoring blockchain {"version": "0.2.8"}
2023-04-11T16:34:12.495+0300 INFO initialize state reset {"target height": 600000}
2023-04-11T16:34:12.513+0300 INFO trying to reset blocks, transactions and AERs
2023-04-11T16:35:03.582+0300 INFO intermediate batch of removed blocks, transactions and AERs is persisted {"batches persisted": 1, "took": "51.087226195s", "keys": 321895}
2023-04-11T16:35:31.302+0300 INFO intermediate batch of removed blocks, transactions and AERs is persisted {"batches persisted": 2, "took": "27.719871393s", "keys": 334823}
2023-04-11T16:35:41.309+0300 INFO last batch of removed blocks, transactions and AERs is persisted {"batches persisted": 3, "took": "10.007017388s", "keys": 208913}
2023-04-11T16:35:41.309+0300 INFO blocks, transactions ans AERs are reset {"took": "1m28.814245057s", "overall persisted keys": 865631}
2023-04-11T16:35:41.309+0300 INFO trying to reset contract storage items
2023-04-11T16:37:38.315+0300 INFO intermediate batch of contract storage items and IDs is persisted {"batch": 1, "took": "1m57.00650253s", "keys": 200000}
2023-04-11T16:39:29.704+0300 INFO intermediate batch of contract storage items and IDs is persisted {"batch": 2, "took": "1m51.385224725s", "keys": 200000}
2023-04-11T16:41:14.991+0300 INFO intermediate batch of contract storage items and IDs is persisted {"batch": 3, "took": "1m45.287483794s", "keys": 200000}
2023-04-11T16:41:31.667+0300 INFO last batch of contract storage items and IDs is persisted {"batch": 4, "took": "16.675347478s", "keys": 56945}
2023-04-11T16:41:31.667+0300 INFO contract storage items and IDs are reset {"took": "5m50.357775401s", "keys": 656944}
2023-04-11T16:41:31.667+0300 INFO trying to reset headers information
2023-04-11T16:42:16.779+0300 INFO headers information is reset {"took": "45.111354262s", "keys": 528438}
2023-04-11T16:42:16.784+0300 INFO trying to reset state root information and NEP transfers
2023-04-11T16:42:35.778+0300 INFO state root information and NEP transfers are reset {"took": "18.99373117s", "keys": 602578}
2023-04-11T16:42:35.781+0300 INFO trying to remove stale storage items
2023-04-11T16:42:43.884+0300 INFO stale storage items are reset {"took": "8.103929306s", "keys": 1350176}
2023-04-11T16:42:43.885+0300 INFO trying to remove state reset point
2023-04-11T16:42:43.926+0300 INFO stale reset point is removed {"took": "41.858883ms", "keys": 2}
2023-04-11T16:42:43.932+0300 INFO reset finished successfully {"took": "8m31.437493325s"}
```
Signed-off-by: Anna Shaleva <shaleva.ann@nspcc.ru>
Everywhere including examples, external interop APIs, bindings generators
code and in other valuable places. A couple of `interface{}` usages are
intentionally left in the CHANGELOG.md, documentation and tests.
Do not add them directly to chain, it will be done by the block queue
manager. Close https://github.com/nspcc-dev/neo-go/issues/2923. However,
this commit is not valid without
https://github.com/roman-khimov/dbft/pull/4.
It's the neo-go's duty to initialize consensus after subsequent block
addition; the dBFT itself must wait for the neo-go to complete the block
addition and notify the dBFT, so that it can initialize at 0-th view to
collect the next block.
It has a special `requestF` and a special initialization function, but other
than that it's an absolutely regular WSClient. Can be used to call, can be
used to subscribe. Fixes#2909.
According to docs, `Server` uses provided error channel only to write
encountered error to it. In this case, there is no need to accept rw
channel to create `Server` instance. Strengthening the type to
write-only will allow the caller to ensure control of reading errors
from the provided channel.
The change is backward compatible since any `chan` is `chan<-`.
Signed-off-by: Leonard Lyubich <ctulhurider@gmail.com>
`Actor.MakeUnsignedUncheckedRun` method imposes restriction to
`CalculateNetworkFee` method's implementations: `Hash` or `Size` methods
must not be called on the pointer to the given transaction.
Add docs to adjust described requirement.
Signed-off-by: Leonard Lyubich <ctulhurider@gmail.com>
We have both from and to here, so technically we can either drop the neg/neg
trick from the processTokenTransfer() or drop one field from the structure
(the other side is a part of the key). Drop the field since this can make the
DB a bit more compact. Change Amount to be a pointer along the way since
that's the "native" thing for big.Int, we've used non-pointer field
specifically to avoid Neg/Neg problems, but it looks like this is not
necessary.
This structure is only used by the RPC server and I doubt anyone uses it via
the *Blockchain.
In some cases n.Add() can reuse the []Word buffer and n.Sub() reallocate it
away. If that happens, we're out of luck with 0.99.0+ versions (since
3945e81857). I'm not sure why it does that, bit
width doesn't change in most of the cases and even if it does, we still have
enough of it in cap() to hold the old Abs() value (when we have a negative
value we in fact decreate its Abs() first and increase it back
afterwards). Still, that's what we have.
So when we have processTokenTransfer() doing Neg/Neg in-place its value is not
affected, but the original []Word bits that are reused by amount value are
(they're shared initially, Amount: *amount).
name old time/op new time/op delta
ToPreallocatedBytes-8 65.8ns ± 2% 45.6ns ± 2% -30.73% (p=0.008 n=5+5)
name old alloc/op new alloc/op delta
ToPreallocatedBytes-8 0.00B 0.00B ~ (all equal)
name old allocs/op new allocs/op delta
ToPreallocatedBytes-8 0.00 0.00 ~ (all equal)
d5a9af5860 is incompatible with the NeoFS
mainnet sidechain, so we add the old logic to the pre-Aspidochelone
behaviour. Changing flags at newMethodAndPrice() is a bit less convenient
unfortunately because this will affect interop validity checks, so let's have
this kludge here.
6b4dd5703e made it to be a uint16 which was
somewhat important for RPC, but now it's irrelevant and the fact that it was a
string in some cases may lead to errors like these:
failed to unmarshal config YAML: yaml: unmarshal errors:
line 48: cannot unmarshal !!str `20011` into uint16
line 52: cannot unmarshal !!str `40001` into uint16
So for maximum backwards compatibility we better have string here and
eventually it'll be deleted anyway.
It directly affects node security and the default here MUST BE the safe choice
which is to do the verification. Otherwise it's just dangerous, absent any
VerifyBlocks configuration we'll get an insecure node. This option is not
supposed to be frequently used and it doesn't affect the ability to process
blocks, so breaking compatibility (in a safe manner) should be OK here.
And include some node-specific configurations there with backwards
compatibility. Note that in the future we'll remove Ledger's
fields from the ProtocolConfiguration and it'll be possible to access them in
Blockchain directly (not via .Ledger).
The other option tried was using two configuration types separately, but that
incurs more changes to the codebase, single structure that behaves almost like
the old one is better for backwards compatibility.
Fixes#2676.
It doesn't store id->hash mappings for native contracts. We need blockchain's
GetContractScriptHash to serve both anyway, so it was changed a bit. The only
other direct user of native.GetContractScriptHash is the VM CLI, but I doubt
anyone will use it for native contracts (they have ~zero VM code anyway).
There are no changes visible from the user side (at least for those
users who doesn't put Prometheus's or pprof's port in quotes), just
internal refactoring. From now and on, BasicService configuration is
used by RPC server config, TLS for RPC server, pprof and Prometheus.
It's more generic and convenient than MillisecondsPerBlock. This setting is
made in backwards-compatible fashion, but it'll override SecondsPerBlock if
both are used. Configurations are specifically not changed here, it's
important to check compatibility.
Fixes#2675.
Follow neo-project/neo#2807. Notice that this data is not cached, our previous
implementation wasn't too and it shouldn't be a problem (not on the hot path).
They can stay in the memory pool forever because consensus process will never
accept these transactions (and maybe even block consensus process at all).
We're paging these hashes, so we need a previous full page and a current one
plus some cache for various requests. Storing 1M of hashes is 32M of memory
and it grows quickly. It also seriously affects node startup time, most of
what it's doing is reading these hashes, the longer the chain the more time it
needs to do that.
Notice that this doesn't change the underlying DB scheme in any way.
If we only have genesis block (or <2000 headers) then we might as well use
generic logic below with zero targetHash because genesis block has zero
PrevHash (and its hash will naturally be the last on the chain going
backwards).
