* Add CLI flags to config LevelDB table/total sizes
I wired up CLI flags to allow configuring LevelDB table and total sizes:
- `--leveldb.compaction.table.size`, LevelDB SSTable file size factor in MiB (default: 2)
- `--leveldb.compaction.table.multiplier`, multiplier on LevelDB SSTable file size (default: 1)
- `--leveldb.compaction.total.size`, total size factor in MiB of LevelDB levels (default: 10)
- `--leveldb.compaction.total.multiplier`, multiplier on LevelDB total level size (default: 10)
N.B. that the default values for these configs are exactly the same as
before this changset and so Bor behavior should not change unless these
flags are deliberately overridden. Bor/Geth inherited the default values
from [the `goleveldb`
defaults](126854af5e/leveldb/opt/options.go).
We (Alchemy) found it necessary to override these configs as follows to
keep Bor archive nodes tracking the canonical chain:
- `--leveldb.compaction.table.size=4`
- `--leveldb.compaction.total.size=20`
These overrides double the size of LevelDB SSTable files (2 MiB -> 4
MiB) and also the total amount of data in each level (100 MiB -> 200
MiB, 1,000 MiB -> 2,000 MiB, etc.). The idea is to have LevelDB read and
write data in larger chunks while keeping the proportional frequency of
compaction operations the same as in the original defaults defined by
Dean and Ghemawat.
Without these overrides we found that our archive nodes would tend to
fall into a "LevelDB compaction loop of death" where the incoming stream
of blockchain data could not be flowed into LevelDB's structure quickly
enough, resulting in the node blocking writes for long periods of time
while LevelDB's single-threaded compaction organized the data. Over
time the nodes would fall farther and farther behind the canonical chain
head, metaphorically dying a slow node's death.
These configs can be changed on existing node databases (resyncing is
not necessary). LevelDB appears to work correctly with SSTable files of
different sizes. Note that the database does not undergo any sort of
migration when changing these configs. Only newly-written files (due to
new data or compaction) are affected by these configs.
* Update docs
* Adjust line spacing for linter
* Replace map with `ExtraDBConfig`
* Rename `LevelDbConfig` to `ExtraDBConfig`
* Regenerate docs
* core/blobpool: implement txpool for blob txs
* core/txpool: track address reservations to notice any weird bugs
* core/txpool/blobpool: add support for in-memory operation for tests
* core/txpool/blobpool: fix heap updating after SetGasTip if account is evicted
* core/txpool/blobpool: fix eviction order if cheap leading txs are included
* core/txpool/blobpool: add note as to why the eviction fields are not inited in reinject
* go.mod: pull in inmem billy form upstream
* core/txpool/blobpool: fix review commens
* core/txpool/blobpool: make heap and heap test deterministic
* core/txpool/blobpool: luv u linter
* core/txpool: limit blob transactions to 16 per account
* core/txpool/blobpool: fix rebase errors
* core/txpool/blobpool: luv you linter
* go.mod: revert some strange crypto package dep updates
This change removes PoW header syncing related code from LES and also deletes
duplicated packages les/catalyst, les/downloader and les/fetcher. These package copies
were created because people wanted to make changes in their eth/ counterparts, but weren't
able to adapt LES code to the API changes.
It is usually best to set GOMAXPROCS to the number of available CPU cores. However, setting
it like that does not work well when the process is quota-limited to a certain number of CPUs.
The automaxprocs library configures GOMAXPROCS, taking such limits into account.
This simplifies the code that initializes the discovery a bit, and
adds new flags for enabling/disabling discv4 and discv5 separately.
---------
Co-authored-by: Felix Lange <fjl@twurst.com>
This change adds back the 'geth --dev' mode of operation, using a cl-mocker.
---------
Co-authored-by: Martin Holst Swende <martin@swende.se>
Co-authored-by: rjl493456442 <garyrong0905@gmail.com>
Co-authored-by: lightclient <14004106+lightclient@users.noreply.github.com>
The clean trie cache is persisted periodically, therefore Geth can
quickly warmup the cache in next restart.
However it will reduce the robustness of system. The assumption is
held in Geth that if the parent trie node is present, then the entire
sub-trie associated with the parent are all prensent.
Imagine the scenario that Geth rewinds itself to a past block and
restart, but Geth finds the root node of "future state" in clean
cache then regard this state is present in disk, while is not in fact.
Another example is offline pruning tool. Whenever an offline pruning
is performed, the clean cache file has to be removed to aviod hitting
the root node of "deleted states" in clean cache.
All in all, compare with the minor performance gain, system robustness
is something we care more.
The state availability is checked during the creation of a state reader.
- In hash-based database, if the specified root node does not exist on disk disk, then
the state reader won't be created and an error will be returned.
- In path-based database, if the specified state layer is not available, then the
state reader won't be created and an error will be returned.
This change also contains a stricter semantics regarding the `Commit` operation: once it has been performed, the trie is no longer usable, and certain operations will return an error.
This PR adds server-side limits for JSON-RPC batch requests. Before this change, batches
were limited only by processing time. The server would pick calls from the batch and
answer them until the response timeout occurred, then stop processing the remaining batch
items.
Here, we are adding two additional limits which can be configured:
- the 'item limit': batches can have at most N items
- the 'response size limit': batches can contain at most X response bytes
These limits are optional in package rpc. In Geth, we set a default limit of 1000 items
and 25MB response size.
When a batch goes over the limit, an error response is returned to the client. However,
doing this correctly isn't always possible. In JSON-RPC, only method calls with a valid
`id` can be responded to. Since batches may also contain non-call messages or
notifications, the best effort thing we can do to report an error with the batch itself is
reporting the limit violation as an error for the first method call in the batch. If a batch is
too large, but contains only notifications and responses, the error will be reported with
a null `id`.
The RPC client was also changed so it can deal with errors resulting from too large
batches. An older client connected to the server code in this PR could get stuck
until the request timeout occurred when the batch is too large. **Upgrading to a version
of the RPC client containing this change is strongly recommended to avoid timeout issues.**
For some weird reason, when writing the original client implementation, @fjl worked off of
the assumption that responses could be distributed across batches arbitrarily. So for a
batch request containing requests `[A B C]`, the server could respond with `[A B C]` but
also with `[A B] [C]` or even `[A] [B] [C]` and it wouldn't make a difference to the
client.
So in the implementation of BatchCallContext, the client waited for all requests in the
batch individually. If the server didn't respond to some of the requests in the batch, the
client would eventually just time out (if a context was used).
With the addition of batch limits into the server, we anticipate that people will hit this
kind of error way more often. To handle this properly, the client now waits for a single
response batch and expects it to contain all responses to the requests.
---------
Co-authored-by: Felix Lange <fjl@twurst.com>
Co-authored-by: Martin Holst Swende <martin@swende.se>
* core/txpool: abstraction prep work for secondary pools (blob pool)
* core/txpool: leave subpool concepts to a followup pr
* les: fix tests using hard coded errors
* core/txpool: use bitmaps instead of maps for tx type filtering
* cryto/kzg4844: pull in the C and Go libs for KZG cryptography
* go.mod: pull in the KZG libraries
* crypto/kzg4844: add basic becnhmarks for ballpark numbers
* cmd, crypto: integrate both CKZG and GoKZG all the time, add flag
* cmd/utils, crypto/kzg4844: run library init on startup
* crypto/kzg4844: make linter happy
* crypto/kzg4844: push missing file
* crypto/kzg4844: fully disable CKZG but leave in the sources
* build, crypto/kzg4844, internal: link CKZG by default and with portable mode
* crypto/kzg4844: drop verifying the trusted setup in gokzg
* internal/build: yolo until it works?
* cmd/utils: make flag description friendlier
Co-authored-by: Martin Holst Swende <martin@swende.se>
* crypto/ckzg: no need for double availability check
* build: tiny flag cleanup nitpick
---------
Co-authored-by: Martin Holst Swende <martin@swende.se>
rename parameter
In this case, the naming of "extapi" might create some confusion. Although it represents an External Signer Backend, its name could be mistaken for an API. In reality, it is a backend instance used for communicating with external signers. A better naming choice could be "extBackend" or "externalBackend" to more accurately describe that it is a backend instance rather than an API.
Makes the `geth account ... ` commands usable even if a geth-process is already executing, since the account commands do not read the chaindata, it was not required for those to use the same locking mechanism.
---
Signed-off-by: jsvisa <delweng@gmail.com>
Co-authored-by: Martin Holst Swende <martin@swende.se>
Co-authored-by: Sina Mahmoodi <itz.s1na@gmail.com>
* cmd/utils, node: switch to Pebble as the default db if none exists
* node: fall back to LevelDB on platforms not supporting Pebble
* core/rawdb, node: default to Pebble at the node level
* cmd/geth: fix some tests explicitly using leveldb
* ethdb/pebble: allow double closes, makes tests simpler
This change prints out more information about the problem, in the case where geth detects a gap between leveldb and ancients, so we can determine more exactly where the gap is (what the first missing is). Also prints out more metadata.
---------
Co-authored-by: Martin Holst Swende <martin@swende.se>
