-
# What
Since selector traversals are deterministic, the simplest way to resume a previous request is to simply tell the responder to not send the first N blocks, corresponding to the blocks you h…
-
### Issue
I am encountering stuck filecoin deal transfers, as well as I/O timeouts from dealbot. Upon further investigation, it looks like GraphSync's `attemptSendAndRecovery`'s resends have malforme…
-
This issue documents our initial goals for profiling go-graphsync performance with TestGround (https://github.com/testground/testground)
# Goals
Assess Graphsync performance in a variety of scen…
-
Define an interface for establishing a transfer session in a specific transport. (e.g. data-transfer/graphsync or http)
-
Can be seen @ https://github.com/ipfs/go-graphsync/pull/332
Relevant bit of schema looks like:
```ipldsch
type GraphSyncRequest struct {
id GraphSyncRequestID (rename "ID") …
rvagg updated
2 years ago
-
# What
We can simplify the ResponseManager by refactoring it to use the extracted TaskQueue logic. Even better if we can successfully extract the queryExecutor to its own module and unit test it (n…
-
Many legit use cases rely on sparse DAGs, understood as DAGs containing blocks not present in the responder's blockstore. Examples include federated/decentralised DAGs, blockchains nodes with limited …
raulk updated
2 years ago
-
Using Lotus markets version v1.11.3-rc1 graphsync v0.9.1 logs show these warnings during data transfers. It does not seem to affect the transfers.
2021-09-09T20:00:12.172Z WARN graphsync …
-
Two racy tests showed up in https://github.com/ipfs/go-graphsync/pull/244:
* `TestValidationAndExtensions/test_update_hook_processing/can_send_extension_data/when_paused`
* `TestValidationAndExten…
rvagg updated
2 years ago
-
# What
Request allocation backpressure was thrown in quite ad-hoc to respond to miner concerns about memory usage. (as opposed to response allocation backpressure which has seen pretty careful deve…