Boost Slow in API Response Causing Lotus-miner Info Command Timeout

[William8Work]

Checklist

• This is not a question or a support request. If you have any boost related questions, please ask in the discussion forum.
• This is not a new feature or enhancement request. If it is, please open a new idea discussion instead. New feature and enhancement requests would be entertained by the boost team after a thorough discussion only.
• I have searched on the issue tracker and the discussion forum, and there is no existing related issue or discussion.
• I am running the Latest release, or the most recent RC(release canadiate) for the upcoming release or the dev branch(master), or have an issue updating to any of these.
• I did not make any code changes to boost.

Boost component

• boost daemon - storage providers
• boost client
• boost UI
• boost data-transfer
• boost index-provider
• Other

Boost Version

boostd version 1.5.1-rc2+git.2c5604e
boost version 1.5.1-rc2+git.2c5604e

Describe the Bug

once a while, the lotus-miner info command errors out on getting information back from boost API.

Lotus-miner info command

Workers: Seal(10) WdPoSt(0) WinPoSt(0)
ERROR: %!s(PANIC=Error method: runtime error: invalid memory address or nil pointer dereference)

Logging Information

whenever that lotus-miner info command erroring out, there is a corresponding error entry in boost log:


Boost log:
2023-02-14T10:55:05.891-0800    ERROR   rpc     go-jsonrpc@v0.0.0-20221118154859-f3f8e9b1cb11/websocket.go:129  handle me:writev tcp4 10.1.18.168:8787->10.1.18.180:45346: writev: connection reset by peer

### Repo Steps

This happens frequently when both miner and boost are stopped, then restart miner, once it's up and running, start boostd. And then execute the command lotus-miner info.

This also happens once a while when both lotus-miner and boost are running fine, workload are light (no sealing activities). 
...

[namoo-cloud]

lotus-miner info

2023-02-22T19:07:15.732+0900 WARN rpc go-jsonrpc@v0.2.1/websocket.go:763 websocket long time no response {"lastAction": "incoming(,3)", "time": 15.084658323}
Retrieval Deals (complete): 462700, 32.73 TiB

[dirkmc]

ERROR: %!s(PANIC=Error method: runtime error: invalid memory address or nil pointer dereference)

The error message is printed out by lotus cli helper when an error is returned from a CLI Action:

fmt.Fprintf(os.Stderr, "ERROR: %s\n\n", err)

The format of the error message %!s(PANIC=...nil pointer indicates that fmt.Fprintf encountered a nil pointer when trying to call the Error() method on the golang error object that was passed to it. This is most likely because the error object wraps a nil error. An example golang Test that reproduces this behaviour:

type myerr struct {
	err error
}

func (e *myerr) Error() string {
	// if e.err is nil, this will panic
	return e.err.Error()
}

func TestPrintErr(t *testing.T) {
	fmt.Printf("%s\n", &myerr{})
}

=== RUN   TestPrintErr
%!s(PANIC=Error method: runtime error: invalid memory address or nil pointer dereference)
--- PASS: TestPrintErr (0.00s)

I would like to reassign this bug to go-jsonrpc because it looks like the underlying implementation of go-jsonrpc is returning an error object that wraps a nil error.

[dirkmc]

Added an issue to the lotus repo to Remove markets API calls from lotus-miner info

[dirkmc]

I opened filecoin-project/go-jsonrpc#96 to track the underlying go-jsonrpc issue.
filecoin-project/lotus#10364 takes care of this issue in lotus.