fix: check if reserve is configured

[acha-bill]

Checklist

• I have read the coding guide.
• My change requires a documentation update, and I have done it.
• I have added tests to cover my changes.
• I have filled out the description and linked the related issues.

Description

Check if reserve is configured before attempting to put chunk

Related Issue (Optional)

#4829

[istae]

this is not a real solution, i think. we need to first understand why the reserve is nil. Chances are this is a light node and it was unable to find a peer to push the chunk to so it tried storing it. In that case, we need to plug that hole, where a light node should not try to store a chunk in the above scenario. but this is one theory. we need to ask the issue creator the config options he was running with before hastily jumping onto a solution.

[acha-bill]

I also got this from this failed beekeeper test

[pod/bootnode-0-0/bee] "time"="2024-10-21 06:07:39.750144" "level"="debug" "logger"="node/pusher" "v"=1 "msg"="chunk stays here, i'm the closest node" "chunk_address"="d43711cd5b3436c09ef09cb4e554eba026719af857220febc41f1db9fa311602"
[pod/bootnode-0-0/bee] panic: runtime error: invalid memory address or nil pointer dereference
[pod/bootnode-0-0/bee] [signal SIGSEGV: segmentation violation code=0x1 addr=0x50 pc=0xf33756]
[pod/bootnode-0-0/bee] 
[pod/bootnode-0-0/bee] goroutine 2528 [running]:
[pod/bootnode-0-0/bee] github.com/ethersphere/bee/v2/pkg/storer/internal/reserve.(*Reserve).Put(0x0, {0x1a55ca0, 0xc0043c9470}, {0x1a64ba0, 0xc004684870})
[pod/bootnode-0-0/bee] 	github.com/ethersphere/bee/v2/pkg/storer/internal/reserve/reserve.go:104 +0x76
[pod/bootnode-0-0/bee] github.com/ethersphere/bee/v2/pkg/storer.(*DB).ReservePutter.func1({0x1a55ca0?, 0xc0043c9470?}, {0x1a64ba0, 0xc004684870})
[pod/bootnode-0-0/bee] 	github.com/ethersphere/bee/v2/pkg/storer/reserve.go:300 +0x53
[pod/bootnode-0-0/bee] github.com/ethersphere/bee/v2/pkg/storage.PutterFunc.Put(0x4ed33c?, {0x1a55ca0?, 0xc0043c9470?}, {0x1a64ba0?, 0xc004684870?})
[pod/bootnode-0-0/bee] 	github.com/ethersphere/bee/v2/pkg/storage/chunkstore.go:56 +0x37
[pod/bootnode-0-0/bee] github.com/ethersphere/bee/v2/pkg/storer.putterWithMetrics.Put({{0x1a486c0, 0xc003b07470}, {{0xc0008c74d0}, {0xc0008c7560}, {0x1a641a8, 0xc0008ec340}, {0x1a641a8, 0xc0008ec3c0}, {0x1a5bf18, 0xc0008f8240}, ...}, ...}, ...)
[pod/bootnode-0-0/bee] 	github.com/ethersphere/bee/v2/pkg/storer/metrics.go:181 +0xb4
[pod/bootnode-0-0/bee] github.com/ethersphere/bee/v2/pkg/pusher.(*Service).pushDirect(0xc0008e84e0, {0x1a55ca0, 0xc0043c9470}, {0x1a64008, 0xc00137c730}, 0xc0043c9440)
[pod/bootnode-0-0/bee] 	github.com/ethersphere/bee/v2/pkg/pusher/pusher.go:321 +0x372
[pod/bootnode-0-0/bee] github.com/ethersphere/bee/v2/pkg/pusher.(*Service).chunksWorker.func2(0xc0043c9440)
[pod/bootnode-0-0/bee] 	github.com/ethersphere/bee/v2/pkg/pusher/pusher.go:161 +0x1f4
[pod/bootnode-0-0/bee] created by github.com/ethersphere/bee/v2/pkg/pusher.(*Service).chunksWorker in goroutine 300
[pod/bootnode-0-0/bee] 	github.com/ethersphere/bee/v2/pkg/pusher/pusher.go:230 +0x619

Config

• full node mode: true
• bootnode-mode: true

The reverse is not configured because

bee/pkg/node/node.go

Line 730 in 0e42d83

if o.FullNodeMode && !o.BootnodeMode {

But the pushsync only uses o.fullNodeMode

bee/pkg/node/node.go

Line 953 in 0e42d83

pushSyncProtocol := pushsync.New(swarmAddress, networkID, nonce, p2ps, localStore, waitNetworkRFunc, kad, o.FullNodeMode, pssService.TryUnwrap, validStamp, logger, acc, pricer, signer, tracer, warmupTime)

bee/pkg/pushsync/pushsync.go

Lines 401 to 406 in 0e42d83

	if ps.fullNode {
	if cac.Valid(ch) {
	go ps.unwrap(ch)
	}
	return nil, topology.ErrWantSelf
	}

So initializing pushsync consistently with o.FullNodeMode && !o.BootnodeMode should fix it

[istae]

does that mean that an upload was attempted from a bootnode?

[martinconic]

This is a very good finding. But I think that when I investigated the issue, actually the reserve seamed not to be nil, else it would have crashed even earlier. Please can you confirm that the reserve is not used in earlier places?

[acha-bill]

does that mean that an upload was attempted from a bootnode?

yes. See https://github.com/ethersphere/bee/actions/runs/11434065514/job/31807011143

msg="soc: submitting soc chunk xx to node bootnode-0"

Please can you confirm that the reserve is not used in earlier places?

I don't see any other place. The stack trace indicates that the receiver is nil.

(*Reserve).Put(0x0,....)