can not find candidate parents when peer proxy a image download

[bigeyefish]

Bug report:

when I use client proxy a image download, print a "can not find candidate parents" log and then download from source
It accours when pieces are cached on other peers。when I download from peer which has cached file, then not down from source

Expected behavior:

download from other peer who has cached pieces use P2P

How to reproduce it:

1. download a image on peer A
2. download same image on peer B (can not find candidate parents and download from source)

Environment:

• Dragonfly version: v2.2.0
• OS: linux
• Kernel (e.g. uname -a):
• Others:
  some schedule log

2025-01-11T02:43:06.956Z	INFO	scheduling/scheduling.go:136	send NeedBackToSourceResponse, because of scheduling exceeded RetryBackToSourceLimit 5	{"hostID": "192.168.100.5-server-1", "taskID": "c4356d4db905467cc4fbe52661a44b746e1f996401e9972888a29a6dedea585a", "peerID": "192.168.100.5-server-1-677912b8-c0a4-4181-ba1c-83e2cd3c6af0"}

2025-01-11T02:43:06.953Z	INFO	standard/peer.go:274	peer state is BackToSource	{"hostID": "192.168.100.5-server-1", "taskID": "cb40510188c4b553e053fda04ee5fcb2dcc2276a32954b5555922af6cbc9aafa", "peerID": "192.168.100.5-server-1-a6b84a64-b010-497e-ac43-f1f883869f84"}

2025-01-11T02:43:06.953Z	INFO	service/service_v2.go:123	receive DownloadPeerBackToSourceStartedRequest	{"hostID": "192.168.100.5-server-1", "taskID": "cb40510188c4b553e053fda04ee5fcb2dcc2276a32954b5555922af6cbc9aafa", "peerID": "192.168.100.5-server-1-a6b84a64-b010-497e-ac43-f1f883869f84"}

2025-01-11T02:43:06.952Z	INFO	scheduling/scheduling.go:136	send NeedBackToSourceResponse, because of scheduling exceeded RetryBackToSourceLimit 5	{"hostID": "192.168.100.5-server-1", "taskID": "cb40510188c4b553e053fda04ee5fcb2dcc2276a32954b5555922af6cbc9aafa", "peerID": "192.168.100.5-server-1-a6b84a64-b010-497e-ac43-f1f883869f84"}

2025-01-11T02:43:06.251Z	INFO	scheduling/scheduling.go:173	scheduling failed in 5 times, because of candidate parents not found	{"hostID": "192.168.100.5-server-1", "taskID": "cb40510188c4b553e053fda04ee5fcb2dcc2276a32954b5555922af6cbc9aafa", "peerID": "192.168.100.5-server-1-a6b84a64-b010-497e-ac43-f1f883869f84"}

2025-01-11T02:43:06.251Z	INFO	scheduling/scheduling.go:396	can not find candidate parents	{"hostID": "192.168.100.5-server-1", "taskID": "cb40510188c4b553e053fda04ee5fcb2dcc2276a32954b5555922af6cbc9aafa", "peerID": "192.168.100.5-server-1-a6b84a64-b010-497e-ac43-f1f883869f84"}

2025-01-11T02:43:06.251Z	INFO	scheduling/scheduling.go:575	filter candidate parents is []string(nil)	{"hostID": "192.168.100.5-server-1", "taskID": "cb40510188c4b553e053fda04ee5fcb2dcc2276a32954b5555922af6cbc9aafa", "peerID": "192.168.100.5-server-1-a6b84a64-b010-497e-ac43-f1f883869f84"}

2025-01-11T02:43:06.251Z	DEBUG	scheduling/scheduling.go:515	parent 192.168.100.5-server-1-a6b84a64-b010-497e-ac43-f1f883869f84 host 192.168.100.5-server-1 is not selected because it is in blocklist	{"hostID": "192.168.100.5-server-1", "taskID": "cb40510188c4b553e053fda04ee5fcb2dcc2276a32954b5555922af6cbc9aafa", "peerID": "192.168.100.5-server-1-a6b84a64-b010-497e-ac43-f1f883869f84"}

[bigeyefish]

it seems that the task dag only has one peer, that is request peer, so after exclude, candidate parents is empty

infact, cluster has 3 seed peer and 12 client peer, all seems healthy