Giters

node-real / bsc-erigon

Ethereum implementation on the efficiency frontier

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Erigon drop peers and stop synchronization

ivan-genev opened this issue · comments

commented

System information

Erigon version: v1.1.12

OS & Version: Linux/Docker

Commit hash: git_branch=HEAD git_tag=v1.1.12-dirty git_commit=f46bb114004c20f63e0d252c9d9c650b6da66c25

Erigon Command (with flags/config):
erigon --chain=bsc --private.api.addr=0.0.0.0:9090 --http=false --metrics --metrics.addr=0.0.0.0 --metrics.port=6060 --torrent.download.rate=70mb --datadir=/data/erigon --p2p.protocol=66 --bodies.cache=214748364800 --batchSize=4096M --db.pagesize=16k --sentry.drop-useless-peers

Chain/Network: bsc

Expected behavior

To reconnect to peers

Actual behavior

At some point erigon start drooping GoodPeers and synchronization stops

Steps to reproduce the behavior

Backtrace

[INFO] [01-22|14:00:49.080] [p2p] GoodPeers                          eth66=33
[INFO] [01-22|14:02:49.080] [p2p] GoodPeers                          eth66=32
[INFO] [01-22|14:04:49.080] [p2p] GoodPeers                          eth66=23
[INFO] [01-22|14:06:49.080] [p2p] GoodPeers                          eth66=11
[INFO] [01-22|14:08:49.080] [p2p] GoodPeers                          eth66=6
[INFO] [01-22|14:10:49.080] [p2p] GoodPeers                          eth66=3
[INFO] [01-22|14:12:49.080] [p2p] GoodPeers                          eth66=4
[INFO] [01-22|14:14:49.080] [p2p] GoodPeers                          eth66=8
[INFO] [01-22|14:16:49.080] [p2p] GoodPeers                          eth66=5
[INFO] [01-22|14:18:49.080] [p2p] GoodPeers                          eth66=3
[INFO] [01-22|14:20:49.080] [p2p] GoodPeers                          eth66=4
[INFO] [01-22|14:22:49.080] [p2p] GoodPeers                          eth66=4
[INFO] [01-22|14:24:49.080] [p2p] GoodPeers                          eth66=4
[INFO] [01-22|14:26:49.080] [p2p] GoodPeers                          eth66=1

even with one peer it will continue to synchronize for some time but at some point it will fail without errors and will print

[INFO] [01-23|09:17:05.773] [2/15 Headers] No block headers to write in this log period block number=35490443
[INFO] [01-23|09:17:05.773] Req/resp stats                           req=0 reqMin=0 reqMax=0 skel=18 skelMin=35490442 skelMax=35490634 resp=18 respMin=35490442 respMax=35490443 dups=36
[INFO] [01-23|09:17:25.773] [2/15 Headers] No block headers to write in this log period block number=35490443
[INFO] [01-23|09:17:25.774] Req/resp stats                           req=0 reqMin=0 reqMax=0 skel=18 skelMin=35490442 skelMax=35490634 resp=18 respMin=35490442 respMax=35490443 dups=36

restart fix the problem.

commented

Hi there, Try removing this --sentry.drop-useless-peers as it seems to work fine after that. JFYR: #249 (comment).

commented

I try that but synchronization didn't start at all even when it say that GoodPeers was 32.
Now I test one of out nodes without --p2p.protocol=66 and our second node is with only one GoodPeer for last 24H and still working.

commented

Try out these:

  • integration state_stages --unwind=100 --chain=bsc --datadir yourdatadir (JFYR: #258 (comment)),
  • Stop and run : integration stage_headers --reset, as it will load the missing header. (JFYR: #258 (comment))
  • Try not specify the body cache and batch size, these will bring heavy to RAM caches :--bodies.cache=214748364800 --batchSize=4096M (JFYR: #39 (comment))

Do share more logs if it still doesn't work.

commented

Use the latest commit in devel branch.

commented

I started it on our backup node, for now everything looks good.
ps: It was still working with this 1 peer :) my hero

commented

Reopen if still have problem