System information

Geth version: geth version 1.5.16
OS & Version: Linux
Commit hash : (if develop)

ISSUE

We run a node service platform and have several BSC nodes all running on the same hardware. NVMe drives, 32 Core Ryzen CPUs, 192GB DDR4 Ram.

Since updating to the latest version not a single one of our nodes will stay in sync for more than a few hours. We are seeing several logs relating to finding "too slow peer" and also several "LargeTx execution time". This has crippled our service. Can anyone please help me find the issue?

Actual behaviour

Nodes fall out of sync after syncing up within 1 - 2 hours even when no RPC requests are made to the node.

LOGS
t=2025-06-13T19:27:46+0000 lvl=info msg="LargeTX execution time" block=51011185 tx=0xa08fb7e70c8a61be947088d86186090d7d4a20183c78d9c311a44bd3b7f67bc0 gasUsed=18155382 elapsed=12.801763ms
t=2025-06-13T19:27:46+0000 lvl=info msg="LargeTX execution time" block=51011188 tx=0x44d6db9cfc273dcf57e85335c5b14b468766b202f178380a69f7aae8920acdf6 gasUsed=18183717 elapsed=12.388213ms
t=2025-06-13T19:27:47+0000 lvl=info msg="LargeTX execution time" block=51011197 tx=0xf87d96ad5d709b08d9cc5230a52efceac7df3d789f19361dde5887e2ef702f3b gasUsed=18194952 elapsed=13.144402ms
t=2025-06-13T19:27:47+0000 lvl=info msg="LargeTX execution time" block=51011201 tx=0x251fcbb13e443fe4418d3d39a04e8c58b221983e99ced19d068424fd7e097b5d gasUsed=18177852 elapsed=12.949802ms
t=2025-06-13T19:27:50+0000 lvl=info msg="Imported new chain segment" number=51011233 hash=0x0ad30411ac2c099989c257f0dc6356c373a90c80503ed22ee2f35c9b30beb96d miner=0x5009317FD4F6F8FeEa9dAe41E5F0a4737BB7A7D5 blocks=88 txs=26005 blobs=32 mgas=5587.062619 elapsed=8.003s mgasps=698.0984494670513 age=6d17h40m snapdiffs="13.97 MiB" triedirty="0.00 B" trieimutabledirty="0.00 B"

t=2025-06-13T19:26:59+0000 lvl=info msg="Imported new chain segment" number=51399140 hash=0x4437074e5f3c431395650affe75d1636f1b72aba06e982389579523cd7827bf5 miner=0x5009317FD4F6F8FeEa9dAe41E5F0a4737BB7A7D5 blocks=1 txs=137 blobs=0 mgas=22.68884 elapsed=44.228ms mgasps=512.9890424324674 snapdiffs="10.30 MiB" triedirty="0.00 B" trieimutabledirty="0.00 B"
t=2025-06-13T19:26:59+0000 lvl=warn msg="find a too slow peer" id=19ccac7f7eb168b4e3ed26df770e52e3dfd9335471906f2590c69df602d8ad86 peer=37.27.112.222:30305 latency=1168
t=2025-06-13T19:27:00+0000 lvl=info msg="Imported new chain segment" number=51399141 hash=0xd040dd20d131347a9253acddc59119c2d2dda0b2db063a6360d2fdc2e42467e2 miner=0x5009317FD4F6F8FeEa9dAe41E5F0a4737BB7A7D5 blocks=1 txs=141 blobs=0 mgas=22.740673 elapsed=41.602ms mgasps=546.6181538583129 snapdiffs="10.34 MiB" triedirty="0.00 B" trieimutabledirty="0.00 B"

I have also noticed this log every few minutes.

t=2025-06-13T19:25:34+0000 lvl=error msg="Error in block freeze operation" err="canonical hash missing, can't freeze block 48574071"

Flags:

--port 30304 --http.port 8533 --ws.port 8534 --config /mnt/data/bnb/config.toml --datadir /mnt/data/bnb/fast --history.transactions=90000 --syncmode=snap --db.engine=pebble --tries-verify-mode=none --http --http.api eth,net,web3 --ws --ws.api eth,net,web3