Pruning fullnodes / node managament

[ThomasBlock]

Hey. I am Still having massive problems keeping my nodes in sync
( as a remidner: my archival node dilutes the size by a fator of two . the only availbale snapshot is from last year )

Now i realize my fullnodes are also have problems: the size is growing and growing ( 2.3 TB after a month ) so i tought i will just use --init.prune full . but no, see error below. i can say fur quite certain that my node never crashed. so why can i not prune?

my only solution for now is two have TWO arbitrum nodes of each type

• i run node A until its too big
• then i start node B from scratch and switch clients to node B
• when node B is too big, i start a fresh node A and switch clients
• etc

Started Arbitrum Nitro Service.
INFO [11-24|21:40:43.804] using existing jwt file                  filename=/root/.arbitrum/jwtsecret
INFO [11-24|21:40:43.804] Running Arbitrum nitro node              revision=d81324d-modified vcs.time=2024-09-24T15:12:36Z
INFO [11-24|21:40:43.805] connected to l1 chain                    l1url=http://192.168.128.46:8545 l1chainid=1
INFO [11-24|21:40:43.806] Starting metrics server                  addr=http://0.0.0.0:6070/debug/metrics
INFO [11-24|21:40:44.250] Using leveldb as the backing database
INFO [11-24|21:40:44.250] Allocated cache and file handles         database=/root/.arbitrum/arb1/nitro/l2chaindata cache=16.00MiB handles=16 readonly=true
INFO [11-24|21:40:46.173] Using LevelDB as the backing database
INFO [11-24|21:40:46.173] Found legacy ancient chain path          location=/root/.arbitrum/arb1/nitro/l2chaindata/ancient
INFO [11-24|21:40:46.202] Opened ancient database                  database=/root/.arbitrum/arb1/nitro/l2chaindata/ancient readonly=true
INFO [11-24|21:40:46.663] Using leveldb as the backing database
INFO [11-24|21:40:46.663] Allocated cache and file handles         database=/root/.arbitrum/arb1/nitro/l2chaindata cache=2.00GiB handles=512
INFO [11-24|21:40:49.699] Using LevelDB as the backing database
INFO [11-24|21:40:49.700] Found legacy ancient chain path          location=/root/.arbitrum/arb1/nitro/l2chaindata/ancient
INFO [11-24|21:40:49.704] Opened ancient database                  database=/root/.arbitrum/arb1/nitro/l2chaindata/ancient readonly=false
INFO [11-24|21:40:49.751] Using pebble as the backing database
INFO [11-24|21:40:49.751] Allocated cache and file handles         database=/root/.arbitrum/arb1/nitro/wasm cache=2.00GiB handles=512
INFO [11-24|21:40:49.779] State scheme set by user                 scheme=hash
INFO [11-24|21:40:49.840] Using leveldb as the backing database
INFO [11-24|21:40:49.840] Allocated cache and file handles         database=/root/.arbitrum/arb1/nitro/arbitrumdata cache=16.00MiB handles=16 readonly=true
INFO [11-24|21:40:50.085] Using LevelDB as the backing database
INFO [11-24|21:41:34.940] looking for old block with state to keep current=277,887,000 target=277,889,214
INFO [11-24|21:42:44.753] looking for old block with state to keep current=277,884,000 target=277,889,214
INFO [11-24|21:43:27.616] looking for old block with state to keep current=277,881,000 target=277,889,214
INFO [11-24|21:44:15.098] looking for old block with state to keep current=277,878,000 target=277,889,214
INFO [11-24|21:44:59.493] looking for old block with state to keep current=277,875,000 target=277,889,214
INFO [11-24|21:45:49.865] found pruning target blocks              heights="[22207817 277872915]" roots="[0x7f2bfc4481d02bfcfc606ebb949384ef78d03a0f30a2dc9cccd652eb80926ae1 0x7a7b73f0ffb7058e41d56aaa55a3a7f92d118d8dbde630033f31f0d3d5a1b6d0 0x0000000000000000000000000000000000000000000000000000000000000000]"
WARN [11-24|21:45:49.961] Loaded snapshot journal                  diskroot=7c1cb4..56d391 diffs=unmatched
WARN [11-24|21:45:49.961] Failed to load snapshot                  err="head doesn't match snapshot: have 0x7c1cb46d7236d12731d16ddeeff4a89c6f43850763c846bdf52a7606f356d391, want 0xea82ebdaf345562a2cb2f1815ded233fc55c28811a84ece969a7d9b6515a8f45"
Usage of /root/nitro/target/bin/nitro:
ERROR[11-24|21:45:49.961] error initializing database              err="error pruning: head doesn't match snapshot: have 0x7c1cb46d7236d12731d16ddeeff4a89c6f43850763c846bdf52a7606f356d391, want 0xea82ebdaf345562a2cb2f1815ded233fc55c28811a84ece969a7d9b6515a8f45"
nitro.service: Main process exited, code=exited, status=1/FAILURE

it will reboot indefinetly with the same snapshot error

[snf]

I just found this issue with a freshly downloaded snapshot from https://snapshot-explorer.arbitrum.io/ :

Size: 1.06 TB

Snapshot Date: January 16, 2025

Download Links:
[Download Part 0000 (500 GB)](https://snapshot.arbitrum.io/arb1/2025-01-16-38995543/pruned.tar.part0000)

SHA256:

724f3b8034fb6af0cf2d3065f1ac63d72a3276985bb9dc79f93954e0c36d22dc
[Download Part 0001 (500 GB)](https://snapshot.arbitrum.io/arb1/2025-01-16-38995543/pruned.tar.part0001)

SHA256:

0fb4f1a90c4e9cb215d958c132b061dc4d1f460a73dfccd6660b3f048d25bc1d
[Download Part 0002 (84.8 GB)](https://snapshot.arbitrum.io/arb1/2025-01-16-38995543/pruned.tar.part0002)

SHA256:

dc9ffd0ab41f48a418e825f1aa342fdc94597e9ba60cd1b404371bcd234c6e68

Is it possible that there is something wrong and this is not a database corruption? I got this error and after re-downloading I got the same error again.

[joshuacolvin0]

Be certain to recombine the parts when extracting tar file

[snf]

Yes, that's fine. The node works fine. The pruning doesn't work.

To reproduce:

1. Download latest snapshot
2. Start with pruning
3. It fails

[joshuacolvin0]

The downloaded snapshot is already pruned, we don't support pruning already pruned snapshots

[snf]

That doesn't sound right. Pruning should always succeed. Worst case there shouldn't be anything to prune and the node should start without removing anything instead of failing to start, right?