As it stands, we are continuously making improvements to both stability and memory usage. So please make sure you keep your client up to date! This means restarting your node and updating your software regularly from the
master branch. If you can't find a solution to your problem here, feel free to hit us up on our discord!
Note: While the
masterbranch of the
nim-beacon-chainrepository is more stable, the latest updates happen in the
develbranch which is (usually) merged into master every week on Tuesday. If you choose to run Nimbus directly from the
develbranch, be prepared for instabilities!
To update and restart, run
make update, followed by
cd nim-beacon-chain git pull make update # Update dependencies make medalla # Restart using same keys as last run
If you find that
make update causes the console to hang for too long, try running
make update V=1 or
make update V=2 instead (these will print a more verbose output to the console which may make it easier to diagnose the problem).
Note: rest assured that when you restart the beacon node, the software will resume from where it left off, using the validator keys you have already imported.
The directory that stores the blockchain data of the testnet is
build/data/shared_medalla_0 (if you're connecting to another testnet, replace
medalla with that testnet's name). Delete this folder to start over (for example, if you started building medalla with the wrong private keys).
If you’re experiencing sync problems, we recommend running
make clean-medalla to delete the database and restart your sync (make sure you’ve updated to the latest
master first though).
make clean-medallawill erase all of your syncing progress so far, so it should only be used as a last resort -- if your client gets stuck for a long time (because it's unable to find the right chain and/or stay with the same head value) and a normal restart doesn't improve things.
If you're running out of storage, you can prune the database of unnecessary blocks and states by running:
make ncli_db build/ncli_db pruneDatabase --db=build/data/shared_medalla_0/db --verbose=true
This will create
nbc_pruned.sqlite3 files in
build/data/shared_medalla_0/db, which you can use in place of the orginal
nbc.sqlite3 files. We recommend you hold onto the originals until you've verified that your validator is behaving as expected with the pruned files.
--keepOldStates(boolean): Keep pre-finalisation states; defaults to
--verbose(boolean): Print a more verbose output to the console; defaults to
As it stands, logging seems to be slowing down the client, and quite a few users are experiencing trouble either catching up or keeping up with the head of the chain. You can use either the
LOG_LEVEL=NOTICE options to reduce verbosity and speed up the client (
NOTICE is even less verbose than
make LOG_LEVEL=INFO medalla
If you're experiencing a low peer count, you may be behind a firewall. Try restarting your client and passing
NODE_PARAMS="--nat:\"extip:$EXT_IP_ADDRESS\"" as an option to
make medalla, where
$EXT_IP_ADDRESS is your real IP. For example, if your real IP address is
18.104.22.168, you'd run:
make NODE_PARAMS="--nat:\"extip:22.214.171.124\"" medalla
If you're experiencing RAM related resource leaks, try restarting your client (we recommend restarting every 6 hours until we get to the bottom of this issue). If you have a local Grafana setup, you can try monitoring the severity of these leaks and playing around with the restart interval.
If you're seeing an error that looks like:
Error: unhandled exception: (98) Address already in use [TransportOsError]
It's probably because you're running multiple validators -- and the default base port
9000 is already in use.
To change the base port, run:
make BASE_PORT=9100 medalla
(You can replace
9100 with a port of your choosing)