Pruning Settings

[andreclaro]

Hello,

We just want to keep history for the last month, which in the case of Acala and Karura is about 200k blocks.

What are the correct parameters/flags to enable pruning? --pruning 200000 or --state-pruning 200000 --blocks-pruning 200000

Thanks!

[andreclaro]

Looking at the code (https://github.com/paritytech/polkadot-sdk/blob/7f7f5fa857502b6e3649081abb6b53c3512bfedb/substrate/client/cli/src/params/pruning_params.rs#L38-L39), looks like --pruning is equivalent to --state-pruning.

I started my node with --pruning 200000. Can I know restart the node with --blocks-pruning 200000 to enable blocks pruning? does it take affect or do I need to delete the existing db and start syncing from genesis?

[xlc]

This is not Acala specific so best to ask at http://github.com/paritytech/polkadot-sdk or http://substrate.stackexchange.com as there will be more people able to help.

To answer your question: --pruning is same as --state-pruning and it does not enable --block-pruning so you will both flag to prune both state and blocks.

Pruning mode cannot be changed AFAIK so you will need to resync the node

[andreclaro]

@xlc, Do I need to start the node from scratch to enable that pruning settings in polkadot db ?

[xlc]

you should resync the polkadot db by deleting the polkadot folder in the base path. but can keep the acala one

[andreclaro]

Ok! Thanks! I already deleted both databases... It should take a full week to have both nodes in-sync... is there any quick way to resync the nodes with pruning settings? 🤔

these are the settings I used (karura example):

/usr/local/bin/acala \
    --base-path=/var/lib/karura/data \
    --chain=karura \
    --name=xLabs \
# TODO(andre): this condition will remove when we deprecate the current nodes
# paritydb database is recommend when pruning is enabled
    --database=paritydb \
    --state-pruning=200000 \
    --blocks-pruning=200000 \
    --rpc-cors=all \
    --rpc-external \
    --rpc-port=9933 \
    --rpc-max-connections=500 \
    --port=30333 \
    --prometheus-external \
    --prometheus-port=9615 \
    -- \
    --database=paritydb \
    --state-pruning=200000 \
    --blocks-pruning=200000

Used --database=paritydb because I got the following warning message:

acala[1810229]: 2024-06-15 10:49:42 [Relaychain] Large pruning window of 200000 detected! THIS CAN LEAD TO HIGH MEMORY USAGE AND CRASHES. Reduce the pruning window or switch your database to paritydb.

[xlc]

I had same request and it isn't supported yet paritytech/polkadot-sdk#8

[andreclaro]

Acala mainnet is currently stuck on finalized block #5446885 .. For details consult #2774