Cardano-db-sync - almost tables are empty

HamsterStakePool · 29 June 2022 18:37

I set up cardano_db_sync with docker-compose. So far everything works and I don’t get any errors displayed. The table block and epoch are gradually being filled, but I have the feeling that something is wrong.

All tables pool_hash, pool_owner, pool…, as well as reward, epoch_stake, epoch_param and ada_pots are not filled. I read in the documentation that you can turn this off with --disable-ledger.

github.com

input-output-hk/cardano-db-sync/blob/master/doc/configuration.md#--disable-ledger

# Configuration

The initial design of db-sync was a one size fits all approach. It served as a general-purpose
backend for Cardano applications, including light wallets, explorers, etc. Over time, many new
features have been added, including historic rewards with Shelley, scripts with Allegra, multiassets
with Mary, Plutus scripts and redeemers with Alonzo, stake pool metadata with the integration of
SMASH, etc.

While db-sync needs to use the resources that all these features require, many applications use only
a small fraction of these features. Therefore, it is reasonable to introduce flags and options that
turn off some of these features, especially the most expensive ones in terms of performance. These
configuration options require proper documentation, which is presented below.

### --disable-ledger

One of the db-sync features that uses the most resources is that it maintains a ledger state and
replays all the ledger rules. This is the only way to get historic reward details and other data
that is not included in the blocks (ie. historic stake distribution, ada pots, epoch parameters,
etc). The flag --disable-ledger provides the option to turn off these features and significantly
reduce memory usage (by up to 10GB on mainnet) and sync time. Another benefit of this option is

This file has been truncated. show original

But I didn’t turn that off consciously. I can’t find anything in the config files either. Can someone give me a tip on how to activate this?

I think the pool_* tables are populated by the smash_server. Am I right?

DinoDude · 30 June 2022 04:32

You need to run the cardano node with smash server and db-sync to get everything: cardano-db-sync/smash.md at master · input-output-hk/cardano-db-sync · GitHub

Also note this could takes days or weeks to fully sync depending on your internet bandwidth and you will probably need 100GB free to grow the SQL database alongside the ledger which is why most opt to disable as much data as possible.

HamsterStakePool · 30 June 2022 19:06

Thanks for your answer. Seems that the docker image (Docker Hub) is not maintained anymore. Therefore I will compile it on myself an try it out. I let you know how far I get.

HamsterStakePool · 1 July 2022 07:08

@DinoDude thank you for your helpful advice. I followed the instruction you have mentioned and additionally started the smash server. Now the tables get filled. Btw. I started cardano_db_sync two days ago. Now, I am at epoch 246. So there are above100 left

DB diskspace usage (with a rising trend):
du -s -h /workspace/postgresql/data/
11G /workspace/postgresql/data/

I think it will be around 20GB when the sync is done.

Plus ~70GB node data

DinoDude · 2 July 2022 17:28

Yup, I was just guessing on the amount of data and time it will take but it’s A LOT so hopefully you got plenty of hard drive space and patience

Topic		Replies	Views
Using cardano-db-sync-extended with docker-compose Cardano Integration cardano-node	1	464	16 November 2021
Db-sync cexplorer db has no relations Cardano Integration	1	412	28 April 2022
No data in Rewards table of database Cardano Integration mainnet	2	648	1 October 2020
Get Historic Number of Stake Pools via cardano-db-sync Education analysis	0	452	8 August 2021
Cardano-db-sync: Offline pool metadata fetch: fetch errors Cardano Integration	3	838	15 March 2022

Cardano-db-sync - almost tables are empty

Related Topics