Cncli.sh leaderlog is killing the block producing node

Staking & Delegation Operate a Stake Pool
1

Every time i run cncli.sh leaderlog the node restarts. Is this expected?

2

If u have not enough RAM, it is normal

Type free -m

3

Plenty of RAM:
total used free shared buff/cache available
Mem: 16008 222 14771 0 1014 15520
Swap: 2047 35 2012

4

Ok, 16G + 2 is not enough

Increase the swap file to 6G

sudo swapoff /swapfile
sudo fallocate -l 6G /swapfile
sudo chmod 600 /swapfile
sudo mkswap /swapfile
sudo swapon /swapfile

Now test again with CNCLI

5

The node didn’t restart this time. However, I’m not sure that cncli did anything this time:
Leaderlogs already calculated for epoch 317, skipping!

I’ll try it again in 5 days to see if the increasing the swap file size resolved the issue.

May the swap size impact the regular node operations? Would it make sense to increase it on the relay nodes as well?

Thank you for your time, Alex!

6

Try ./cncli.sh leaderlog force (it will force to recalculate)

May the swap size impact the regular node operations? Would it make sense to increase it on the relay nodes as well?

Nope, for a relay 16G is more than enough

U must use swap on BP because the node already using ~9G of RAM plus CNCLI another ~9G of RAM when it run (u can check with top or htop while CNCLI is running)

You are welcome, :beers:

7

Success, no restart. Thanks again!

Is the book “Cardano Pool for Dummies” coming soon?

1 Like