You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Summary
A node (and dbsync and cardano-wallet) eats more and more memory day per day.
Count of tokens is increasing, so ledgerState is growing too.
After Alonzo release it will become more worse, I guarantee.
It's not a perfect idea to keep in memory the thing which grows every day with no limits. Its a long-distance memory leak as a part of software design.
Its not okay when you have to buy enterprise-level servers to run couple cardano applicarions.
Its not okay when you cannot plan load level.
Its not okay when you have to pay for hardware ten times more than just a year go.
Its not okay when you think how to run apps on different servers with shared by local netfs socket to buy more time between hardware upgrades.
Its not okay when you know that everything written above its not enough, because RAM utilization grows every day.
RAM is not unlimited. And not cheap when we talk about VPS.
You can not ignore this.
Steps to reproduce
Steps to reproduce the behavior:
Run a node over the mainnet
See the "7.5GB" in the htop
Run dbsync
See the "16GB" in the htop
Try to not cry
Expected behavior
Run a node over the mainnet
See the "300M" in the htop
Be happy, enjoy your life
The text was updated successfully, but these errors were encountered:
Summary
A node (and dbsync and cardano-wallet) eats more and more memory day per day.
Count of tokens is increasing, so ledgerState is growing too.
After Alonzo release it will become more worse, I guarantee.
It's not a perfect idea to keep in memory the thing which grows every day with no limits. Its a long-distance memory leak as a part of software design.
Its not okay when you have to buy enterprise-level servers to run couple cardano applicarions.
Its not okay when you cannot plan load level.
Its not okay when you have to pay for hardware ten times more than just a year go.
Its not okay when you think how to run apps on different servers with shared by local netfs socket to buy more time between hardware upgrades.
Its not okay when you know that everything written above its not enough, because RAM utilization grows every day.
RAM is not unlimited. And not cheap when we talk about VPS.
You can not ignore this.
Steps to reproduce
Steps to reproduce the behavior:
Expected behavior
The text was updated successfully, but these errors were encountered: