Hacktoberfest 2026：メンテナが10月に向けて印を付けた、オープンで初心者向けの issue。 Hacktoberfest の issue を見る

[BUG] - Cardano-node 11.1.3 crashes due to increasing memory consumption

オープン
  • 何か問題がありますか？
#6,734 コメント 2 件 リアクション 0 件 担当者 0 名 GitHub で見る

メンテナーはふだん 1 日以内に返信

まだ誰も着手していません。

評価

難易度
4/5
見積もり時間
3〜5日
初心者へのやさしさ
44/100
issue の種類
バグ
明瞭さ
おおむね明確
活発さ
活発
技術スタック
docker, docker-compose, grafana, haskell
領域
backend, performance

調査の方向性

Start by comparing cardano-node 11.1.3 with 11.1.2 under the reported Docker setup, using docker stats to track memory and the supplied RTS options. The issue does not identify a source file or failing test, so tracing the memory growth will require familiarity with the node. Done means reproducing and locating the regression and confirming memory stays within the reported 16–24 GB range.

索引モデルが issue の本文から書いたものです。

説明

needs triage

External

Area
Other cardano-node 11.1.3

Summary
cardano-node 11.1.3 consumes over 28 Gb memory within 6 hours , exhausts resources and crashes (restarts my container). This is on a bare metal machine with 64 Gb memory running two nodes. On the same machine running one node this problem doesn't occur. But it should never consume more than 28 Gb memory in the first place and the memory consumption crawls up over time.

Running this in my own docker environment on basis of Debian-Slim. Switching back to cardano-node 11.1.2 shows that the node stays within the 24Gb memory with the same docker environment, just change of the binary.

So there is definitely a problem with Cardano-node 11.1.3 binary.

Steps to reproduce
I think it should be easy to reproduce on your end by monitoring the memory or by installing two nodes on a machine with 64 Gb and restricting the container with the following settings in .env NODE_MEM_LIMIT=32g
NODE_MEM_RESERVATION=28g and under services relay in docker-compose.yaml
services:
relay:
mem_limit: ${NODE_MEM_LIMIT:-0}
mem_reservation: ${NODE_MEM_RESERVATION:-0}

I increased the values of NODE-MEM-LIMIT and NODE_MEM_RESERVATION because of this bug, but this didn't help still crashing the node after exhausting the memory.

cardano-node started with the following rts_opts="${RTS_OPTS:--N4 -xn -T -A128m -n1m -I0.3 -Iw3600 -F1.5 -H8000m}"

I switched back to cardano-node 11.1.2 with the exact same setup and settings and it stays within the boundaries of 24 Gb memory consumption, the memory consumption doesn't crawl up like with cardano-node 11.1.3

Checked memory consumption with: docker stats --no-stream
Monitoring the memory consumption of my nodes with Grafana and get an alert when the node exceeds 24 Gb or crashes.

Expected behavior
The cardano-node stays within the range of 16 - 24 Gb memory consumption, one shouldn't be forced to run with the disk option when they have memory enough to run it comfortably in memory.

Screenshots and attachments
Can provide if required, but I think on the above you should have enough to reproduce the behavior.

Additional context
It seems the node was not tested well and rushed out to fix an issue, or assuming everyone runs on local storage instead of memory.

主要言語
Haskell
スター
3.2k
フォーク
762
平均マージ
2日 9時間
マージ済み PR（30日）
19

環境構築

はじめの一歩

  1. issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
  2. 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
  3. リポジトリをフォークし、ブランチを切って変更します。
  4. issue 番号を参照したプルリクエストを送ります。

IntersectMBO/cardano-node のほかの issue

IntersectMBO/cardano-node の issue をすべて見る

似ている issue

Haskell の issue をもっと見る

新しい issue をメールで受け取る

初心者向けの GitHub issue を短くまとめたダイジェスト。