apache / apache/datafusion

Memory held on in `datafusion-cli` (reported memory size increases over time)

Open
#15,939 5 comments 0 reactions 0 assignees View on GitHub
bug
Dominant language
Rust
Stars
9.3k
Forks
2.4k
Avg merge
3d 7h
Merged PRs (30d)
344

Description

### Describe the bug

When running several memory consuming queries in datafusion-cli, the system memory won't be released after the query ends, the memory usage from different queries are accumulated, until `datafusion-cli` is exited.

### To Reproduce

1. Compile and run datafusion-cli (commit 8c42a65a4)
2. I'm using MacOS's `Application Monitor` GUI to check memory usage of `datafusion-cli` process
3. Execute several memory-intensive queries, and memory usage accumulates over time.
```
select *
from generate_series(1,100000000) as t1(v1)
order by v1;

-- datafusion-cli process: 1.5G memory

select *
from generate_series(1,200000000) as t1(v1)
order by v1;

-- datafusion-cli process: 3.2G memory

select 1;

-- Still 3.2G
```

### Expected behavior

_No response_

### Additional context

_No response_

Contributor guide

Open the contributing guide

Research direction

Compile and run datafusion-cli at commit 8c42a65a4, then reproduce the two memory-intensive generate_series queries followed by SELECT 1 while monitoring the process in macOS Application Monitor. Trace the datafusion-cli query lifecycle to determine why memory remains allocated after each query; done means repeated queries no longer cause memory usage to accumulate after completion.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust
Domain
cli, performance
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
42/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.