[MacOS] "device not configured"

[BitPhinix]

What happened:

Running juicefs bench on macos with a default s3 + redis setup causes

2024/08/01 23:58:22.065604 juicefs[1221] : Failed to open file /Volumes/datasets/juicefs_benchmark_1722581901987909000/bigfile.0.0: open /Volumes/datasets/juicefs_benchmark_1722581901987909000/bigfile.0.0: device not configured [bench.go:121]

How to reproduce it (as minimally and precisely as possible):

Setup a new juicefs volume using s3 + redis

juicefs format \
          --storage s3 \
          --trash-days 7 \
          --bucket xxxx \
          --access-key xxxx \
          --secret-key xxx \
          "redis://xxx:xxxx@redis-10710.c113.us-west-2-mz.ec2.redns.redis-cloud.com:10710" \
          datasets

Mount it using

sudo juicefs mount --backup-meta 0 -d "redis://xxxx:xxxxxx@redis-10710.c113.us-west-2-mz.ec2.redns.redis-cloud.com:10710" /Volumes/datasets

Run juicefs bench /Volumes/datasets/

Anything else we need to know?

Environment:

• JuiceFS version (use juicefs --version) or Hadoop Java SDK version: juicefs version 1.2.0+2024-06-18.873c47b922ba
• Cloud provider or hardware configuration running JuiceFS: local laptop
• OS (e.g cat /etc/os-release): macos 14.5 (23F79)
• Kernel (e.g. uname -a): Darwin Erics-MacBook-Pro.local 23.5.0 Darwin Kernel Version 23.5.0: Wed May 1 20:12:58 PDT 2024; root:xnu-10063.121.3~5/RELEASE_ARM64_T6000 arm64
• Object storage (cloud provider and region, or self maintained): aws s2 us-west-2
• Metadata engine info (version, cloud provider managed or self maintained): redis labs managed redis
• Network connectivity (JuiceFS to metadata engine, JuiceFS to object storage): JuiceFS to object storage
• Others: Macfuse version 4.8.0

[BitPhinix]

Logs around the time the error happens:

2024/08/02 00:31:49.640262 juicefs[1934] <WARNING>: slow request: PUT chunks/0/12/12329_1_4194304 (req_id: "FDK0NZATAZC71JEJ", err: <nil>, cost: 29.636684583s) [cached_store.go:667]
2024/08/02 00:31:50.558047 juicefs[1934] <DEBUG>: PUT chunks/0/12/12340_4_65536 (req_id: "S7ZK131GEVEB14AG", err: <nil>, cost: 917.302708ms) [cached_store.go:669]
2024/08/02 00:31:51.541151 juicefs[1925] <DEBUG>: send FUSE fd: 14 [passfd.go:110]
2024/08/02 00:31:51.543918 juicefs[1925] <DEBUG>: msg:  fds: [] [passfd.go:139]
2024/08/02 00:31:51.550047 juicefs[1934] <WARNING>: flush 3074 interrupted after 120014983000 [writer.go:371]
2024/08/02 00:31:51.550281 juicefs[1934] <INFO>: slow operation: flush (3074,14,0000000000000000) - interrupted system call <120.015203> [accesslog.go:83]
2024-08-02 00:31:51.550797 I | writer: Write/Writev failed, err: 9=bad file descriptor. opcode: FLUSH

[zhijian-pro]

@BitPhinix Does this happen once in a while or every time?

[jiefenghuang]

The issue has been reproduced, operation timed out, causing FUSE to exit, and JuiceFS forgot to unmount.