fsx test failed on bus error with meta engine tikv

[zhoucheng361]

What happened:

https://github.com/juicedata/juicefs/actions/runs/9058635700/job/24884731815

Run:
  if [[ "pull_request" == "schedule" || "pull_request" == "workflow_dispatch"  ]] ; then
    duration=900
  else
    duration=300
  fi
  sudo touch /jfs/fsx.out
  sudo rm -f /tmp/fsx.out
  sudo ln -s /jfs/fsx.out /tmp/fsx.out
  sudo secfs.test/tools/bin/fsx -d $duration -p 10000 -F 10000000 /tmp/fsx.out
  # sudo secfs.test/tools/bin/fsx -d 180 -p 10000 -F 100000 /jfs/fsx.out
  shell: /usr/bin/bash -e {0}
Running for 300 seconds
Using file /tmp/fsx.out
command line: secfs.test/tools/bin/fsx -d 300 -p 10000 -F 10000000 /tmp/fsx.out
Seed set to 1715601[2](https://github.com/juicedata/juicefs/actions/runs/9058635700/job/24884731815#step:11:2)91
truncating to largest ever: 0x2aa19c
truncating to largest ever: 0x83579c
truncating to largest ever: 0x86cea8
truncating to largest ever: 0x878a70
truncating to largest ever: 0x9250d8
truncating to largest ever: 0x9[3](https://github.com/juicedata/juicefs/actions/runs/9058635700/job/24884731815#step:11:3)ac5c
truncating to largest ever: 0x950c1c
truncating to largest ever: 0x97898[4](https://github.com/juicedata/juicefs/actions/runs/9058635700/job/24884731815#step:11:4)
truncating to largest ever: 0x989680
/home/runner/work/_temp/cebaf[5](https://github.com/juicedata/juicefs/actions/runs/9058635700/job/24884731815#step:11:5)03-2b15-489f-b24d-3b8a[6](https://github.com/juicedata/juicefs/actions/runs/9058635700/job/24884731815#step:11:6)c4c[7](https://github.com/juicedata/juicefs/actions/runs/9058635700/job/24884731815#step:11:7)f10.sh: line 9: 17105 Bus error               sudo secfs.test/tools/bin/fsx -d $duration -p [10](https://github.com/juicedata/juicefs/actions/runs/9058635700/job/24884731815#step:11:10)000 -F 10000000 /tmp/fsx.out

What you expected to happen:

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

Anything else we need to know?

Environment:

• JuiceFS version (use juicefs --version) or Hadoop Java SDK version:
• Cloud provider or hardware configuration running JuiceFS:
• OS (e.g cat /etc/os-release):
• Kernel (e.g. uname -a):
• Object storage (cloud provider and region, or self maintained):
• Metadata engine info (version, cloud provider managed or self maintained):
• Network connectivity (JuiceFS to metadata engine, JuiceFS to object storage):
• Others:

[SandyXSD]

Can't reproduce on the upterm after the CI failed.

[zhoucheng361]

fix by #4847