Multiple mount on the same mount point

[cat0dog]

What happened:
After setting the juicefs file system to mount automatically on boot, every time you execute mount -a manually, there will be one more juicefs mount processes, corresponding to X umount

However, when NFS automounts are recorded in /etc/fstab, no matter how many times mount -a is executed, there is only 1 NFS mounted process, corresponding to 1 umount

What you expected to happen:
mount -a will not mount an existing juicefs file system repeatedly

How to reproduce it (as minimally and precisely as possible):
Set JuiceFS to mount automatically on boot, then repeat mount -a several times

Anything else we need to know?

Environment:

• JuiceFS version (use juicefs --version) or Hadoop Java SDK version:

juicefs version 1.0.0+2022-08-08.cf0c269

• OS (e.g cat /etc/os-release):

PRETTY_NAME="Ubuntu 22.04 LTS"
NAME="Ubuntu"
VERSION_ID="22.04"

• Kernel (e.g. uname -a):

Linux node75 5.15.0-33-generic #34-Ubuntu SMP Wed May 18 13:34:26 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux

[cat0dog]

And after executing 1 umount, it will cause the client monitoring data of this node to be lost

[SandyXSD]

mount -a will compare the entry in fstab and mountinfo (/proc/self/mountinfo or /run/mount/utab), and ignore already mounted file systems. JuiceFS use META-URL as the first argument in fstab, and volume name as the FsName of FUSE, while both of them are considered to be SRC in mount commands. Since they are usually different, so whenever mount -a is executed it will always call juicefs mount.
There is no plan to change this behavior of JuiceFS (META-URL is not a good option to display as FsName because it's usually pretty long), so for now we may just leave this issue as it is.

[SandyXSD]

If this issue is really a concern to you, you can try recording the META-URL and setting it as the FsName here: https://github.com/juicedata/juicefs/blob/release-1.0/pkg/fuse/fuse.go#L434