Linux – how to get the number of inotify watches in use

inotifylinux

I use inotifywait for event trigger which put file.

When many files are watched by inotifywait, when max_user_watches is exceeded, the following error occurs.

Terminating since out of inotify watches.#012Consider increasing /proc/sys/fs/inotify/max_user_watches

It is necessary to tune /proc/sys/fs/inotify/max_user_watches, but is it possible to check the current file watch number?

Is there a way to check like file-nr in file descriptor?

Best Answer

I cobbled together this little script based on @mosvy's answer. Since the initial conception, it has since seen quite a few improvements (stability on older systems, total count, speed). On most normal machines, running it should take a less than 100 ms. Even though it was much faster than the other alternatives when I made it initially, Simon Matter added some speed enhancements for heavily loaded Big Iron Linux (hundreds of cores) that sped it up immensely, taking it down from ten minutes (!) to 15 seconds.

   INOTIFY
   WATCHER
    COUNT     PID USER     COMMAND
--------------------------------------
    3044   3933 myuser node /usr/local/bin/tsserver
    2965   3941 myuser /usr/local/bin/node /home/myuser/.config/coc/extensions/node_modules/coc-tsserver/bin/tsserverForkStart /hom
     979   3954 myuser /usr/local/bin/node /home/myuser/.config/coc/extensions/node_modules/coc-tsserver/node_modules/typescript/li
       1   3899 myuser /usr/local/bin/node --no-warnings /home/myuser/dev/dotfiles/common-setup/vim/dotvim/plugged/coc.nvim/build/i

    6989  WATCHERS TOTAL COUNT

Related Solutions

Efficiency of lots of inotify watches or stat calls

When you read() an inotify fd, the name field of the returned struct tells you which file was modified relative to the directory being watched, so you shouldn't have to stat every file in a directory after the event.

See http://linux.die.net/man/7/inotify

Specifically:

struct inotify_event {
    int      wd;       /* Watch descriptor */
    uint32_t mask;     /* Mask of events */
    uint32_t cookie;   /* Unique cookie associating related
                          events (for rename(2)) */
    uint32_t len;      /* Size of 'name' field */
    char     name[];   /* Optional null-terminated name */
};
The name field is only present when an event is returned for a file inside a watched directory; it identifies the file pathname relative to the watched directory. This pathname is null-terminated, and may include further null bytes to align subsequent reads to a suitable address boundary.

How does inotify work

Inotify is an internal kernel facility. There is no “inotify file”. There are dedicated system calls inotify_init, inotify_add_watch and inotify_rm_watch that allow processes to register themselves to be notified when certain filesystem events happen. When the event happens, the process receives a description of the event through the file descriptor returned by inotify_init.

The OS isn't “told” that a file has been changed: it knows, because it's doing the changing. It's the application that's told that a file has been changed instead of having to go looking.

The program inotifywait provides a simple way to use inotify from the command line.

Best Answer

Related Solutions

Efficiency of lots of inotify watches or stat calls

How does inotify work

Related Question