Jekyll, Bundler and the “No space left on device - Failed to watch” error
Today I started my blog locally, to inspect some posts, and the system refused to start.What did I change since last time? Nothing!
Apparently, however, I was running the system with an higher load than usual, having several other applications already opened. Let’s face the error:
% bundle exec jekyll serve
...
Generating...
done in 12.299 seconds.
jekyll 3.9.2 | Error: Unable to monitor directories for changes because iNotify max watches exceeded. See https://github.com/guard/listen/blob/master/README.md#increasing-the-amount-of-inotify-watchers .
/var/lib/gems/3.0.0/gems/listen-3.7.1/lib/listen/adapter/linux.rb:32:in `rescue in _configure': Unable to monitor directories for changes because iNotify max watches exceeded. See https://github.com/guard/listen/blob/master/README.md#increasing-the-amount-of-inotify-watchers . (Listen::Error::INotifyMaxWatchesExceeded)
...
/var/lib/gems/3.0.0/gems/rb-inotify-0.10.1/lib/rb-inotify/watcher.rb:74:in `initialize': No space left on device - Failed to watch "/fluca1978.github.io/_site/2019/12/10": The user limit on the total number of inotify watches was reached or the kernel failed to allocate a needed resource. (Errno::ENOSPC)
...
What the hell is happening?
The first thing that caught my attention was that
No space left on device: did I run out of disk space? Of course I was not.
Then I have a second look at the error, in particular at the root cause, and followed the suggested link to solve the issue.
The sysctl mentioned there,
fs.inotify.max_user_watches determines the maximum value of files that the inotify subsystem can handle for changes on the filesystem tree.
Once this value has been raised, the Jekyll local site started up smoothly.