[arch-general] systemd claims /run out of space
For some reason on my home server, systemd is often telling me I'm out of space, but I can't find a problem. Just now, I stopped httpd, and when I try to start it, again, systemd says I'm out of space, the status reports: systemd[1]: Failed to set a watch for httpd.service's PID file /run/httpd/httpd.pid: No space left on device. However, df shows: Filesystem Size Used Avail Use% Mounted on /dev/sda4 107G 33G 73G 31% / dev 7.9G 0 7.9G 0% /dev run 7.9G 572K 7.9G 1% /run tmpfs 7.9G 0 7.9G 0% /dev/shm tmpfs 7.9G 0 7.9G 0% /sys/fs/cgroup tmpfs 7.9G 1.4M 7.9G 1% /tmp di reports: run /run 7.8G 0.0G 7.8G 0% tmpfs Does anyone know where I can look to see why systemd returns this? Thanks, =-Jameson
Am 13.01.2014 17:54, schrieb Jameson:
For some reason on my home server, systemd is often telling me I'm out of space, but I can't find a problem. Just now, I stopped httpd, and when I try to start it, again, systemd says I'm out of space, the status reports: systemd[1]: Failed to set a watch for httpd.service's PID file /run/httpd/httpd.pid: No space left on device.
That error message is confusing. The inotify_add_watch() system call returns ENOSPC. Usually, ENOSPC means "No space left on device", and the strerror() library call translates it to that. However, in the context of inotify_add_watch(), it means: ENOSPC The user limit on the total number of inotify watches was reached or the kernel failed to allocate a needed resource. So, it seems that you allocated many many inotify watches and need to increase the limit.
On Mon, Jan 13, 2014 at 2:10 PM, Thomas Bächler <thomas@archlinux.org> wrote:
Am 13.01.2014 17:54, schrieb Jameson:
For some reason on my home server, systemd is often telling me I'm out of space, but I can't find a problem. Just now, I stopped httpd, and when I try to start it, again, systemd says I'm out of space, the status reports: systemd[1]: Failed to set a watch for httpd.service's PID file /run/httpd/httpd.pid: No space left on device.
So, it seems that you allocated many many inotify watches and need to increase the limit.
Thanks, man. You nailed it. Do you think I should file a bug report somewhere to see if I can have the devs work out a better error message? Is it a kernel bug, a bug with the strerror library, or a systemd issue?
Am 14.01.2014 00:35, schrieb Jameson:
Thanks, man. You nailed it. Do you think I should file a bug report somewhere to see if I can have the devs work out a better error message? Is it a kernel bug, a bug with the strerror library, or a systemd issue?
It is in the function service_watch_pid_file here: http://cgit.freedesktop.org/systemd/systemd/tree/src/core/service.c#n2783 I guess systemd shouldn't blindly print the strerror message and provide a better one.
participants (2)
-
Jameson
-
Thomas Bächler