# grep NGINX_PIDFILE /etc/conf.d/nginx NGINX_PIDFILE="/run/nginx/nginx.pid" # /etc/init.d/nginx --verbose start * Caching service dependencies ... [ ok ] * Executing: /lib/rc/sh/openrc-run.sh /lib/rc/sh/openrc-run.sh /etc/init.d/nginx start * Checking nginx' configuration ... [ ok ] * Starting nginx ... * start-stop-daemon: fopen `/run/nginx/nginx.pid': No such file or directory * Detaching to start `/usr/sbin/nginx' ... * start-stop-daemon: fopen `/run/nginx/nginx.pid': No such file or directory * start-stop-daemon: did not create a valid pid in `/run/nginx/nginx.pid' * Failed to start nginx [ !! ] * ERROR: nginx failed to start This makes nginx problematic to run completely rootless. Normally this supposed to work (for instance it works in Alpine Linux): echo 'command_user="nginx"' >> /etc/conf.d/nginx setcap CAP_NET_BIND_SERVICE=+eip /usr/sbin/nginx # NGINX_PIDFILE is already set to location writable by user nginx in Alpine Linux /etc/init.d/nginx start ps uax | grep nginx | grep -v grep # none of the processes are running by root It seems to be hardcoded at "configure" step with "--pid-path" in both distributions. To mitigate this issue it seems we can make "/run/nginx/nginx.pid" a default pid file (where "nginx:nginx" owns "/run/nginx", same as in Alpine). Reproducible: Always
Created attachment 902813 [details] emerge --info
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=108a642ef4da9fa6ae4f0d33580a05cea3f70678 commit 108a642ef4da9fa6ae4f0d33580a05cea3f70678 Author: Alexander Lopatin <alopatindev@codonaft.com> AuthorDate: 2024-09-13 17:01:51 +0000 Commit: Joonas Niilola <juippis@gentoo.org> CommitDate: 2024-10-17 07:35:53 +0000 www-servers/nginx: fix directory path to pid file Closes: https://bugs.gentoo.org/939560 Signed-off-by: Alexander Lopatin <alopatindev@codonaft.com> Closes: https://github.com/gentoo/gentoo/pull/38593 Signed-off-by: Joonas Niilola <juippis@gentoo.org> www-servers/nginx/files/nginx.confd | 5 ++++- www-servers/nginx/files/nginx.initd-r4 | 4 +++- www-servers/nginx/files/nginx.logrotate-r1 | 2 +- www-servers/nginx/files/nginx.service-r1 | 2 +- www-servers/nginx/nginx-1.26.2-r3.ebuild | 2 +- www-servers/nginx/nginx-1.27.2.ebuild | 2 +- 6 files changed, 11 insertions(+), 6 deletions(-)
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=6f136d56a588639697e42a27391f83483d333e0e commit 6f136d56a588639697e42a27391f83483d333e0e Author: Louis Sautier <sbraz@gentoo.org> AuthorDate: 2024-10-19 14:36:26 +0000 Commit: Louis Sautier <sbraz@gentoo.org> CommitDate: 2024-10-19 14:44:24 +0000 www-servers/nginx: fix systemd unit to create /run/nginx/ (#941784) 108a642ef4da broke the systemd service, by moving the PID file to /run/nginx/ without creating the directory, resulting in 'open() "/run/nginx/nginx.pid" failed (2: No such file or directory)' Bug: https://bugs.gentoo.org/939560 Closes: https://bugs.gentoo.org/941784 Signed-off-by: Louis Sautier <sbraz@gentoo.org> www-servers/nginx/files/nginx.service-r1 | 1 + www-servers/nginx/{nginx-1.26.2-r3.ebuild => nginx-1.26.2-r4.ebuild} | 0 www-servers/nginx/{nginx-1.27.2.ebuild => nginx-1.27.2-r1.ebuild} | 0 3 files changed, 1 insertion(+)
i did not had this issue until i just updated to -r4! with -r4 i get this error: did not create a valid pid in `/run/nginx/nginx.pid' and i landed here, so this patch does not fix an error, it actually produces an error! (i created an overlay, reverted back to -r3 and the error is gone!) i have this line in etc/nginx.conf pid /run/nginx.pid;
please reopen - needs fix!
/run/nginx.pid is not a valid path anymore, it's a hardcoded value that has been changed to /run/nginx/nginx.pid due to original issue https://bugs.gentoo.org/939560#c0 Does it work for you if you change /run/nginx.pid to /run/nginx/nginx.pid in nginx.conf?
Just installed -r5 and changed nginx.conf to "pid /run/nginx/nginx.pid;" Now start/stop/restart works fine - Thanks!
