You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Does a report exist for this bug with the project's home (upstream) and/or another distro?
No response
Expected behaviour
If the directories /var/log/socklog and /var/log/socklog/* do not exist, they should be automatically recreated by the shell script /etc/sv/socklog-unix/run before starting the socklog executable.
Actual behaviour
If the directories /var/log/socklog and /var/log/socklog/* do not exist when the socklog-unix service starts, it leads to unexpected behavior where all login attempts, including sudo and su, stall indefinitely. This effectively locks users out of the system, rendering it inaccessible for any user interaction. The socklog-unix service seems to depend on the existence of its log directories for normal operation. Without them, processes like login, which need to log activities, hang waiting for the log to be written. This isn't just a minor inconvenience but a critical system access issue, as no user can log in to fix the problem without rebooting into single-user mode or using a live USB to manually disable the service.
Additionally, 'xbps-reconfigure -f socklog-void' fails to recreate the subdirectories of /var/log/socklog. The package must be removed and reinstalled. The socklog-conf executable also seems to fail.
This issue prohibits the use of socklog on systems that, for example, mount a tmpfs on /var/log, like home PCs.
It is xbps-install -f socklog-void that you should run, and it should recreate the directory.
I agree we could add a check somewhere in the service for the existence of dir /var/log/socklog.
Is this a new report?
Yes
System Info
Void 6.12.10_1 x86_64-musl
Package(s) Affected
socklog-void-20200115_2
Does a report exist for this bug with the project's home (upstream) and/or another distro?
No response
Expected behaviour
If the directories /var/log/socklog and /var/log/socklog/* do not exist, they should be automatically recreated by the shell script /etc/sv/socklog-unix/run before starting the socklog executable.
Actual behaviour
If the directories /var/log/socklog and /var/log/socklog/* do not exist when the socklog-unix service starts, it leads to unexpected behavior where all login attempts, including sudo and su, stall indefinitely. This effectively locks users out of the system, rendering it inaccessible for any user interaction. The socklog-unix service seems to depend on the existence of its log directories for normal operation. Without them, processes like login, which need to log activities, hang waiting for the log to be written. This isn't just a minor inconvenience but a critical system access issue, as no user can log in to fix the problem without rebooting into single-user mode or using a live USB to manually disable the service.
Additionally, 'xbps-reconfigure -f socklog-void' fails to recreate the subdirectories of /var/log/socklog. The package must be removed and reinstalled. The socklog-conf executable also seems to fail.
This issue prohibits the use of socklog on systems that, for example, mount a tmpfs on /var/log, like home PCs.
Steps to reproduce
The text was updated successfully, but these errors were encountered: