Summary: | nilfs2 is mounted without cleanderd started | ||
---|---|---|---|
Product: | systemd | Reporter: | Peter <pva> |
Component: | general | Assignee: | systemd-bugs |
Status: | RESOLVED INVALID | QA Contact: | systemd-bugs |
Severity: | normal | ||
Priority: | medium | ||
Version: | unspecified | ||
Hardware: | Other | ||
OS: | All | ||
Whiteboard: | |||
i915 platform: | i915 features: |
Description
Peter
2013-10-12 17:01:35 UTC
We have an analogous sitiuation with nfs, which also requires a bunch of services to run. With nilfs, I guess that we could a) modify fstab-generator to add a dependency on nilfs-cleanerd@<device>.service. If the daemon is non-essential, Wants=nilfs-cleanerd@<device>.service is probably the right dependency. b) modify mount.nilfs2 to run the equivalent of 'systemctl start nilfs-cleanerd@<device>.service', probably directly through d-bus. It shouldn't be run directly from mount.nilfs2, because we want systemd to supervise it. I guess that the first option would be easier to implement, and resembles what we do with nfs and lvm. What do you think about option a? This is similar I figure to file systems using FUSE which also require a per-mount userspace process to stay around while the mount is active. This should more or less work, .mount units will keep track of processes forked off. So, if mount.nilfs2 works correctly this should also just work with systemd. Except of course that daemon is supposed to be a singleton that is shared by all such mounts. (In reply to comment #2) > This should more or less work, .mount units will keep track of processes > forked off. So, if mount.nilfs2 works correctly this should also just work > with systemd. Except of course that daemon is supposed to be a singleton > that is shared by all such mounts. Lennart you are right. My nilfs package was built without libmount support and that's why it failed. After I rebuilt it with such support everything works. Sorry for the noise. |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.