I'm using nilfs2 for my root file system. It seems as if mount.nilfs2 manages to start a cleanerd, because "ps" shows this:
[root@robot1 /]# ps alx | grep ni[l] 1 0 222 2 20 0 0 0 nilfs_ S ? 0:03 [segctord] 4 0 223 1 20 0 14692 864 wq_sle Ss ? 0:00 /sbin/nilfs_cleanerd -n /dev/sda1 / [root@robot1 /]# nilfs-clean /dev/sda1 No cleaner found on /dev/sda1. [root@robot1 /]#
However, as you can see, trying to clean the /dev/sda1 volume doesn't work.
If, however, I manually start nilfs_cleanerd, then I can clean the volume:
[root@robot1 /]# /sbin/nilfs_cleanerd [root@robot1 /]# nilfs-clean /dev/sda1 [root@robot1 /]#
I'd like to understand why the nilfs-clean utility doesn't think that the running, auto-launched nilfs_cleanerd is suitable for cleaning the device.
I have another post in the server config forum about systemctl for starting a nilfs_cleanerd; that's not what this question is about. This question is specifically: Why doesn't nilfs-clean find the running, seemingly auto-started, nilfs_cleanerd instance?
Nobody knows about nilfs2 or how to diagnose this problem?