Hi,
Some of our embedded units enter a strange state on booting after a power-down:
We have noticed this on other units that have an overlay filessystem - and sometimes it can take a long time before the unit properly boots again.
[ 9.448158] init: - preinit -
[ 10.659347] nicvf 0000:05:00.1 eth0: Link is Up 1000 Mbps Full duplex
[ 14.795447] mount_root: loading kmods from internal overlay
[ 14.837603] kmodloader: loading kernel modules from //etc/modules-boot.d/*
[ 14.847751] kmodloader: done loading kernel modules from //etc/modules-boot.d/*
[ 15.137347] block: attempting to load /etc/config/fstab
[ 15.147132] block: unable to load configuration (fstab: Entry not found)
[ 15.153905] block: no usable configuration
[ 15.380795] F2FS-fs (loop0): recover_inode: ino = baf, name = syslog, inline = 1
[ 15.388293] F2FS-fs (loop0): recover_data: ino = baf (i_size: recover) recovered = 0, err = 0
[ 15.396827] F2FS-fs (loop0): recover_inode: ino = baf, name = syslog, inline = 1
[ 15.404279] F2FS-fs (loop0): recover_data: ino = baf (i_size: recover) recovered = 0, err = 0
[ 15.412810] F2FS-fs (loop0): recover_inode: ino = baf, name = syslog, inline = 1
[ 15.420260] F2FS-fs (loop0): recover_data: ino = baf (i_size: recover) recovered = 0, err = 0
[ 15.428789] F2FS-fs (loop0): recover_inode: ino = baf, name = syslog, inline = 1
[ 15.436240] F2FS-fs (loop0): recover_data: ino = baf (i_size: recover) recovered = 0, err = 0
[ 15.444768] F2FS-fs (loop0): recover_inode: ino = baf, name = syslog, inline = 1
[ 15.452227] F2FS-fs (loop0): recover_data: ino = baf (i_size: recover) recovered = 0, err = 0
[ 15.460757] F2FS-fs (loop0): recover_inode: ino = baf, name = syslog, inline = 1
[ 15.468209] F2FS-fs (loop0): recover_data: ino = baf (i_size: recover) recovered = 0, err = 0
[ 15.476746] F2FS-fs (loop0): recover_inode: ino = baf, name = syslog, inline = 1
[ 15.484213] F2FS-fs (loop0): recover_data: ino = baf (i_size: recover) recovered = 1, err = 0
[ 15.492744] F2FS-fs (loop0): recover_inode: ino = baf, name = syslog, inline = 1
[ 15.500197] F2FS-fs (loop0): recover_data: ino = baf (i_size: recover) recovered = 0, err = 0
Once complete it seems happy enough again:
[ 16.164241] F2FS-fs (loop0): recover_inode: ino = baf, name = syslog, inline = 1
[ 16.171702] F2FS-fs (loop0): recover_data: ino = baf (i_size: recover) recovered = 1, err = 0
[ 16.180240] F2FS-fs (loop0): recover_inode: ino = baf, name = syslog, inline = 1
[ 16.187695] F2FS-fs (loop0): recover_data: ino = baf (i_size: recover) recovered = 0, err = 0
[ 17.213819] F2FS-fs (loop0): checkpoint: version = 7f7acc28
[ 17.219537] F2FS-fs (loop0): Mounted with checkpoint version = 7f7acc28
[ 17.351770] block: attempting to load /etc/config/fstab
[ 17.357116] block: unable to load configuration (fstab: Entry not found)
[ 17.363886] block: no usable configuration
[ 17.369894] mount_root: switching to f2fs overlay
[ 17.410907] overlayfs: "xino" feature enabled using 32 upper inode bits.
[ 17.508656] urandom-seed: Seeding with /etc/urandom.seed
[ 17.813323] procd: - early -
[ 17.816305] procd: - watchdog -
[ 18.381135] procd: - watchdog -
[ 18.384626] procd: - ubus -
[ 18.454204] procd: - init -
[ 18.819612] urngd: v1.0.2 started.
[ 18.919262] kmodloader: loading kernel modules from /etc/modules.d/*
# df -h
Filesystem Size Used Available Use% Mounted on
/dev/root 48.0M 48.0M 0 100% /rom
tmpfs 495.3M 14.9M 480.3M 3% /tmp
/dev/loop0 918.0M 167.9M 750.1M 18% /overlay
overlayfs:/overlay 918.0M 167.9M 750.1M 18% /
tmpfs 512.0K 0 512.0K 0% /dev
/dev/mmcblk1p4 28.7G 44.0M 27.1G 0% /media
#
Does anyone know how we can prevent this behaviour?
If we shrunk the overlay partition would it at least reduce the time it takes?
I know the above example only took 2 seconds but sometimes it can take 30 minutes.
Not sure what "name = syslog" is about either because there's not one on OpenWRT.
Thanks in advance for any advice.
Stuart