Clean install of odroid-c2 image is slow to boot [SOLVED]

[graysky]

Seems as though I have two 10+ second delays in my init:
$this->bbcode_second_pass_code('', '
% systemd-analyze critical-chain
The time after the unit is active or started is printed after the "@" character.
The time the unit takes to start is printed after the "+" character.

graphical.target @11.482s
└─multi-user.target @11.478s
└─systemd-resolved.service @11.416s +56ms
└─network.target @11.382s
└─systemd-networkd.service @11.217s +130ms
└─dbus.service @11.086s
└─basic.target @11.049s
└─sockets.target @11.039s
└─dbus.socket @11.034s
└─sysinit.target @11.029s
└─systemd-timesyncd.service @10.745s +278ms
└─systemd-tmpfiles-setup.service @10.718s +15ms
└─local-fs.target @10.711s
└─mnt-data.mount @10.684s +21ms
└─systemd-fsck@dev-mmcblk0p3.service @1.409s +9.255s
└─dev-mmcblk0p3.device @1.401s

')

The culprits are: systemd-resolved, systemd-networkd, and strangely enough, a data partition mount service. When I inspect the `journalctl -u xxx` output for each of these, I see nothing out of the ordinary. I have more or less the same setup on a RPi2 and it boots up insanely fast by comparison. Do others running this hardware see the same delays in the same places? I should note that I am using a very fast microSD card so I doubt the problem lies with that. Thanks all!

[cmsigler]

Hi,

My Odroid-C2 seems to be booted up in just over 1.5 secs:

$this->bbcode_second_pass_code('', '
$ systemd-analyze critical-chain
The time after the unit is active or started is printed after the "@" character.
The time the unit takes to start is printed after the "+" character.

graphical.target @1.541s
`-multi-user.target @1.537s
`-systemd-resolved.service @1.484s +46ms
`-network.target @1.423s
`-systemd-networkd.service @1.298s +83ms
`-dbus.service @1.244s
`-basic.target @1.238s
`-sockets.target @1.233s
`-dbus.socket @1.228s
`-sysinit.target @1.222s
`-systemd-timesyncd.service @978ms +237ms
`-systemd-tmpfiles-setup.service @945ms +21ms
`-systemd-journal-flush.service @832ms +103ms
`-systemd-journald.service @491ms +287ms
`-systemd-journald-dev-log.socket @363ms
`--.slice @196ms
')
My setup is pretty vanilla, except for repo and AUR packages I've added *shrug* I sprung for the 64GB eMMC memory module, FWIW.

Clemmitt

[WarheadsSE]

I am going to take a random guess, and say that /mnt/data from /dev/mmcblk0p3 is being fsck'd for 9.255s .. and that is slowing things down.

[graysky]

Yes, I too thought that... but even removing it from /etc/fstab doesn't explain the network service issues :/

[graysky]

Ah! The problem is in fact with the mount. If I add 'noauto,x-systemd.automount' to the entry in /etc/fstab I get the expected times and fast boot. Not sure why the fsck is triggered on every boot....

$this->bbcode_second_pass_code('', '/dev/mmcblk0p3 /mnt/data ext4 defaults,relatime,noauto,x-systemd.automount 0 2')

$this->bbcode_second_pass_code('', '% systemd-analyze critical-chain
The time after the unit is active or started is printed after the "@" character.
The time the unit takes to start is printed after the "+" character.

graphical.target @6.507s
└─multi-user.target @6.502s
└─fail2ban.service @2.618s +3.877s
└─network.target @2.589s
└─systemd-networkd.service @2.489s +84ms
└─dbus.service @2.130s
└─basic.target @1.992s
└─sockets.target @1.987s
└─dbus.socket @1.982s
└─sysinit.target @1.977s
└─systemd-timesyncd.service @1.721s +249ms
└─systemd-tmpfiles-setup.service @1.668s +29ms
└─local-fs.target @1.661s
└─tmp.mount @1.642s +13ms
└─swap.target @1.632s
└─dev-mmcblk0p2.swap @1.595s +30ms
└─dev-mmcblk0p2.device @1.588s')