Why does my Linode using SELinux have issues booting after a Backup restore?
After I restored from a backup, my Linode using SELinux appears to get stuck during the boot sequence and doesn't show a login prompt in the LISH console. The boot sequence seems to stop after output that looks like this:
[ 10.082030] AVX2 version of gcm_enc/dec engaged.
[ 10.083092] AES CTR mode by8 optimization enabled
[ 10.090444] iTCO_wdt: initialized. heartbeat=30 sec (nowayout=0)
[ 10.503246] alg: No test for __gcm-aes-aesni (__driver-gcm-aes-aesni)
[ 10.506607] alg: No test for __generic-gcm-aes-aesni (__driver-generic-gcm-aes-aesni)
[ 10.540377] Decoding supported only on Scalable MCA processors.
[ 34.579467] random: crng init don
I tried rebooting the Linode, but the same thing happened.
After waiting a while, the Linode eventually rebooted itself and was powered back on by the Lassie watchdog service. The Linode's boot sequence completed normally after that, and I was able to login.
What happened?
1 Reply
✓ Best Answer
Since our Backup Service does not restore security contexts for SELinux, the Linode's internal file system must be relabeled in order to rebuild these security contexts after a backup is restored.
This process may take a considerable amount of time if the Linode contains a large number of files, and you may not see any output in LISH indicating that this happening, so you'll need to wait on this process to complete. The operating system should try to automatically reboot the Linode after relabeling is finished, but please be aware that if Lassie is not enabled, you'll need to power on the Linode manually since Linodes remain powered off when rebooted internally.
Alternatively, you can switch to a Linode provided kernel if you don't need to use SELinux since our kernels don't support SELinux. This will allow the Linode to boot up normally when powered on without having to wait for the file system to be relabeled as SELinux is no longer active.