coreos-install hangs after failure to reread partition table

[bgilbert]

Issue Report

Bug

Container Linux Version

$ cat /etc/os-release
NAME="Container Linux by CoreOS"
ID=coreos
VERSION=1675.0.0+2018-02-05-2246
VERSION_ID=1675.0.0
BUILD_ID=2018-02-05-2246
PRETTY_NAME="Container Linux by CoreOS 1675.0.0+2018-02-05-2246 (Rhyolite)"
ANSI_COLOR="38;5;75"
HOME_URL="https://coreos.com/"
BUG_REPORT_URL="https://issues.coreos.com"
COREOS_BOARD="amd64-usr"

Environment

Any

Expected Behavior

If coreos-install can't reread the partition table after installing, it immediately wipes the partition table and exits.

Actual Behavior

gpg: Signature made Thu 01 Feb 2018 12:57:37 PM PST using RSA key ID EF4B4ED9
gpg: key 93D2DCB4 marked as ultimately trusted
gpg: checking the trustdb
gpg: 3 marginal(s) needed, 1 complete(s) needed, PGP trust model
gpg: depth: 0  valid:   1  signed:   0  trust: 0-, 0q, 0n, 0m, 0f, 1u
gpg: Good signature from "CoreOS Buildbot (Offical Builds) <buildbot@coreos.com>"
blockdev: ioctl error on BLKRRPART: Invalid argument
Failed to reread partitions on /dev/loop0
blockdev: ioctl error on BLKRRPART: Invalid argument
Failed to reread partitions on /dev/loop0
blockdev: ioctl error on BLKRRPART: Invalid argument
Failed to reread partitions on /dev/loop0
blockdev: ioctl error on BLKRRPART: Invalid argument
Failed to reread partitions on /dev/loop0

and then it hangs.

Reproduction Steps

1. dd if=/dev/zero of=img bs=1M count=1 seek=8192
2. losetup /dev/loop0 img
3. coreos-install -d /dev/loop0

Other Information

write_to_disk has no way to report a failure to the parent process when running as a subshell, so the parent process blocks until timeout. coreos/init#262 increased the timeout to two hours.