tests/boot-mirror: sleep 20s after detaching primary block device

For some reason, if we try to reboot too quickly after detaching the
primary block device, the OS will sometimes crash with:

```
[  100.662358] watchdog: watchdog0: watchdog did not stop!
[  100.969436] watchdog: watchdog0: watchdog did not stop!
[  100.998017] BUG: Unable to handle kernel data access at 0x5deadbeef0000100
[  100.998158] Faulting instruction address: 0xc000000000f219d4
[  100.998264] Oops: Kernel access of bad area, sig: 11 [coreos#1]
[  100.998348] LE PAGE_SIZE=64K MMU=Radix SMP NR_CPUS=2048 NUMA pSeries
[  100.998454] Modules linked in: rfkill crct10dif_vpmsum binfmt_misc raid1 xfs zram virtio_net net_failover vmx_crypto
 crc32c_vpmsum pseries_wdt virtio_console failover virtio_blk scsi_dh_rdac scsi_dh_emc scsi_dh_alua ip6_tables ip_tables dm_multipath fuse
[  100.998822] CPU: 0 PID: 1 Comm: systemd-shutdow Not tainted 6.4.15-200.fc38.ppc64le coreos#1
[  100.998947] Hardware name: IBM pSeries (emulated by qemu) POWER9 (raw) 0x4e1203 0xf000005 of:SLOF,HEAD hv:linux,kvm pSeries
[  100.999107] NIP:  c000000000f219d4 LR: c000000000f219c8 CTR: c0000000001c60e0
[  100.999229] REGS: c0000000085a3860 TRAP: 0380   Not tainted  (6.4.15-200.fc38.ppc64le)
[  100.999352] MSR:  800000000280b033 <SF,VEC,VSX,EE,FP,ME,IR,DR,RI,LE>  CR: 44442404  XER: 00000092
[  100.999511] CFAR: c000000000f2191c IRQMASK: 0
[  100.999511] GPR00: c000000000f219c8 c0000000085a3b00 c000000001eea800 c000000002c88ac8
[  100.999511] GPR04: c000000002c88ac8 0000000000000001 0000000000000001 fffffffffffe0000
[  100.999511] GPR08: 0000000000000001 0000000000000001 5deadbeef0000100 0000000000002000
[  100.999511] GPR12: 0000000000000000 c000000002ca0000 0000000000000000 0000000000000000
[  100.999511] GPR16: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
[  100.999511] GPR20: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
[  100.999511] GPR24: 0000000000000002 0000000000000000 c000000002c888e0 c00000000294e6e8
[  100.999511] GPR28: c000000002c88ac8 5deadbeeeffffd28 c00000001f2f1218 c0000000b8d9d000
[  101.000532] NIP [c000000000f219d4] md_notify_reboot+0x154/0x250
[  101.000643] LR [c000000000f219c8] md_notify_reboot+0x148/0x250
[  101.000748] Call Trace:
[  101.000793] [c0000000085a3b00] [c000000000f219a0] md_notify_reboot+0x120/0x250 (unreliable)
[  101.000922] [c0000000085a3b60] [c000000000199e30] notifier_call_chain+0xc0/0x1b0
[  101.001049] [c0000000085a3bc0] [c00000000019a114] blocking_notifier_call_chain+0x64/0xa0
[  101.001176] [c0000000085a3c00] [c00000000019dbb8] kernel_restart+0x38/0xe0
[  101.001282] [c0000000085a3c70] [c00000000019dfbc] __do_sys_reboot+0x12c/0x2c0
[  101.001409] [c0000000085a3dd0] [c000000000030f34] system_call_exception+0x174/0x320
[  101.001537] [c0000000085a3e50] [c00000000000d05c] system_call_vectored_common+0x15c/0x2ec
[  101.001681] --- interrupt: 3000 at 0x7fffb995aa88
[  101.001767] NIP:  00007fffb995aa88 LR: 0000000000000000 CTR: 0000000000000000
[  101.001888] REGS: c0000000085a3e80 TRAP: 3000   Not tainted  (6.4.15-200.fc38.ppc64le)
[  101.002010] MSR:  800000000280f033 <SF,VEC,VSX,EE,PR,FP,ME,IR,DR,RI,LE>  CR: 48442403  XER: 00000000
[  101.002167] IRQMASK: 0
[  101.002167] GPR00: 0000000000000058 00007fffd8caa5c0 00007fffb9a76f00 fffffffffee1dead
[  101.002167] GPR04: 0000000028121969 0000000001234567 0000000000003a5d 0000000000000020
[  101.002167] GPR08: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
[  101.002167] GPR12: 0000000000000000 00007fffba163680 0000000000000000 0000000000000040
[  101.002167] GPR16: 0000000000000138 0000000000000001 000000010b9213c8 000000010b921a90
[  101.002167] GPR20: 0000000000000000 000000010b9212e0 0000000000000010 000000010b921318
[  101.002167] GPR24: 000000010b9212d0 00007fffd8caa6e0 00007fffd8caa6f8 00007fffd8caa6d8
[  101.002167] GPR28: 00007fffd8caada8 00007fffd8caa6e8 00007fffd8caa6c8 0000000000000000
[  101.003150] NIP [00007fffb995aa88] 0x7fffb995aa88
[  101.003234] LR [0000000000000000] 0x0
[  101.003299] --- interrupt: 3000
[  101.003363] Code: 7f84e378 48423c31 60000000 2c030000 4182000c 7fe3fb78 4bff859d 7f83e378 48465645 60000000 39000001 395d03d8 <e93d03d8> 7fbfeb78 7c2ad800 3929fc28
[  101.003604] ---[ end trace 0000000000000000 ]---
[  101.014145] pstore: backend (nvram) writing error (-1)
[  101.014255]
[  102.014301] note: systemd-shutdow[1] exited with irqs disabled
[  102.014503] Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
```

The `md_notify_reboot` implicates the MD code in the kernel.

Obviously this is likely a bug that needs fixing

mantle/kola/tests/misc/boot-mirror.go

@@ -222,6 +222,7 @@ func detachPrimaryBlockDevice(c cluster.TestCluster, m platform.Machine) {
 		if err := m.(platform.QEMUMachine).RemovePrimaryBlockDevice(); err != nil {
 			c.Fatalf("failed to delete the first boot disk: %v", err)
 		}
+
 		// Check if we can still SSH into the machine. We've noticed sometimes
 		// that after removing the primary device, we lose connectivity.
 		if err := ut.Retry(5, 3*time.Second, func() error {
@@ -230,6 +231,10 @@ func detachPrimaryBlockDevice(c cluster.TestCluster, m platform.Machine) {
 		}); err != nil {
 			c.Fatalf("Failed to retrieve boot ID: %v", err)
 		}
+
+		// Give some time to the host before doing the reboot.
+		time.Sleep(20 * time.Second)
+
 		err := m.Reboot()
 		if err != nil {
 			c.Fatalf("Failed to reboot the machine: %v", err)