Overview

A Pod that will not boot usually points to a filesystem problem, a bad kernel or a broken configuration. Because you cannot SSH into a Pod that is not running, the ALYPOD Console's VNC console is your primary tool.

Before you begin

• Access to the ALYPOD Console.
• Any recent snapshot or backup you can restore if repairs fail.

Steps

1. Open the VNC console from the ALYPOD Console and watch the boot process. Note the exact point where it stops — that message tells you a lot.
2. If it stops at the bootloader, the boot entry may be pointing at a missing kernel. Boot the previous kernel or a recovery entry where available.
3. If it drops to a recovery or maintenance shell, the root filesystem may need checking:

bash fsck -y /dev/vda1

Replace /dev/vda1 with the correct device from lsblk. 4. If it hangs on a service, boot into rescue/single-user mode and disable the offending unit:

bash systemctl disable <service>

1. If the disk is full, the boot can stall while services fail to write. Free space from a recovery shell — see Disk is full: how to free space.
2. If it is a bad update or kernel, roll back using the package manager's history or select the older kernel at boot.
3. Check the configuration you changed last. A broken /etc/fstab or network config changed just before the failure is a frequent culprit. Correct it from the recovery shell.
4. Restore a snapshot if repairs are not working. In the ALYPOD Console, restore the most recent good snapshot or redeploy from your image and reattach data.

Troubleshooting

• Problem: The console shows no output at all. Solution: Power-cycle the Pod from the ALYPOD Console and try again; if there is still no output, contact support.
• Problem: fsck reports unrecoverable errors. Solution: Restore from a backup or snapshot; continuing to run the filesystem risks further data loss.
• Problem: It boots but a service fails to start. Solution: Read journalctl -u <service> -b for the specific error.

Related articles

• [Disk is full: how to free space]
• [My server is unreachable (network troubleshooting)]
• [Keeping your server patched]