How to fix an unreachable DreamCompute Instance


This guide is a checklist of common issues that can be investigated before contacting support. Please keep in mind that DreamHost support does not have access to your instances and cannot view your data. To help resolve the issue, please provide any error messages, logs or other details you may see.

Ensure the instance is running

Log in to the DreamCompute dashboard and examine the instances page. Examine the Status and Power State columns for the instance. Some common combinations of these values and their meaning are below:

Status Meaning
Active If 'Power State' shows as 'Running', the instance is up and running. Everything should be configured properly within DreamCompute.
Error The instance has encountered a serious error on the DreamCompute side. The exact error can be seen by clicking on the instance name, then the Overview tab. This error can only be reset by an administrator. Please contact support.
Revert Resize/Migrate 
A resize was attempted on the instance. Please confirm or revert the resize by using the dropdown menu to the right in order to Confirm or Revert.
Paused The instance was paused from within the dashboard or with the API. Click the dropdown menu item to resume the instance. If you did not do this, consider changing your dashboard/API password.

The instance was shut down. This can happen if someone with root privileges on the instance ran a command like 'shutdown -h now', there was a kernel panic, or some internal error occurred within DreamCompute itself.

Before starting the instance up again, click the instance name, then the Log tab. It shows the last console output before shutdown. This can be used to see if it was an intentional shutdown or another issue. If you contact support about this, please click the 'View Full Log' button and include the output in the support ticket.

Contact support for any other status you may see.

Check the instance log for boot errors

On the Instances page, click your instance. Then select the Log tab. This displays the last 35 lines of the consoles output. A properly booted instance has a login prompt which could be slightly buried in the log. An Ubuntu-16 example would look like this:

[[0;32m  OK  [0m] Started Execute cloud user/final scripts.
[[0;32m  OK  [0m] Reached target Cloud-init target.

Ubuntu 16.04.1 LTS test ttyS0

instancename login:

Error messages such as failing to find a disk, file system errors, or syntax errors in config files will also generally display there. These type errors are more difficult to fix without creating a new instance, however a ticket to support is recommended.

Check the instance log for network errors

On the same 'Log' page as above, click the View Full Log button on the top right. This opens a new browser tab with the full console log output. Search for 'eth', 'ens, or 'cloud-init'. Some problems may include not finding a network device, or if a snapshot was used, the network not being re-configured.

Compare the output of cloud-init network configuration (if available), to the assigned IPv4 and IPv6 addresses of the instance. An example output snippet for Ubuntu-16 looks like this:

Starting Initial cloud-init job (metadata service crawler)...
cloud-init[982]: Cloud-init v. 0.7.8 running 'init' at Tue, 02 Apr 2024 21:32:28 +0000. Up 8.86 seconds.
cloud-init[982]: ci-info: ++++++++++++++++++++++++++++++++++++++Net device info+++++++++++++++++++++++++++++++++++++++
cloud-init[982]: ci-info: +--------+------+------------------------------+---------------+-------+-------------------+
cloud-init[982]: ci-info: | Device |  Up  |           Address            |      Mask     | Scope |     Hw-Address    |
cloud-init[982]: ci-info: +--------+------+------------------------------+---------------+-------+-------------------+
cloud-init[982]: ci-info: |  ens3  | True |        | |   .   | fa:16:3e:a5:4c:1c |
cloud-init[982]: ci-info: |  ens3  | True | fe80::f816:3eff:fea5:4c1c/64 |       .       |  link | fa:16:3e:a5:4c:1c |

This example instance has an IPv4 address of Confirm this is the correct IP on the DreamCompute dashboard instances page.

Check security group rules

On the Instances page to the far right, you'll see a column titled Actions. Under this is a dropdown. Click this and select 'Edit Security Groups'. Confirm the necessary security groups are in the 'Instance Security Groups' list on the right. Make a note of the list of security groups to check next.

Navigate back to the Instances page. In the left panel, click 'Network > Security Groups' page. Check that the security group(s) assigned to the instance have the ports open that are being checked. To do this, click on the Manage Rules button to see the list of rules. All instances should have rules for unrestricted egress (outbound) traffic like so:

Egress      IPv6    Any     Any     ::/0            -
Egress      IPv4    Any     Any       -

The other default rules that DreamHost puts in the 'default' security group allow for ping (ICMP), SSH (port 22), HTTP (port 80) and HTTPS (port 443):

Ingress     IPv6     58     Any             ::/0            -
Ingress     IPv4    ICMP    Any          -
Ingress     IPv6    TCP     22 (SSH)        ::/0    -
Ingress     IPv4    TCP     22 (SSH)       -
Ingress     IPv6    TCP     80 (HTTP)       ::/0    -
Ingress     IPv4    TCP     80 (HTTP)       -
Ingress     IPv6    TCP     443 (HTTPS)     ::/0    -
Ingress     IPv4    TCP     443 (HTTPS)       -

If any of these rules are missing, consider adding them to restore the default functionality. Sometimes the 'allow everything' of IPv4 and IPv6 ::/0 is modified to allow just specific IP blocks. If so, confirm the IP range is sufficient for the connectivity desired.

Try a reboot, just in case

It is preferred to find the cause of an issue before rebooting an instance, but if none of the solutions above fixed the issue, you should now attempt a reboot. If this fixes the issue, include the previously gathered console logs for support if a support ticket is opened.

Consider changes made before the issues

Lastly, if all the above seems correct but there are still issues, consider what changes were last made. Some examples of changes that can cause issues are additional firewalls, large system upgrades that could modify kernels or network systems, or modifying config files that could disrupt the boot process if a program won’t start.


Should there be no other indication of issues in the above logs and configurations, please contact support with all available information.


Did this article answer your questions?

Article last updated PST.

Still not finding what you're looking for?