Category: Computing

In December 2017, I wrote a post about how to use Quad9 DNS resolution with Algo VPN, a free and open-source project by Trail of Bits that lets you easily set up your own non-logging VPN server.

My earlier tutorial no longer works with versions of Algo VPN since 25 April 2018, when the Algo VPN code base was changed to enforce encrypted DNS lookups (and five days later, WireGuard support was added, yay!). By default, Algo now uses Cloudflare’s secure DNS lookup. Quad9, which I prefer because of its option to block known malicious sites, implemented DNS over HTTPS in October 2018, and you can now use it with Algo VPN instead of Cloudflare if you like. I will provide step-by-step instruction on how to do this below.

Bear in mind that for these instructions to work, you must be using an Algo installation made with the code base since the April 2018 revision implementing secure DNS lookup:

1. SSH to your Algo server via your terminal app:

ssh root@ip -i ~/.ssh/algo.pem

2. Open /etc/dnscrypt-proxy/dnscrypt-proxy.toml for editing:

nano /etc/dnscrypt-proxy/dnscrypt-proxy.toml

3. Find the line:

server_names = ['cloudflare', 'cloudflare-ipv6']

4. Replace with:

server_names = ['quad9-dnscrypt-ip4-filter-pri', 'quad9-dnscrypt-ip4-filter-alt', 'quad9-doh-ip6-filter-pri', 'quad9-doh-ip6-filter-alt']

5. Hit Control-X to exit, choosing Y to save and keeping the same file name.

6. Type reboot at the command line to reboot.

That’s all! Your Algo VPN server will now use Quad9 for secure DNS resolution.

Note that there are other secure DNS servers that you can set as parameters for server_names. You’ll find them listed here.

After upgrading an instance of NextCloudPi from version 13 to version 14, I received the following error notification:

In short, the table “oc_share” was missing the index “parent_index” and the table “oc_filecache” was missing the index “fs_mtime”.

I tried SSH-ing to my NextCloudPi server and running the suggested command, “occ db:add-missing-indices” but it generated an error.

Despite the missing database entries, the instance still seemed to run fine, but I found the constant error message vexing.

After some online research and examination of my NextCloudPi instance’s directory structure, I found the following solution, and not seeing it posted elsewhere, thought I’d share it here:

1. SSH to your NextCloudPi instance.

2. Get a root prompt by entering:

sudo bash

3. Go up to the root directory by typing the following command twice (or as many times as necessary):

cd ../

4. Switch to the nextcloud directory:

cd /var/www/nextcloud

5. Run the following command:

sudo -u www-data php occ db:add-missing-indices

6. You should receive something like the following prompts (which I received):

Check indices of the share table.
Adding additional parent index to the share table, this can take some time...
Share table updated successfully.
Adding additional mtime index to the filecache table, this can take some time...
Filecache table updated successfully.

7. That’s all! Type “exit” and hit enter to terminate the SSH session. The error message should no longer appear.