nookup in web console

[fehrin1]

In the DB free web console on nslookup de.wikipedia.org i get

REFUSED

Is this a hard restriction of the free edition or can i work around this?

In some documentation i found that the free edition has external access to a limited set of sites only, is there a list of these sites?

[balajij8]

You can download the file, upload it & process from volume for files in blocked sites

[szymon_dybczak]

Hi @fehrin1 ,

In my case it works without issue on Free Edition:

[fehrin1]

Strange, for me it is 

 

[fehrin1]

@szymon_dybczak telnet 192.168.200.5 53 works. It seems like some port/protocol filtering/blocking. Maybe tcp access to 53 is allowed, but udp (or whatever nslookup uses) is not? Can i view and check which adresses, ports and protocols are blocked?

[szymon_dybczak]

Unfortunately, I don't know if such a list exists. But you can try to use nslookup with vc option which forces nslookup to use TCP protocol instead of UDP to verify that udp protocol is blocked.

Another reason why in your case you may be hitting this error is following limitations.

Maybe my free edition is hosted in different region where this domain is not blocked. 

[fehrin1]

I am us-east-2 and nslookup -vc gives also REFUSED only. Could this be a very recent hardening (my Databricks free edition is only a few days old)?

[fehrin1]

This is not a port/protocol problem, but a whitelist thing. nslookup wikipedia.org fails, while nslookup pypi.org yields valid results.

It still would be interesting to see the whitelist of allowed hosts/domains. I will ask this as a separate question.

[szymon_dybczak]

HI @fehrin1 ,

That was my second guess in above answer. So basically wikipedia domain is not added to allowed host/domains in region your Free Editon is hosted.

[fehrin1]

Hi @szymon_dybczak my region is us-east-2, what is yours?

[szymon_dybczak]

My is eu-west-1. So I guess de wikipedia is here added to a whitelist 🙂

[pradeep_singh]

there is no list of sites .

[fehrin1]

So why is wikipedia.org accessible with a free edition hosted in eu-west-1 (-> @szymon_dybczak), but not with a free edition hosted in us-east-2 (-> me)?

[SteveOstrowski]

Hi @fehrin1,

This is expected behavior on the Databricks Free Edition. The Free Edition documentation confirms that "outbound internet access is restricted to a limited set of trusted domains."

https://docs.databricks.com/aws/en/getting-started/free-edition-limitations

This means DNS resolution (nslookup) and actual network connectivity will only succeed for domains on that trusted allowlist. Package repositories like pypi.org are included because they are needed for installing Python libraries, but general-purpose sites like wikipedia.org are not on the list. That is why your nslookup for de.wikipedia.org returns REFUSED while pypi.org resolves successfully.

A FEW THINGS TO NOTE

1. There is no publicly documented list of the specific trusted domains. The allowlist includes domains required for core Databricks functionality and common package repositories.

2. The behavior you observed where another community member (@szymon_dybczak) in a different region could resolve wikipedia.org may reflect differences in how the restrictions are applied across regions or rollout timing. The restriction is the documented intended behavior.

3. This restriction applies specifically to the Free Edition's serverless compute environment. Paid Databricks workspaces on AWS, Azure, or GCP do not have these outbound network restrictions (though administrators can configure their own network policies).

WORKAROUNDS

If you need to work with data from a site that is blocked:

1. Download the data to your local machine first, then upload it to your Databricks workspace using the file upload feature in the UI or through a Volume.

2. If you are working with public datasets, check whether they are available through a package repository that is already on the allowlist (for example, datasets available via pip-installable Python packages).

3. For more flexible network access, consider upgrading to a Databricks trial with a cloud provider account (AWS, Azure, or GCP), which provides a 14-day full-featured trial without these egress restrictions.

https://docs.databricks.com/aws/en/getting-started/free-trial

* This reply used an agent system I built to research and draft this response based on the wide set of documentation I have available and previous memory. I personally review the draft for any obvious issues and for monitoring system reliability and update it when I detect any drift, but there is still a small chance that something is inaccurate, especially if you are experimenting with brand new features.