Stack Manager for Windows (Do you want to run your own domain?)


Hi All,

We now have a Stack Manager available for Windows and OSX, allowing you to easily run your own local domain. You can download it here:



When you run the Stack Manager for the first time it will download all the dependencies that it needs. You will then be presented with the following screen:

This panel allows you to stop and start all the services that you need.

Have fun and if your domain is public then make sure to update the known domains on the wiki so we can come and check them out.

See you in world.



Why did you put the windows version behind MAC: ? Small typo i think.

Anyway, cool !
Need to take a look at it soon.


thanks @Richardus I just corrected it


Other important question, wich ports need to be open ?
Good place to write it down in this thread.


To run locally and connect locally you do not need any. To access externally you need to make sure port 40102 is open and forwarded to the machine running the domain-server.


@chris Is it possible to enable the logs? If so, how do we do that?

  • Dave


Well with a ton of @thoys help I got it working

Heres what I did
download and install the link from above
run it and turn all the buttons on
then in hifi go to @localhost , if it says u have servers running then its working,
next part is to access it over the net
first u need to log into your router mines a netgear one to log in i put into my browser password admin and what ever your password is
then i head to advanced-portforwarding, click add custom service
i called mine hifidomain and entered 40102 as both port numbers so from 40102 to 40102
you also need to know if you have more than one computer connecting to your router what the internal ip of it is
that’s found under attached devices and will look like
ok once u have that all entered and saved
google whats my ip adress and it will give u a number take that number add hifi:// infront of it
giving you
paste that into the go to place thing in hifi press @ to open it
and you should be there

getting a custom url is a bit trickier and thoys did that for me lol so thanks man


The stack manager runs smooth.

Started it up a few days ago, rezzed a few voxels and mesh.


Going to try it in hour. run it local on other server special assigned for hifi. Judus already explained my last question. just use hifi://domainname to get there.


you better use only domain name in goto domain. but still it doing weird. i can get to my server. Seeing myself double also it looks like some avatar is faiiling constantly when am on my own server. interface console is spamming with hash mismatch. so its working, but with lots of ductape :smile:


We provided a Windows 8.1 pro system based domain server at
hifi:// only port 40102 has been opened externally through our University firewall as requested.

for tests and that will shortly be available at stable domain name of


Visit in this corner to see a mesh model…



I also see the “double” and “ghostly glow” copy of an avatar at the last location it was at before moving between domains. It seems to be there until the same avatar arrives back and then it refreshed to the solid coloured “real” version. In the image below, @Be is actually in another domain at the time.


As others have asked… is there any way to get some feedback on what, if anything, is happening on a domain server via some (minimal) activity in the Stack Manager console windows or on the control panel itself. Its unclear what is happening if anything and whether a domain is in use if it is to be shutdown.

Showing the number of avatars present (other than via the http: status interface “agents” rows) might be helpful.

Also, I see 6 servers running for the domain, but that stays at 6 even if I close down oe or two,… e.g. the Audio Mixer and Particle server. Even after logging off and back on to a copy of the Interface. is tha correct?

If I close down services, the http status web page shows that they actually were restarted… even though I have said Stop and the control panel shows that.

If I stop the entire domain server in the control panel it seems to still run and the status page shows that, and avatars can get there via the Interface. Yet there is no indication on the server machine that things are active.

Its not clear to me what is being added to the system in the way of data and content and where its stored or how much space is allocated to that. can we get a description please? Its fine just now while we have s small testing group, but t make this scalable and manageable more information and tools to manage and control things will be needed.

Can @Chris pass on these issues to the Domain Server folks?


It looks like there are a few (no doubt several) bugs on windows. I have put some of these into a worklist job . There are a lot of additional features that we can add to the stack manager. This is a great thread to get that discussion going.


I am now testing from outside our firewall to the windows domain server on hifi:// (which is also hifi:// and I am getting masses of packet mishmash message which @Judas also reported seeing. And Error getting domain settings - “Connection refused” - retrying

[2014-08-01T20:43:03] Packet hash mismatch on 8 - Sender QUuid("{3d8b37a3-35b6-1eb-9431-e3e37fc119d5}")
[2014-08-01T20:43:03] Packet hash mismatch on 11 - Sender QUuid("{a1b30aad-daf74048-91dc-9c3e7974bdc1}")
[2014-08-01T20:43:03] Packet hash mismatch on 8 - Sender QUuid("{3d8b37a3-35b6-1eb-9431-e3e37fc119d5}")
[2014-08-01T20:43:03] Packet hash mismatch on 5 - Sender QUuid("{a1b30aad-daf7-048-91dc-9c3e7974bdc1}")
[2014-08-01T20:43:03] Removing Avatar with UUID QUuid("{af32147b-8511-48bb-a078e4e027bcc834}") from AvatarHashMap.
[2014-08-01T20:43:03] Packet hash mismatch on 8 - Sender QUuid("{3d8b37a3-35b6-1eb-9431-e3e37fc119d5}")
[2014-08-01T20:43:04] Error getting domain settings - “Connection refused” - retrying
[2014-08-01T20:43:04] Requesting domain-server settings at “”
[2014-08-01T20:43:04] Error getting domain settings - “Connection refused” - retrying
[2014-08-01T20:43:04] Failed to retreive domain-server settings 5 times. Re-setting connection to domain.

Also note typo on spelling of "retreive " in
Failed to retreive domain-server settings 5 times. Re-setting connection to domain.


@Chris, I think :40100 MUST also be open as well as :40102?

This is being requested by the Interface and failing if its not open I think…

But when I look at this WITHIN our firewall it serves this content…

“voxels”: {
“per-meter-cubed-credits”: “0.0”,
“per-voxel-credits”: “0.0”,
“voxel-wallet”: “”

If so the admin/server status/shutdown page on MUST be moved into a different port so that is not exposed outside firewalls.

Can we get the DEFINITIVE set of firewall ports required to be open to the world for the domain server on Windows and whether they are for TCP, UDP or both, and the protocols used on them, This sort of detail is required to request firewall holes.

Even if :80 for http via TCP is needed, please say so, as it wrong to assume that port even is open on such domain servers unless its necessary.

I have to open such holes at server, school and university levels so it would be good if we can get it right first time.

I only opened :40102 as request today via a priority request. Can’t do that all the time.


I assume that json served setting information needs to be set somewhere on the domain server… its not asked for when the server is started.

Is some sort of info needed

…user…\AppData\Local\High Fidelity\Stack Manager\resources\web\settings


Have the packet mismatch to, but only opened port 40102, thats what im told on this forum. Now i remeber its possible safer to add it manual in the windows firewall to. because the automatic handling from windows only works for one way. If we need to open more ports i like to know. 40100 ?? We need to avoid at all costs that HiFy wants to use port 80 for server things. because isp’s can block that port. Otherwise routers go make troubles


I see that there is a settings web page on :40100
where the Destination Wallet ID can be set. but that makes me realise
more that :40100 cannot be left open to an external connection until
there is much greater security of that page in place.

But it does seem that the interface assumes port 40100 needs to be- open to READ the JSON delivered settings.


Other ports seem to be in use according to the status page. As some of us observe issue with the audio services trying and failing to connect, could it be that even more ports than 40100 and 40102 must be open through firewalls?

58213 particle-server
58214 metavoxel-server

58954 audio-mixer
58955 avatar-mixer
58956 voxel-server
58957 model-server

Interface connection…
Users IP:50460