Hello,
Would have a chance to set up Linux
OpenSuse 42.1 Leap to be “multiseat”?
Thank you a lot!!
And what is your definition of “multiseat”? Do you perhaps mean multi-user? If so, openSUSE, being a linux distribution, and unix-like, is multi-user naturally.
In my case it would be the following.
Two people to use a single computer on different monitors each with its keyboard and mouse.
If I understand what you are saying, “seats” typically is a term often used describing licenses… Typically the number of Users who are permitted to access and use the install instance according to various definitions (eg concurrent, total, logged in, etc).
Linux is also free, open source and open to <use> without restriction (there are restrictions on copying and distributing), and is even multi-User (multiple Users can log in and use concurrently).
TSU
Would you like to share any more experiences around software evolution for multiseat, multi-station or multiterminal configurations?
https://en.wikipedia.org/wiki/Multiseat_configuration
By that definition,
I’m not sure how practical that implementation is today compared to multiple Users remoting into a common system using SSH or similar, or a variety of virtual desktop software.
The defined “multi-seat” in the above wikipedia article requires extensive hardware setup and configuration. It’s marginally cheaper than using fully independent dumb clients which today might be as cheap as a RPi Model B (approx $35 with ethernet ports), maybe cheaper (Maybe the RPi zero with a USB network dongle? Maybe a $9 CHIP with similar?).
When hardware was more costly, a “multiseat” might have made more sense, I don’t know that is the case today.
IMO,
TSU
- How do you think about to distribute computation power and management efforts between local workstations and bigger machines (or clusters) to use various services and resources in a more efficient and safer way?
- How flexible would you like to handle such a mixture?
Further examples:
Have you got any expectations or special requirements for the connection of such devices to your seats?
Thanks for the help and clarification.
Based on the given definition of multiseat, it’s severely restricted by how many devices can be connected directly into the Server… connections to monitors, connectors to I/O devices like keyboards and mice. Off the shelf I/O hardware generally assumes a limited number of connections per hardware expansion card and your backplane will limit the number of expansion hardware, eg 1 or 2 video connections, single mouse and keyboard. If you want something with greater capacity, you need expansion hardware and perhaps unusual, custom hardware.
Compare that with the cost I quoted for a “thin client” (not a full workstation) connecting over standard, high capacity media like ethernet. Especially for very high numbers of clients(equivalent of seats) alternatives are nearly the same or even cheaper. You’d largely still use the same or similar necessary Linux multi-User features without the higher hardware costs.
So, I doubt you’ll find much effort or “success stories” within the past 4 years or so.
TSU
- There are some constraints to consider for the use of such simple clients.
- Would you often like to increase the usage of resources that ordinary workstations can provide?
So, I doubt you’ll find much effort or “success stories” within the past 4 years or so.
Would any other contributor like to point further activities out from this application domain?
[quote="“elfring,post:11,topic:114733”]
Would any other contributor like to point further activities out from this application domain?[/QUOTE]Before xrandr gained the ability to bridge multiple graphics adapters into a contiguous output, I used to run a multiseat configuration on my workstation. Two monitors in Screen 0 attached to X server Display :1, and three monitors in Screen 0 attached to X server Display :0. I used a single keyboard and mouse for both, accessing the second Display by using http://synergy-project.org/
I don’t recall over which openSUSE releases this was done, but it was for a couple of years and releases. I talked about it several times here in the forums IIRC
How do you think about to help in achieving progress on issues like the following?
Best asked in a Xephyr forum. Why ask her were few if any use it??
Thin clients don’t necessarily mean simple in computing functionality. In this case, since the User is logged in remotely to the Server, the end result is almost no difference compared to your multiseat configuration because the remote thin client is only providing input/output for the User. All actual computing is done on the Server/Workstation. If you think there is a real constraint, please state what it might be.
Are you saying that you’re trying to increase the utilization of workstations so are considering configuring workstations as a multiseat? If so, this is one of the fundamental reasons for deploying the multiseat alternatives I described… Virtual desktops (This is VMware’s VDI, but all virtual technologies can be configured similarly), Terminal Services, and in the case of Linux multiple clients can remote in using SSH. You <might> avoid an incremental amount of overhead implementing an all-hardware, local-only setup, but likely run into a number of obstacles scaling like I described… So, today multi-seat is not in favor compared to the alternatives I listed.
TSU
Can it be that this software is used more often than we might know at the moment?
- Which network speed and how many data transfer traffic can you afford (for example)?
- Do you need any special devices for your work flow?
Are you saying that you’re trying to increase the utilization of workstations so are considering configuring workstations as a multiseat?
When would you start to turn a computer for a single user into a server for a small workgroup by extending its device connections?
If there were more using it that would have jumped in. Most will go a thin clients route today. And I’d point out that the multi-user set ups are normally USB based and that has a lot less bandwidth then Ethernet. It is also more complex to set up and may require special hardware for really large numbers of seats.
Which connection types do you prefer for your pointing devices and screens?
And I’d point out that the multi-user set ups are normally USB based and that has a lot less bandwidth then Ethernet.
Is the popularity and speed evolving for both technologies?
It is also more complex to set up and may require special hardware for really large numbers of seats.
Are you looking for updates to a description like “Build a Six-Headed, Six-User Linux System” by Bob Smith](Build a Six-headed, Six-user Linux System)?
Don’t use multi seat or thin client I have no need. If I did I’d go with thin client since it does not require special hardware
Got no clue
Nope
Personally I think you are being a troll and will stop trying to help you in the future