Forum Replies Created
-
Posts
-
Well, it’s hard to quantify because UDP (https://kb.nomachine.com/AR10T01174) is not available in SSH sessions and it also depends on the characteristics of the network. In terms of CPU usage, or when UDP is not in use, the difference between NX/SSH is negligible. You could take a look at the session statistics and check the bitrate/latency in the two set ups in your own test environments that you have set up. More about session statistics is available here: https://kb.nomachine.com/AR02N00875.
“We do not synchronize databases between the two ECS servers, but they can have the same nodes added to them.”
By this I mean that two ECS are completely separate hosts, they are independent from each other. The rules you create on one are separate from the same/different rules you create on the other. They can however share the same nodes. The shared nodes can be in different groups of the two independent cloud servers.
I understand that you want to synchronize DBs of two ECS. Synchronizing DBs of two ECS (Enterprise Cloud Server) is beyond the scope of their design. If you need to setup a cluster failover, you need to consider ECSC (Enterprise Cloud Server Cluster). Fine-tuning either, or any Enterprise server, in order to integrate with third-party products and discussing what architecture is best for your needs is ideally a job for our pre-sales channel. Please contact our sales team https://nomachine.com/contact-request with all the details.
Hi, we have tested with HAProxy, although not heavily. We aim to test a variety of third-party products to check that the NoMachine server can work out-of-the-box with NoMachine and that no special configuration is required on our side (https://kb.nomachine.com/DT04O00138). We cannot test all products on the market of course and for advanced configurations of these products, this is ultimately the responsibility of the customer or through a professional services contract. Our own previous tests have shown that 1. the client IP was not showing up, but rather the HA Proxy; 2. HAProxy on SSH works. It doen’t require any additional settings. If HAProxy has been configured to let terminal/console SSH connections work, SSH connections in NoMachine will work too; 3. we have not tested this mmproxy you mention, but it’s interesting to know that using it allows the client IP to show rather than the HAProxy IP. About what other customers are using, there have been a few cases of HAProxy, and also a number of others like Zscaler, Google Cloud Load Balancer (for web sessions in this case, https://kb.nomachine.com/AR03Q01019).
Hi, read this article which highlights the benefits of using NX over SSH.
What is the difference between the NX and SSH protocols?
https://kb.nomachine.com/AR11K00739Hi, this is a known issue reproduced on where Wayland is runnng, and there is a workaround.
As workaround change the following key in the nomachine node.cfg on the Ubuntu host to:
WaylandModes drm,compositor,eglYou can monitor the Trouble Report at the following link: https://www.nomachine.com/TR0110731.
This is an LDAP issue related to some mis-configuration which we were unable to pinpoint. We were not able to reproduce this issue even with the user’s same configuration.
In a working and correctly configured environment, LDAP and local users have separate home folders location. Home directories for LDAP users are mounted dynamically using autofs from an NFS share. Based on the logs provided, there is a problem with create file ‘options’ and setting ownership for it in the case of LDAP users (so those who have mounted homedir during login to system).
After considerable investigation, the workaround proposed in this case is to set the following key in the node configuration file to point to a local directory instead of the remotely mounted one: UsersDirectoryPath /some/local/path.
On ECS V8.16, when nxDB is modified and by who ?
We store there almost all data added by the administrator (such as nodes and rules), as well as data created when new sessions or connections are established. There is also some data related to this instance’s installation, such as a unique UUID
Are all the modifications under control of an human (administrator) or some modifications are done by internal processes ?
Some of the data is added by the administrator, while other data is generated while the server is running
We’d like to know if it’s possible, for a cluster of 2 ECS (not ECSC) to keep their nxDB synchronized just by running twice the admin commands (add node, add group of nodes, …).
Two ECS servers work as two separate machines, whereas a V8 cluster (ECSC) is designed to keep the host always available using active-passive heartbeat mechanics. We do not synchronize databases between the two ECS servers, but they can have the same nodes added to them.
Hi, thanks for the detailed report. This looks like a known issue. https://kb.nomachine.com/TR05V11141, specifically with KDE 6 onwards. What KDE version is it?
Hi, testing on the fly with default settings (out-of-the-box) we were not able to reproduce.
Can you send us the logs from the server host? You can extract them using the instructions here: https://kb.nomachine.com/DT07S00243.
We’d also need any backtrace which you can extract following the steps here: https://kb.nomachine.com/AR09L00810
Send everything to forum[at]nomachine[dot]com. Please use the title of this topic as the subject of your email. Thanks!
The macOS packages are suitable for both Intel and M1 architectures. You can downloaded the package you need from this link: https://downloads.nomachine.com/download/?id=46.
The NoMachine for ARM/Enterprise Desktop for ARM packages are RPM and DEB, so suitable for Linux devices such as the Jetson Nvidia.
Hi, did you already try restarting the nxserver from the server settings on the Ubuntu host?
Is Ubuntu running on a physical machine or is in a virtual machine? Is the display server X org or Wayland?
Are you able to send us the server-side logs so we can check them? You can extract them using the instructions here: https://kb.nomachine.com/DT07S00243. Please enable debug as explained in the instructions, then try restarting the server and extract the logs.
Send them to forum[at]nomachine[dot]com. Please use the title of this topic as the subject of your email. Thanks!
Support for authentication with smart card has been set-up by relying on the Public Key Infrastructure (PKI) and using an OpenSC compatible smart card, and using the SSH key method (see more about this here: https://kb.nomachine.com/DT07S00230#4, there is also a nice comparison table in chapter 2 which shows you what is supported).
The alternative is to use password authentication. However, we currently have a TR open related to USB forwarding and the remote desktop which shows the login window.