All in all everything is working. Only issue: mouse and keyboard in web VNC and with a VNC viewer for a Windows 10 VM do not work.
I’ll have to investigate for more details. Might be a local issue specific to my NAS. I have been running the nightlies on a test system for a couple of weeks including installing a Windows 10 VM from scratch without that problem surfacing.
I’m also not happy with the decision to pull the remove the shell. Why not just fix it? Oh well.
For me, since I do not keep a terminal session open (putty), the loss of the GUI shell means I must open an ssh putty session to a server & log in. Since I already have to do that for each server’s gui, where the shell feature met my needs as-is, the change, for me, is more work & slower. I can live with it but wish iX would say why the change was needed.
John
Windows 10 and 11 include a SSH client, so PuTTY is no longer needed unless you’re still using Windows 7.
They have: it sucked. Basic control characters didn’t work. Last I checked, the arrow keys didn’t work. Copy and paste didn’t work. Why they couldn’t replace it with the one from SCALE (which works reasonably well) I don’t know, but if they couldn’t fix it, better IMO to get rid of it.
Nice slip of the keyboard… 
Replication is broken: [NAS-128866] - iXsystems TrueNAS Jira
I can confirm this too, the web VNC for the VMs does not work correctly - it only shows a static image although keyboard inputs seems to go trough as verified by refreshing the page manually
Same here. Display doesn’t update until page is reloaded.
No Cloudflare ACME option?
Good news the fix is in the nightly build now and we had someone deploy the nightly to their system and confirmed the replication with NETCAT is now running successfully.
Does this include SSH+NETCAT-based replication between systems of different builds, i.e. 13.3 to 13.0U6.1?
Also, I just found out a great way to nuke 700GB of replication progress - i.e. restart your NAS due to a system upgrade. All progress is lost and the task has to start anew.
Yeah, I thought this was supposed to be in 13.3, but it sure isn’t in the BETA.
The bug makes the local execution of the zettarepl script fail - on both ends if they run BETA1 and the method is SSH+NETCAT. Reverting to SSH only fixes it.
I think it’s a reasonable assumption that if one end is BETA1 the version of the other end is not really relevant. It fails with SSH+NETCAT, it works with SSH. I assume.
Well, of course it is not hard. And that’s my choice anyway when I’m home. I even have some shortcuts that open the ssh shell of a target machine directly. There is a but: from some locations ssh is blocked e.g. from my company’s offices and getting into shell via the web gui was a nice option. I still have a solution - I can get in via web gui of another server running cockpit on debian and ssh to TrueNas from there. But that is an extra step which was not needed so far.
It is two months later and it is still in beta. So there is no rush. And moreover seems that 13.3 does not bring any important feature in, so no rush at all. Btw, you could run a small vm with e.g. Debian and Cockpit and use it as a web gui entry point to your system. Obviously an extra step in the process but if someone needs the web shell this might be an easy option.
