This have been tested with 1.54 AND the latest perf v7

This issue doesn't occur when testing a headless client on windows connecting to a dedicated server on linux.

Interesting, sqwishy, that means that it's likely the -client parameter in Linux that makes it time out.

Oddly enough, I also don't get this issue doing the reverse; a headless client on linux connecting to a dedicated server on windows.

Maybe it's a problem with isolation? I'm curious what will happen with a linux everything on two different machines, virtual machines/containers, or different users on the same machine?

I tried running everything from a separate folder alltogether to see if that changed things, but unfortunately it didn't. Setting up a VM seemed counterproductive given ARMA's ill nature toward VM:s performancewise.

I tried with the arma 3 server running in a docker container and the headless client on the host and the headless client still froze up. I don't have the resources to try running each on separate linux machines. But that isn't really a viable workaround anyway, and neither is using visualization as you pointed out.

Can confirm. Tested on debian 8.x and ubuntu 14.04 LTS, both either crashes with signature verification timeout or just fails to verify signatures (BIS A3 ones, not mods) just at start.

Related
http://feedback.arma3.com/view.php?id=22914

Encountered the same issue. I was able to get the headless client to connect by making the VM single core. Obviously, not an acceptable solution. I suspect a race condition exists, probably in the handshake.

Additional details: Both the server and the HC are running on a Ubuntu 15.10 VM (running under VMware). I was able to get the HC to connect and FuMS to run correctly, however both processes were running on a single core. When I increased VMs core, the issue manifested.

After additional testing, I was able to get Linux HC to work on a separate (Ubuntu 15.10 on VMware) single core/cpu VM. I was not able to get it to work on the same VM as the server (2 core/cpu VM), even when I used taskset to run both processes on the same core/cpu. It appears that there is a race condition specific to the Linux HC, when running on multi-core/cpu machines.

This is still broken.

Linux is a 2nd class citizen with these guys. They can't even get the Linux client updated from 1.42, they've teased us with talk a 1.52 version for a few months, but we're not seeing any dev branches updates to help them with testing. Now they have 1.54 version on Windows and no timeline for Linux Battleye support, so a Linux client is pretty much useless for multiplayer.

I get the same problem with ubuntu 14.04 LTS.

Would it not be nice and kind of a solution to allow a IP list that is excluded from signature checks?

BTW did you guys tried the -cpuCount=1 param, for me its looks like its fixing the problem, I know its not great but at least it will work until a fix.
Both HC and Server are on the same Server for me.

never mind after 2,5 hours in the hc get the time out.

Player headlessclient: Signature check timed out

Had it run for about 4 hours with -cpucount=1 without any problems so far. Nevertheless I hope that this will be fixed soon.

same issue here, i am running server and headless client on the same machine, both in seperated chroot environments.

my command line for hc uses -cpucount=1 and i connect with global ip adress not with the local 127.0.0.1
im not loading any mod at hc or server.

i ve scripted an auto reconnect for the hc so that it connects 1 minute after disconnection.
in my logs i can see that it is disconnected after nearly exact 2 and a half hour every time.
if i watch cpu consumption of the hc process then I can see it raising within those 2.5 h from 15-20% at the beginning to 100% at the end of that time.
that was testet with an altis domination mission, mission was started but no one was playing.

BIS, please, do something about this in 1.60.
It's been more than a year without HC :(
Can we just have an option for skipping signature verification for HCs (like you can allow file patching only for HCs)? This will not solve a problem, of course, but at least it will allow us to use HC.

BIS, if you can't fix this issue right away, maybe you can add "skip signature verification for HCs" option, same as you have for file patching for HC?

This will at least allow us to play with acceptable FPS.

cpucount fix doesn't work anymore.

Please give us something to work with here BI.

Hello, still no solutions? BI is reading this?

thanks in advance

This is fixed at least with 1.70 but I guess some versions before.
could be closed...