Rhinox

My host is VMware vSphere ESXi (but VMware Workstation is reported to work too), and VM is Win7/6bit with 2 vCPU and 8GB vRAM. Not much difference compared to native HW. BTW this topic (DCS & virtual server) was discussed extensively here (but the thread is closed now): http://forums.eagle.ru/showthread.php?t=116021&highlight=esxi&page=3

That's strange. I'm running DCS on virtual machine, with no 3d-acceleration, using VNC for connection. Are you really sure you disabled 3d-rendering?

Yeah, that's what I found yesterday evening (~20:00 GMT): - one aerobatic server (~15 players) - plenty of passworded (various number of players) - one public (12of12 slots taken, don't remember which one) - plenty of empty servers Where did all the DCS-community go? Seems to me this sim is dying out. It is high time to make it more attractive!

Do not forget that mobile-CPU/GPU is not the same, as desktop-CPU/GPU! Even the best "gaming laptop" is about as good as mid-range PC... I just tested P34G/v2: quite impressive specs (i7-4710, gtx-860m, 8GB ram), but it was "just enough" for DCS. My year and half old PC (i5-2500k, GTX-670) is in DCS ~25% faster...

It is not (only) about brand, but technology too. Bigger cells (i.e. 25nm) are better than smaller (20nm), and single level cells (SLC) is better than two-level cells (MLC) and both are better than three level cells (TLC).

Exactly. Apart from stability/quality and similar issues, one more topic is power distribution. 300W PSU can not "move" power freerly from 3.3V-branch to 5V or 12V. There are certain limits, and these are sometimes defined as "combined power" (+5V & +3.3). BTW, concerning PCI-express based cards you can check GPU-card design to estimate upper limit of power it can theoretically draw: from pcie 16x slot GPU can draw ~75W (iirc, 10W from 3.3V, rest from 12V). Then there are additional 6-pin power connectors (~75W) and/or 8-pin power connectors (~150W).

So I suppose you do not want to do some gaming on it, right? Then why not to buy some ready-made mini-PC? A friend of mine has "Gigabyte Brix". It has i5 (probably some low-voltage notebook version), slots for so-dimm and 1x mSATA. It has even vesa-adapter for mounting on back side of LCD. Looks very nice and does not take much space on the table...

No matter how it smells to you, it is sadly truth. I remember days of LOMAC, FC1 & FC2, when you could find much more on-line players at any time, than now at prime-time. There were big on-line events, with 50-60 (or even more) players, going on for a few hours (locers, GI/CI, etc). With todays DCS (in)stability you can only dream of it. And what's worse, I do not see any improvement. Might be wrong, but I think with every new version or module released it goes from bad to worse. And there is no end to it...

That's probably why I did not find anything in event-log. But I will try it again... Concerning my configuration: it is Windows7/64bit/SP1 with all critical/recommended patches installed. No antivir.

Rule number one: 1. If something works, do not touch it! Maybe ED-devs should read it a few times. Honestly, I have no idea why ED changed the way users connect to master-servers. There must be probably some reason for it, but by using "Schannel" we are now at mercy of Microsoft. If M$ screw something, we can only wait till it's fixed...

Nothing I can relate to this problem. The only trace is in dcs.log: 00009.584 ERROR NET: Can't connect to Master: SSL connect error 00009.584 ERROR NET: Master login : 503 00012.403 INFO NET: client has stopped Edit: I will check it again. Maybe logs are written in batches. But I remember immediatelly after MP-login failed I found nothing...

*IF* it were so, then no services relying on Schannel would work. But on my home LAN everything works without problem (sharing, remote connection, AD, etc) even without lsass.exe having free access to the net. DCS is different. If I allow lsass.exe outside connection, I can log in to MP. If I block it, I can not log in. I tried to grab the traffic, but it seems to be all encrypted (I could not find any readable strings). All I know up to now it does about 3-4kB traffic per MP-connection, all to Akamai-servers. But it is huge CDN, so it is hard to find what actually it does there... DCS has definitely something to do with lsass.exe, and I do not like the idea of opening my system. Lsass.exe was never meant to be fully opened to the wild! Very critical services depend on it...

Windows "firewall" does not filter system apps at all. For all system services everything is allowed. Thats what Microsoft calls "firewall"...

Neither-nor. And I described it quite clearly in my previous bug-report for 1.2.8-beta... BTW, you started to show colors, so there is no need to wait: yes, I found it myself. You changed the way users log in to "your" server. What I do not know is: are you aware of security implications you caused by this modification? Let me explain it in more details: Yes, you are now using standard OS-service called "Local Security Authentication Server" (Local Security Authority Process). But because of this, one more than generous firewall-rule must be active: For the process "lsass.exe" (which itself is frequently targeted by malware) free communication with any system on the internet must be granted (in addition to that granted to DCS.exe)! Or at least for those destination IP-addresses belonging to Akamai Content Delivery Network which you use (that's one or more B-class networks, if someone knows all them). And to make things even worse, this single particular process handles multiple sensitive services, i.e.: CNG Key Isolation Creditential Manager Encrypting File System Netlogon Protected Storage Security Accounts Manager Maybe I forgot something, but it does not matter, because what I found is more than enough to make me worry. Exposing any service having something to do with system-security and encryption/protection for free access to/from internet? And all of this becaues of DCS? With all do respect, the one in ED who came with this "idea" must be (hard self-censorship follows) complete madman! Any local process (even malware running in user-space!) can now communicate with internet using this service and what's even worse, any node on the internet can communicate with this local process (thus easily bypasing any firewall)! Hey, all you brute-force wanna-be hackers: welcome to my system, and play with it as much as you want! No, do not thank me. Thanks should go to Eagle Dynamics who forced us to open this door for you... Only one question remains: ED, did you force us to open this security-leak purposely? FYI: as soon as I opened firewall for lsass.exe, DCS works. But I'm not going to keep it that way!

Just FYI, I *think* I know where the problem is (at least in "my" case). Let's see how ED's support works, and give them a few working days to react. If nothing comes from ED, I will open the pandora-box...

What we have common is that 1.2.7 worked, and 1.2.8 does not. So I'm pretty sure it is something on ED-side because nothing has been changed on my PC, just DCS was updated...

I have most of DCS-modules active (iirc, all except P51), but I just tested my server (pure DCS-World, nothing more) with the same result: with 1.2.7 all normal, after update to 1.2.8 "Cant' connect to Master"...

I went back with restore-point to 1.2.7: everything works as expected. Then I updated again to 1.2.8 and the same problem again: Cant' connect to Master. Seriously: someone still thinking it is on my side? I think it's time ED spits out the truth and says what is different between 1.2.7 and 1.2.8! Apparently, there is something more than what's in changelog...

I suppose I'm not running it because I do not even know wtf "bit defender" is. But I know with 1.2.7 I did not have this problem. It all started with 1.2.8-beta. It is quite clear it has something to do with this particular version, not with my configuration...

Logs are THE SAME as with beta-version with the only suspicious entry: 00009.584 ERROR NET: Can't connect to Master: SSL connect error 00009.584 ERROR NET: Master login : 503 00012.403 INFO NET: client has stopped I just want to point out that everything worked without any problem with 1.2.7... Logs.zip

It is everything pretty much explained in the link I provided. I do not see any reason to repeat everything once more...

How should I stay calm??? I reported problem long time ago, no one from ED reacted! And now the very same problem is in 1.2.8! So why the hell you release beta-versions when you pay no attention to bugs reported??? Official (ehm, ehm, so called) "support"? I already reported tickets 5 or 6 times with the same "results": I waited 2-3 month, then made loud complaints here, and finally someone answered my problem here. I NEVER got effective support using your portal! Your "record" was 9 months "response time", after which ED just asked "does your problem still persist?". What the hell is such a "support" good for? It is pretty much like non-existent...

OK, it is all again here. I reported it for 1.2.8-beta, but no one from ED reacted, so now with update to 1.2.8 I CAN NOT USE DCS:WORLD AT ALL!!! WTF is this, ED? What did I pay for???

CPU: get "K-model" (unlocked multipier = easy overclocking, trust me, you'll need it) GPU: 2GB might be just enough for full-hd (with future "edge" engine being still question mark). If you want to try higher resolution (or multi-screen), go for GTX770/4GB. SSD: if you can, take 240GB-version (bigger cache, higher i/o-speed, higher iops, more space for over-provisioning, better price_for_capacity ratio)

IIRC TrackIR is auto-configured in DCS. There is no need to play with its configuration there (unless you want to use head-movement for i.e. throttle-control). Just set-up axis and button-controls in TrackIR-software, and you are done...