Server Issues? Questions? Tips for 4.33.1?

[Brain]

I agree with Cobalt, 500 should be solid for clients. 4 players basically in one bubble around the FLOT didn't seem to cause any issues...even with moving mud and tons of artillery fire going on.
Anybody still missing objects regularly (airfields technically only consist of many objects btw) should do a speed test. A bad connection to the host can't be improved by changing numbers in BMS. It sure sucks for those affected but we can't do anything about it unless we spend millions on improving the network infrastructure in certain parts of the world...needless to say not really an option.

Broken AI seems to be a totally different, server side issue that occurs over time. I didn't keep track but can somebody confirm crashes happen more often on weekends? If so the total amount of TEs being played is suspicious. There must be a reason why FO doen't allow hosting of any TEs.

[CobaltUK]

Just now had a server crash. Do I remember correctly ? that the client connected ports as displayed in the chat screen are the bms ports as forwarded ? Noticed a player whose connected port showed as 6625 though displaying the P2P `ok` connection. I do not think that's relevant to the crash as he was there for at least 20 mins prior.

It was very soon after another client joined that it crashed but nothing to say that was the cause.

[=VG= SemlerPDX]

restarted.

[Brain]

I broke it!

Just as I'm talking about things are fixed I'm missing a whole airport. In this case Onjin Airbase. I got shot down and after watching the ACMI I think the launches look valid, however I could visually see the launchers but not the missiles/smoke itself. When I exited the sim the server was gone.

Crap, can I get a hold of that crashlog maybe? :(

[=VG= SemlerPDX]

Restarted - file transfer from the server takes a long time, I will grab the logs during some server downtime in the near future. The problems are Memory related for sure, and the the Page Faults are often near 100 billion just prior to CTD events. The server runs an Intel Core 2 Duo and 4GB of RAM (I assume around 800MHz)

[CobaltUK]

The quite small crash text (<50k) might have some indications

[Derk]

As I researched specs to build my current pc this past June, I took specs from many of the regular users on the BMS forums to gauge what I wanted to build. (user specs were posted at the bottom of their posts). I formulated that I would go with a minimum of 8 GB RAM, and this was for BMS 4.32. Many users listed 16 GB RAM. I was successful running 4.32 with a system that had 4 GB, but now with the steeper system requirements of 4.33, and what I see as a much busier server, is only 4 GB of RAM on the server enough? Wouldn't the page faults on the server be an indication of a lack of RAM?

As for missing objects and other similar anomalies, I've read post after post recommending 1024 BW connections during MP Campaigns during heavy use. (500 may work during slower times). Seems like quite a jump from 4.32 using only 300, but 4.33 is obviously passing more data back and forth in MP than before. This is tiresome reading, but the thread at BMS Forums is here >>>> http://www.bmsforum.org/forum/showthread.php?23781-Multiplayer-Campaign-Issues

There is also info about 'dubious connections' that I don't understand totally, but has to do with those without correct ports set as described in the BMS Manual. (those with "CS" code behind their MP callsign instead of 'P2P'). These type users, and others with bad connections, can cause a number of object and AI issues, and possible host crash. I have seen this on the VG server recently and can be prevented with configuration settings to limit it. Info here >>> http://www.bmsforum.org/forum/showthread.php?24392-Multiplayer-Question-(P2P-CS-differences)&highlight=P2P

So with the combination of server crashes, AI issues, and lost objects, are we experiencing multiple issues at the same time with BMS and the server? Has a VG thread been started with BMS Forums to explain these issues to see if the Devs can shed some light or point in a direction? Perhaps a non-forums PM to someone on the Dev team?

[CobaltUK]

Ive started to notice some players who do not display the `correct` 2935 connected port though they do display P2P. How significant this is I do not know. Most players DO display 2935 including myself (ive asked others to tell me).

[=VG= SemlerPDX]

Server has been updated with the following changes:

-Client Bandwidth Requirement raised to 1024 BW

-Server will now only accept P2P connections, if you can no longer connect, read about port forwarding for Falcon BMS here: http://www.krauselabs.net/dump/Hosting-Falcon-4-BMS.html (and here is a link about routers and forwarding: http://portforward.com/english/routers/port_forwarding/routerindex.htm )

-Server will now enforce Same Tile Set and Same Aircraft Data

-Server set to display a 16 bit forced display, not sure if it will work in helping the server load from graphical demand


Server restarted to day 1

[=VG= SemlerPDX]

The server performed better over the last two days, I'd like to hear if there is more stability in-game.... It was sitting at the Mission Debrief Screen this morning and was restarted to Day 1, anyone notice any funky issues since Tuesday?

[Brain]

Doesn't being in the pit count toward flight hours? If so you might want to take note before each restart to see if it's a timed thing. Still thinking it has something to do with the flight being cancelled and bumping the server back. No idea how to prevent it though.
You are still using the same old day 1 save so you should be getting the same flight in the same conditions as well.
By the way: does BMS write any kind of server log?

On the other hand I was wondering what happened when we use an auto save after restart. It would throw off testing conditions but we know already that AI goes nuts after a while. Would they behave normal (at least for a while) after loading or is it the campaign that breaks?
Don't know what else everybody is up to but I'm running out of ideas what to do. Would need some coordination so somebody can hop straight in and see how things look.

[Derk]

I've been unavailable this week to fly, but plan to be in this weekend to see how the changes are working. As for the VG host exiting to 2D, I found some information on a BMS posting here...... http://www.bmsforum.org/forum/showthread.php?14032-Server-reverts-to-2d-map-on-it-s-own&highlight=server+reverts

Two points in this post is making sure host acft is in Startup with a cold cockpit, and a config setting for the "bump" time. I would offer to test the campaign file your using in offline mode on my pc to see if it gives the same behavior of exiting 2d.

[=VG= SemlerPDX]

Well, then it's cuz someone is jumping in the server plane and trying to fire it up? then leaving?

Moort wrote ..."the server a/c has to remain dark and cold cockpit otherwise it'll get kicked back to the gui after idling for some time on the ground."

!dntknw



*EDIT: to make sure these points are covered, all three atc.ini files in the Falcon BMS 4.33 files have PlayerBumpTime @ 1000000 now. This morning the server had a CTD crash, but the last time was a crash to 2D menu, so let's see how it works now.

[=VG= SemlerPDX]

I'm impressed with this latest round of fixes, touch wood: I think we might have fixed it!!

Server should be on day 4 by now, day 3 at the least, and all is smooth - page faults are approaching 2 billion, but the rise has been very steady over the last several days, and not 2 billion faults in less than 24 hours - a vast improvement.

I'm hoping to see this Campaign run to it's finish or until players say there are no more good missions. Please let me know. Lost one of my workers on the remodel due to sickness (tis the season for cold and flu) but I'm hoping to have it finished before this weekend.

At least I have this laptop to keep up with things, but my PC and monitors are all covered by towels to keep out dust.

[CobaltUK]

Would it be possible at the next re start to up the challenge from Cadet to Ace or at least veteran ? Thanks.

A.I wingman issues still showing up. The ground circling, airborne fast rolling and another I am seeing now. When viewed in external view he is constantly changing power settings in distinct steps between 70 and full AB and back to 70. Perhaps 12 steps in all. This can both be heard and seen in the infobar. As with the other stuff it goes away once he is un tied from flight lead by a RTB command. Human wingmen do not warp or show other signs of connection issues so why the A.I do is odd.

[=VG= SemlerPDX]

I have time soon to find out if other servers have that AI issue or if it's just us.

Also, I was under the impression the difficulty was already set at ACE - if it is not, I will make sure it is. I need one of you to let me know when there are no more good flights here - this stability test succeeded and the server looks like it wants to keep running, so you all may run out of missions worth flying.

Let me know, normally I'd restart it on Day 5 or 6 regardless, but this time (with 4.33) I'd like to hear how far it should go, or if it is different that 4.32 campaigns.

Thanks again for all the helpful input!
!cheers

[Sinu]

It's very interesting when you look at the force levels page for the DPRK - at around day 4 they had 0 tactical aircraft and then it spiked to ~100 aircraft the same day...Same deal with the ground units. Watching the JSTAR replay is pretty laughable.

Our ground forces have failed to punch through the DPRK lines.

As far as difficulty - I've always found Veteran to be just right for operations, but again that's just my opinion.

I have been very impressed with the stability of the server over the week or so it has been up. Great job Semler!

[=VG= SemlerPDX]

Sinu LAN_WROTE ...

It's very interesting when you look at the force levels page for the DPRK - at around day 4 they had 0 tactical aircraft and then it spiked to ~100 aircraft the same day...Same deal with the ground units. Watching the JSTAR replay is pretty laughable.

Our ground forces have failed to punch through the DPRK lines.

As far as difficulty - I've always found Veteran to be just right for operations, but again that's just my opinion.

I have been very impressed with the stability of the server over the week or so it has been up. Great job Semler!

So, then, the big question is: It's day 5 or 6, are there any good missions still available, or do we need to restart the campaign? If it's stable, and if the DPRK still have some butt that needs kicking (and it's still fun) then let's keep going to see how long it takes in 4.33 - otherwise, let me know today and I will restart it to the beginning.

Since this final test succeeded, the next round will involve auto-save and restarts will load from the day it crashed. We will evaluate how well the server recovers from a loaded campaign save file. But I don't want to start that until we know how long a campaign will want to run (with fun/good available missions) in this 4.33 version.

!hi

[CobaltUK]

Anyone else noticed day after day of low cloud despite the briefing saying otherwise ?

I think 6-7 days max allows for those that want to see their efforts compounded to good effect and those wanting more AA action back. Possibly moving from cadet to veteran/ace will prolong the AA action ?

I don't expect it to happen but even if only a 1 day test I would like to see the 433 Israel theatre run because I feel it is a much ligter load on a server (could well be wrong in that) and if there were a total absence of issues on a busy w/e day that just might mean something.

[Brain]

There is no Sinai for 4.33 yet and although the old version seems to work (according to some YT footage) there is no guarantee a campaign will run stable. It's also no longer available on FO. Balkans is on it's way however and might see a release "in 2-4 weeks" (as usual). I'm not sure if Sinai would be lighter on the server anyways. The new KTO seems to have more and larger battalions compared to .32, making it more on par with the FO theaters.

For the weather, I believe it only addresses the cloud plain you usually get between angels 25 and 30. Could be wrong though as I never did pay that much attention to it ... yes it's stupid but I still carry on with LGBs even if I can't even see the end of the runway :D

The AI (apart from its buggy behavior) is damn stupid as far as I can tell. They do know a few new tricks (radar snoozed, GCI, ambushing...nasty) but I can't tell you how often I dragged all kinds of MiGs into friendly AA. Rarely will they fire a missile even with full AA loadout and perfect firing position. Also they don't seem to care about any defensive maneuvers making BVR kills VERY easy. I suggested reviewing the difficulty setting some time ago but TBH I didn't mind testing out the new toys around the FLOT and using the same save for testing made sense anyways. I have to admit though flying directly over an SA-2 site and not getting shot at is getting old :)

Resupply does happen according to the active war production on each side. It has something to do with generated supply points that get spent after a set amount of time ("time to next resupply") but I still don't fully understand how it works. The amount of available aircraft might be calculated with their airfield's status in mind. In some cases a single repaired runway could mean 40 aircraft are suddenly operational again and there are quiet some engineer battalions driving around.
...
Whatever might be the exact case I don't mind the resupply at all as it keeps the airspace busy for both sides and the campaign interesting. We have to see how it plays out in the long run and if it turns out there is no air resistance at some point we should even consider increasing the DPRK airforce level. Depending on the actual change it makes and how the campaign commander reacts to it it may increase the difficulty early but keeps things interesting later...we'll see, probably thinking too far ahead.

With so many squadrons NOT set to HQ it's mostly up to us planning and executing missions to help the war effort. Infrastructure, supply, supply routes, morale ... even if not immediately apparent it all does make a difference and helps pushing north eventually.
Just to give you one example: taking out the bridges north of Seoul (Ryes?ng River?) forces DPRK to take a long detour, causing exhaustion and lowering supply. The result is a large push to the east about 12h later but some CAS will lower their numbers and morale, forcing a retreat or at least giving friendly troops a better chance of repelling the attack (or retreating if necessary). In the end it delays the DPRK assault on Seoul by a few days and with the campaign's main objectives in mind I feel this is highly rewarding...and it's just one plane doing 4 missions up to this point!
I was also very happy when I saw a HMMWV battalion (towed guns I believe) retreat after I repelled a heavy tank push pinning them down.

We are not simply pilots, we are also making a decision every time we select our objectives. Sometimes it might not seem we're making much progress because we are just looking at the icons on the map, not the status of every individual battalion or the big picture in general. It's all about working towards a long time goal with smaller victories in between. Restarting just wipes the entire progress made so far even if it seems like nothing has changed since day 1...we are also missing the surprises that may come in later...

That's what makes BMS so special and we should take full advantage of it. Restarting the same scenario (Day 1, 5:00) "every few hours" is what DCS servers do and doesn't come close in terms of long time motivation.

If everything is stable I'm all for using saves at least until I have to move my FLOT line considerably to the north ;)

Damn ... just wanted to make a simple point and then ... be glad I don't have a collapsible spoiler BBtag or it would have been worse! :D Well, hopefully it was as encouraging as intended. I promise I'll shut up now and get airborne again these days ^^

[CobaltUK]

Are we talking about the same Israel theatre ?
http://www.bmsforum.org/forum/showthread.php?23621-Israeli-Theater-for-BMS-4-33-v1-0


Regarding bridges I USED to give them some priority but when you see vehicles crossing in the water even before bridges are hit there doesn't seem too much benefit.

[Sinu]

Server appears to have crashed today around 15:00 EST.

[=VG= SemlerPDX]

restarted. Will be creating a new save file soon with a higher difficulty setting and other tweaks now that we know it's more stable.

[=VG= SemlerPDX]

Okay... Server Save File rebuilt now, have to leave town today, this weekend will be a difficulty test - auto-save will begin first thing next week. For now, I need input on the difficulty level - it was indeed set very low this entire time - don't get caught off guard now, it's at ACE.

Let's see if the enemy Mig's will fire off a shot or try to evade a BVR attack. Will be back in around 12 hours, will check in then.
!hi

[Sinu]

I will say my first impression is that the allied side is almost defenseless against this new onslaught of enemy aircraft (although this may just be day 1 effects). I have enjoyed the increased threat level of aircraft and SAMs. The airspace just became one of the deadliest I have ever seen.