Jump to content
Dante Unbound: Share Bug Reports and Feedback Here! ×

Dedicated Conclave Servers


[DE]Drew

Recommended Posts

On 2017-08-08 at 8:51 AM, Echoa said:

Hey guys, just trying to make a decision here.

 

My question is does having an SSD improve the server performance or is it just fine in a HDD? I'm trying to decide if I'm going to get 6gb more ram so I can dedicate more ram or get an SSD. Right now I'm running it entirely headless with 

Xeon x5650 (3 cores per server (3 servers))

12gb ram (2.8gb per server)

320gb 7200rpm HDD

 

Only things the server does are warframe and it's my storage server (only 2 reasons I have it). Idk how the warframe dedicated servers are effected by drive speed. I'm not having any issues/performance problems but anything I can do to improve user experience would be nice. 

I use my main server with 7200RPM hdd no problem with that and even running ubuntu host and windows guest in a vm and it is not giving any perf issues

Link to comment
Share on other sites

4 hours ago, maciejs said:

I'd say it's not worth it only for Warframe DS.It'd only affect your loading times, but since:

a) we're not loading gfx data

b) it only happens once, when hosting a mission for the first time

I'd guess the perf improvement would be very modest (avatar stuff would load faster too, for JIP, but again - this is not a deal breaker).

ok, if its not going to be any significant improvement i wont waste my time. Im not running into any issues performance wise was just curious if it would help, thanks 

 

Ill look at grabbing some more ram then so i can start up a 4th VM

Link to comment
Share on other sites

On 31/7/2017 at 3:36 PM, maciejs said:

Yes, I did some initial investigation and it seems doable. Can't give you any hard dates, but I'll try to get something working soon-ish.

Ok, thank you @maciejs!

That is good news though.:thumbup:

i do really hope that you guys also start considering supporting Linux, so it can be hosted in a more professional manner,. 

Link to comment
Share on other sites

4 hours ago, maciejs said:

As you might have noticed in the patch notes, the 'alliance only' mode is in. See the 'advanced guide' on how to enable it -

(TLDR: put allianceOnly=1 in your settings)

Sweet, thank you so much @maciejs

Any recommendations for a VPS server setup for running a large alliance on it? 

Thanks! 

Link to comment
Share on other sites

I've been working on tools for hosting for a while now. As it's quite stable at the moment (no changes since U21), i might as well share it here now. It's still a work in progress, and might not be as easy to set up as i want it to be (there is an app included to do the required setup). Source code Soon (TM).

https://github.com/Spiedie/WarframeServerManager

@dreamquestwhen choosing VPS, make sure it's not IP banned by DE. I know AWS EC2 is banned. Estimated requirements from the my readme:

  • ~500-700 MB ram per server instance + OS ram requirements
  • ~1 core per server instance
  • ~20GB disk space per 6 instances you want to run

If your server is running windows, you don't need any VMs, making ram requirements a lot easier to manage (using 9GB ram for 20 instances on windows 8.1).

 

@maciejs "Free for all" is kinda confusing in the selection, as Annihilation is commonly referred to as Free for all (FFA). I think "Public" would make things less confusing.

 

Link to comment
Share on other sites

On 8/13/2017 at 0:25 PM, spiedie said:

I've been working on tools for hosting for a while now. As it's quite stable at the moment (no changes since U21), i might as well share it here now. It's still a work in progress, and might not be as easy to set up as i want it to be (there is an app included to do the required setup). Source code Soon (TM).

https://github.com/Spiedie/WarframeServerManager

@dreamquestwhen choosing VPS, make sure it's not IP banned by DE. I know AWS EC2 is banned. Estimated requirements from the my readme:

  • ~500-700 MB ram per server instance + OS ram requirements
  • ~1 core per server instance
  • ~20GB disk space per 6 instances you want to run

If your server is running windows, you don't need any VMs, making ram requirements a lot easier to manage (using 9GB ram for 20 instances on windows 8.1).

 

@maciejs "Free for all" is kinda confusing in the selection, as Annihilation is commonly referred to as Free for all (FFA). I think "Public" would make things less confusing.

 

How many instances have you been able to run at once with full speed? If I could condense down to one VM on my server that'd be great but I don't want to hurt user experience.

Link to comment
Share on other sites

On 8/19/2017 at 10:31 PM, Echoa said:

How many instances have you been able to run at once with full speed? If I could condense down to one VM on my server that'd be great but I don't want to hurt user experience.

I run up to 24 instances at 120Hz. At 240Hz i started having issues with the gpu (server hardware integrated, just meant to produce output, its really bad) when running 18 or more instances. Note that this is a dual socket E5-2670 (v1) system, so 16 cores and 32 threads. For comparison, most modern quad-core systems can run 6 instances at 60Hz.

Link to comment
Share on other sites

26 minutes ago, spiedie said:

I run up to 24 instances at 120Hz. At 240Hz i started having issues with the gpu (server hardware integrated, just meant to produce output, its really bad) when running 18 or more instances. Note that this is a dual socket E5-2670 (v1) system, so 16 cores and 32 threads. For comparison, most modern quad-core systems can run 6 instances at 60Hz.

What do you mean GPU issues? I'm running my server entirely headless and SSH into it with a fake display in xorg. Ill give it a go I'm running a bit older hardware (x5650) but have 12 threads to play with.

Link to comment
Share on other sites

13 hours ago, Echoa said:

What do you mean GPU issues?

The 18+ instance windows changing their title (the fps counter) 240 times a second appears to be too much for the gpu to handle. You likely wont ever get this issue tho. Minimizing the windows also fixes this.

Link to comment
Share on other sites

13 hours ago, spiedie said:

The 18+ instance windows changing their title (the fps counter) 240 times a second appears to be too much for the gpu to handle. You likely wont ever get this issue tho. Minimizing the windows also fixes this.

Just want to say thanks alot for the tool Spiedie it is working great

The tool let me condense my VMs down to one virtual machine with 4c/8t and 8gb of ram. All 6 servers running 60fps and while i set it up switched to VirtIO on my VMs to speed things up. Now i just gotta nab an SSD (which now i only need a 128gb) and ill be golden. 

Really appreciate the work

 

Link to comment
Share on other sites

Another question I have for the folks with better network knowledge than me. I have a 5up/45down connection. I pay to have no data cap (got the line specifically for hosting these servers) so data used isn't an issue. My question is should I use 60hz or 30hz for my 6 servers or will I be ok on my upload bandwidth at 60hz? I want to make sure all servers can be used at the same time on my 5mb upload speed.

Link to comment
Share on other sites

On 2017-08-22 at 9:49 AM, Echoa said:

Another question I have for the folks with better network knowledge than me. I have a 5up/45down connection. I pay to have no data cap (got the line specifically for hosting these servers) so data used isn't an issue. My question is should I use 60hz or 30hz for my 6 servers or will I be ok on my upload bandwidth at 60hz? I want to make sure all servers can be used at the same time on my 5mb upload speed.

They do not use much of download speed though if you plan on hosting more then lets say 6 servers than you start to run into issues I guess with your up-speed.

I have myself a 40up/300down connection and that handles it fine so far with 24 instances fully used (PVP event).

 

Nep_Blanc

Link to comment
Share on other sites

For reference, 24 instances at default (60fps, highBandwidth=0) during quick steel event (6 players per instance) was about 12 Mbit/s upload. I think 8 players was 600kbit/s per instance (this was during snowball events).

Total data during last snowball event, 4 day weekend (close to new years eve?) was 300GB uploaded and 120GB downloaded.

Link to comment
Share on other sites

Yes, it's been a while since I measured, but last time I checked, 1 client needs roughly 10kb/s (with some decent action going around, can probably peak higher than this, though). So 8 clients would be around 600kbit as @spiedie mentioned. That's with high-bandwidth=0, 60hz will be more than this (not exactly twice, but more), so let's say 1Mbit/s. That means 5Mbit up for 6 servers would be cutting it close (assuming 6 full games going on). I'd recommend 30 (or a mix of 30 & 60).

Link to comment
Share on other sites

Alright thank you for the info, that's what I figured. I knew down bandwidth wasn't an issue as they probably wouldn't use much but my up being 5mb I figured 60hz on all might saturate it. Unfortunately don't have higher up options without spending a relative fortune in comparison lol. I moved them all to 30hz just to be safe.

Link to comment
Share on other sites

  • 2 weeks later...

I've been having a few issues loading instances lately. Sometimes they freeze as seen in the first spoiler (last part of the log, unedited including the times at the end). This results in following instances to start and freeze like in the second spoiler (full log, freezes after "Found 26 cache blocks..."). Not only does freezing like this cause an instance to be inaccessible, but if more are launched after, it takes down all 5 other instances running on the engine (all players in those 5 instances disconnect and host migrate, see third spoiler).

Is it possible to have these loading actions time out at some point? afaik, disk activity is 0%, so it seems to be a locking issue. Having these instances close themselves would already solve a lot of the issues.

 

Spoiler

3.045 Sys [Info]: Loaded TOC for Windows/H.Misc_xx (took 8.83ms, loaded 6795 entries, found 0 deleted, spot-loaded)

3.053 Sys [Info]: Loaded TOC for Windows/B.Misc_xx (took 5.82ms, loaded 6795 entries, found 0 deleted, spot-loaded)
3.136 Sys [Info]: Budget overrun fetching cache results (78.1 ms)
3.217 Sys [Info]: Loaded TOC for Windows/H.Misc (took 73.9ms, loaded 75696 entries, found 596 deleted)
3.269 Sys [Info]: Loaded TOC for Windows/B.Misc (took 46.7ms, loaded 48609 entries, found 84 deleted)
3.273 Sys [Info]: Loaded TOC for Windows/H.Font (took 1.93ms, loaded 1712 entries, found 172 deleted)
3.277 Sys [Info]: Loaded TOC for Windows/B.Font (took 1.92ms, loaded 1689 entries, found 76 deleted)
3.284 Sys [Info]: Loaded TOC for Windows/H.AnimRetarget (took 3.49ms, loaded 3103 entries, found 130 deleted)
3.288 Sys [Info]: Loaded TOC for Windows/B.AnimRetarget (took 2.75ms, loaded 2759 entries, found 72 deleted)
3.290 Sys [Info]: Loaded TOC for Windows/H.CharacterCodesCache (took 0.013ms, loaded 1 entries, found 2 deleted)
3.292 Sys [Info]: Loaded TOC for Windows/B.CharacterCodesCache (took 0.005ms, loaded 1 entries, found 0 deleted)
3.300 Sys [Info]: Loaded TOC for Windows/F.Misc (took 5.73ms, loaded 2914 entries, found 2 deleted)
3.302 Sys [Info]: Loaded TOC for Windows/H.Misc_en (took 0.132ms, loaded 53 entries, found 6 deleted)
3.304 Sys [Info]: Loaded TOC for Windows/B.Misc_en (took 0.097ms, loaded 53 entries, found 9 deleted)
3.306 Sys [Info]: Loaded TOC for Windows/F.Misc_en (took 0.058ms, loaded 29 entries, found 0 deleted)
3.308 Sys [Info]: Loaded TOC for Dx9/H.Misc (took 0.016ms, loaded 1 entries, found 0 deleted)
3.311 Sys [Info]: Loaded TOC for Dx9/B.Misc (took 0.083ms, loaded 1 entries, found 0 deleted)
3.315 Sys [Info]: Loaded TOC for Dx9/H.CharacterCodesCache (took 0.349ms, loaded 214 entries, found 92 deleted)
3.316 Sys [Info]: Loaded TOC for Dx9/B.CharacterCodesCache (took 0.083ms, loaded 72 entries, found 6 deleted)
3.323 Sys [Info]: Loaded TOC for Dx9/H.ShaderPermutation (took 4.46ms, loaded 4538 entries, found 2000 deleted)
3.335 Sys [Info]: Loaded TOC for Dx9/B.ShaderPermutation (took 10.5ms, loaded 4538 entries, found 745 deleted)
3.373 Sys [Info]: Loaded TOC for Dx9/H.Texture (took 34.3ms, loaded 38279 entries, found 102 deleted)
3.401 Sys [Info]: Loaded TOC for Dx9/B.Texture (took 22.4ms, loaded 36043 entries, found 392 deleted)
3.426 Sys [Info]: Loaded TOC for Dx9/F.Texture (took 19.1ms, loaded 28888 entries, found 317 deleted)
3.427 Sys [Info]: Loaded TOC for Dx9/H.Texture_en (took 0.061ms, loaded 23 entries, found 0 deleted)
3.430 Sys [Info]: Loaded TOC for Dx9/B.Texture_en (took 0.143ms, loaded 23 entries, found 12 deleted)
7120.383 Sys [Info]: Budget overrun fetching cache results (3.81 ms)
15806.951 Sys [Info]: Budget overrun fetching cache results (3.10 ms)

 

Spoiler

0.034 Sys [Diag]: Process Command-line: -fullscreen:0 -dx10:0 -dx11:0 -threadedworker:1 -cluster:public -language:en -allowmultiple -log:DedicatedServer13.log -applet:/Lotus/Types/Game/DedicatedServer /Lotus/Types/GameRules/DefaultDedicatedServerSettings -instance:13 -settings:SpTDMRC
0.034 Sys [Diag]: Build Label: 2017.08.31.11.45 Retail Windows x64 [Stripped]
0.035 Sys [Diag]: Build Unique ID: 3045033953
0.035 Sys [Diag]: Windows user-name: Spiedie
0.035 Sys [Diag]: Windows computer-name: Warframe-Server
0.035 Sys [Diag]: On developer network: no
0.036 Sys [Diag]: Current time: Mon Sep  4 02:01:09 2017 [UTC: Mon Sep  4 00:01:09 2017]
0.037 Sys [Diag]: Current directory: C:\Users\Spiedie\Desktop\alloc0.2.0.2\WFAllocManagerLocal
0.047 Sys [Diag]: Processor: Intel 16-Core 16-Core/32-Thread Xeon E5-2670 0 2600MHz HT/MMX/SSE/SSE2/SSE3/SSSE3/SSE41/SSE42/AVX/POPCNT F/M/S/EF/EM: 6/D/7/0/2
0.048 Sys [Diag]: Physical Memory: 21GB / 32GB free
0.048 Sys [Diag]: Address Space: 128TB / 128TB free
0.048 Sys [Diag]: C:\pagefile.sys: 1,024MB of 1,024MB free
0.048 Sys [Diag]: Page size: 4KB (Granularity: 64KB)
0.048 Sys [Diag]: Operating System: Windows 8.1  [64-bit]
0.049 Sys [Diag]: System Processes: 67
0.049 Sys [Diag]: System Threads: 1,574
0.049 Sys [Diag]: System Handles: 22,024
0.050 Sys [Diag]: System GDI Objects: 230
0.050 Sys [Diag]: System USER Objects: 273
0.051 Sys [Diag]: System Up-Time: 0d 10h 56m 41s
0.051 Sys [Diag]: Free Disk Space on C: 88,730,673,152 bytes
0.052 Sys [Diag]: Free Disk Space on D: 2,338,859,171,840 bytes
0.052 Sys [Diag]: Found Monitor: (0, 0) - (1280, 1024) : 1280x1024
0.053 Sys [Diag]: Clock frequency: 14,318,180Hz
0.065 Sys [Info]: Shutting down content update...
0.070 Gfx [Info]: Private shader options need 35 bits
0.091 Sys [Info]: Loading packages took 0ms
0.091 Sys [Info]: Redirecting package load to: C:\Users\Spiedie\AppData\Local\Warframe\Editor.cfg
0.092 Sys [Info]: Windows caching enabled
0.093 Sys [Info]: Dx9 caching enabled
0.093 Sys [Info]: Dx11 caching enabled
0.093 Sys [Info]: PS4 caching disabled
0.094 Sys [Info]: XBONE caching disabled
0.094 Sys [Info]: Scanning D:\Program Files\Warframe3\Downloaded\Public\Cache.Windows\ for cache blocks...
0.095 Sys [Info]: Found 26 cache blocks...

 

Spoiler

15257.267 Sys [Info]: Budget overrun fetching cache results (4974 ms)
15257.267 Sys [Info]: Budget overrun retiring cache results (4974 ms)
15257.267 Sys [Info]: Budget overrun issuing download callbacks (4974 ms)
15257.268 Net [Info]: 5 (4) : need bps: 6335.14, bcs: WAIT, allocated: 81920, congest %: 8, throttle: 0.999, max bps=31129.6, slow start=0, bad client: 0, rtt: 107
15257.268 Net [Info]: Total upstream: 4429, max total upstream: 33762
15257.268 Net [Info]: 26 hi priority objects.
15257.268 Net [Info]: Sustained upstream: 14581, avg packet loss: 0.334247
15257.268 Net [Info]: 5: Pending reliable messages, channel 0: 62
15257.268 Net [Info]: 5: Retransmit throttle: 1.1, threshold: 1.1s
15257.268 Net [Info]: 5: Lost packets: 2, acked packets: 11620
15257.268 Net [Info]: 5: Ping: 125
15257.268 Net [Info]: 5: Queued reliable: 0
15257.268 Net [Info]: 5: Expected seq: 33296
15262.276 Sys [Info]: Budget overrun fetching cache results (4977 ms)
15262.276 Sys [Info]: Budget overrun retiring cache results (4977 ms)
15262.276 Sys [Info]: Budget overrun issuing download callbacks (4977 ms)
15262.277 Net [Info]: 5: Pending reliable messages, channel 0: 62
15262.277 Net [Info]: 5: Retransmit throttle: 1.1, threshold: 1.1s
15262.277 Net [Info]: 5: Lost packets: 2, acked packets: 11620
15262.277 Net [Info]: 5: Ping: 125
15262.277 Net [Info]: 5: Queued reliable: 0
15262.277 Net [Info]: 5: Expected seq: 33296
15267.283 Sys [Info]: Budget overrun fetching cache results (4974 ms)
15267.283 Sys [Info]: Budget overrun retiring cache results (4974 ms)
15267.283 Sys [Info]: Budget overrun issuing download callbacks (4974 ms)
15267.284 Net [Info]: 5: Pending reliable messages, channel 0: 62
15267.284 Net [Info]: 5: Retransmit throttle: 1.1, threshold: 1.1s
15267.284 Net [Info]: 5: Lost packets: 2, acked packets: 11620
15267.284 Net [Info]: 5: Ping: 125
15267.284 Net [Info]: 5: Queued reliable: 0
15267.284 Net [Info]: 5: Expected seq: 33296

 

Link to comment
Share on other sites

  • 2 months later...

Been having a crash on my TDM and DM servers

WAR-1507945

Quote

 

6908.918 Error [Info]: GPF

Application error messages:
HealthSpawner.lua
HealthSpawner
PVPDeathMatch.lua
RunMode
RequireWeaponEquipped.lua
StartMod
RadialBlindAbility.lua
ActivateAbility

 

 



DM just crashed with this one as well:

WAR-1507915

Quote

 

7288.266 Error [Info]: GPF

Application error messages:
PrimeVectis.lua
UpdateDiscDeco
ElevatingGlow.lua
ConstantGlowAsNoise
Kraken.lua
ActivateAbility
Kraken.lua
TentacleAnimation

 


Another

WAR-1509347

Quote

3095.431 Script [Info]: NULL SOURCE OBJECT AND TYPE IN DAMAGE TRACKING!!!

3099.448 Error [Info]: GPF

Application error messages:
DecoyAbility.lua
DecoyMonitor
BowEffects.lua
StringPullback
ShieldDownEffects.lua
AddScreenEffect
RadialBlindAbility.lua
ActivateAbility


WAR-1510003
WAR-1507943
WAR-1510093



As with spiedie this has been occurring all day / night, seemed to start after update 22.2.5

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...