Please for the love of god, restart the servers weekly
Please for the love of god, read up on modern infrastructure and try to understand that ārestarting the serverā isnāt a thing for this.
1 Like
where do people come up with this nonesense?
Years of Microsoft Windows.
5 Likes
While i agree that netcode and latency nonsense needs to be worked out I wish the solution was as simple as the OP suggests.
1 Like
Either something was implemented wrong, the maintenance is being done incorrectly or theyāre skimping out on the hosting plans, or something unfixable keeps repeatedly cropping up that might be because of the aforementioned reasons.
Why do people immediately latch onto the specific words someone uses instead of responding to the message theyāre trying to convey?
Obviously OP is complaining about poor performance, so why not talk about that instead of berating them for not knowing how modern service hosting works?
9 Likes
Maybe because words convey meanings and we canāt know when one person did not mean what they said.
Also the OPās tone warranted a similarly toned answer.
5 Likes
As a system admin, servers should be patched regularly for cyber security reasons, which often necessitate a restart. IF you need to restart for performance reasons you have a memory leak or something WRONG with the application. At that point we (me) brings this to the vendor as a ticket.
Now, load balancers and high availability is a thing. The technology isnāt that hard. Use Citrix net scalers, AWS ALB, whatever, and just drain user sessions for the targeted servers, And when itās empty you reboot. Just ensure you have enough capacity when you drain a quarter, third, or half, and ensure itās done in low usage days. For my organization thatās Wednesdays 2:00am. But you can automate that. Donāt need to be awake for it.
All this to sayā¦ Yes restart your servers. Is darktide performance issues related to server reboots? Maybe. But if it is, itās just a symptom, not the cause.
8 Likes
Kernel releases are 10 weeks apart. You must have a pretty severe patch to apply if you restart more often than that.
I thought the āserversā were owned by AWS, and itās just temp Instances of the game world (map) that spin up and then get dropped when the mission completes?
Sure I remember a post in the past where someone here pointed to their (AWS) games instance servers, found the ips, and identified they were on the cheaper side of what could be bought.
Iād imagine AWS would be fairly sharp on server best practice, even if the Instances are less impressive.
*edit Gamelift I think?
1 Like
I use AWS for work, it depends on their setup but if you use EC2 instances AWS is just another hypervisor. You manage the OS and everything else higher on the OSI. Maybe AWS has some other products where you only manage the application layer for gaming.
I found the thread when back at keyboard:
by @ggg
Thereās a process that attempts to connect the user to the gamelift host/cluster with the lowest latency. Didnāt catfish or someone say a while ago that there was an issue where that process was sometimes placing users on distant hosts?
To be fair, I donāt think itās the lack of server maintenance. Itās just a lot of entities to handle at the same time, which is what Horde games struggle generally with.
2 Likes
Use some critical thinking. Why do you think theyād want the āservers restartedā?
Bull. No one here has any reason to have taken offense at OPās tone.
1 Like
I donāt want to put more oil on this fire.
So I will only say this: have a nice weekend! 
1 Like
If your game canāt be run well on your servers, perhaps you should look into some optimization, cull the number of entities on screen at a time, or invest in better servers?
Thatās a cop out answer.
1 Like
Iām not an expert, but I doubt itās as simple as weekly restarts. However server stability, hit reg, and etc ARE a problem that they need to address and which will probably get worse with Xbox crossplay.
1 Like