SQL Host Hardware Upgrades

[Water]

On 1/24/2024 at 5:53 AM, PeregrineFalcon said:

How would you know how many people were logged onto the retail servers? I don't remember Cryptic or Paragon ever showing those numbers.

 

Did you work for either company? If not then how do you know what the numbers were?

 

One could extrapolate a rough idea of player numbers based this historical account.

https://cohtitan.com/forum/index.php?PHPSESSID=a755lb64pkb52q9mbbc9nbe345&topic=5309.0;nowap

 

 

[Ridiculous Girl]

1 minute ago, Number Six said:

 

Bigger server go faster.

 

[Farrellite]

6 minutes ago, Ridiculous Girl said:

can someone explain this to me like i am a five year old? 😄

Probably the wrong person to do this, as my expertise is in performant cloud infrastructure (not specifically, but often based around SQL & SAP HANA) and they're on physical analogues to my boxes in the sky, but here I go:

Big monster provides playground for it's fleas on it's back.  But to power playground, Big Monster has to eat a lot. However, Big Monster cannot always find the best things to eat, and sometimes the food gives Big Monster an upset stomach.  Lately, a lot more fleas are playing on Big Monsters playground on his back, so Big Monster has been having to eat more and more, and sometimes runs out of food, giving the playground hiccups or if Big Monster is just too overwhelmed by it all, crashes entirely.   Sometimes, Big Monster has to sleep (maintenance). 

[Number Six]

On 1/30/2024 at 12:03 PM, Hyperstrike said:

But look at how much memory gets held onto (Properties/Memory - Minimum Server Memory)

Don't worry, we have a couple people on the infra/admin team who do enterprise SQL server tuning for their day job. We're very well aware that it will try to consume all available system memory for page cache.  😉

 

But the reason it's so critically important to run SQL on a dedicated host separate from dbserver and/or other game components is exactly so that it can make use of all that RAM without having to artificially limit it to a much lower cache size to accommodate a variable amount of memory demand from other processes.

 

That and dbserver is extremely sensitive to latency. It can't handle a stall if SQL is monopolizing resources and something gets paged out or has to wait on a kernel mutex.

[Water]

12 minutes ago, Ridiculous Girl said:

can someone explain this to me like i am a five year old? 😄

 

Yes, I'll tell you later...

 

[Jordisk]

39 minutes ago, Number Six said:

 

Global services like the auction house, chat server, etc. work by the service itself connecting to each shard, and if one of the shards are down it tries to reconnect every few seconds. But it does this synchronously, freezing the entire process, and causes lag on the other shards if they are waiting for a reply from the service.

 

Also accountserver is not the most well written piece of software and instead tends to get really crashy if shards are down.

And I assume editing the list of shards that it checks is a hassle, so you can't just remove the two that are down and then add them back afterwards.

[Number Six]

46 minutes ago, Jordisk said:

And I assume editing the list of shards that it checks is a hassle, so you can't just remove the two that are down and then add them back afterwards.

 

We considered that but it would mean a second restart for everything else, which is maybe marginally better but also annoying for everyone.

[Lockely]

3 minutes ago, Number Six said:

 

We considered that but it would mean a second restart for everything else, which is maybe marginally better but also annoying for everyone.

 

Sometimes games have to go down for extended maintenance, especially MMOs. It's cool. At least the beta servers are up!

[victusfate]

I'd be happy to contribute to a hardware fundraiser for whatever servers you folks think is best. I suspect I'm not alone.

 

I need to set an alarm for that  Saturday of the month when donations occur. I kinda wish we could donate anytime.

[Hyperstrike]

2 hours ago, Story Archer said:

 

Gotta listen to a man who'll use 'ameliorate' in casual conversation. 🙂

No you don't.  I'm a well read spelling bee champ.

And there's always the chance I could be an idiot.

[Hyperstrike]

1 hour ago, Number Six said:

Don't worry, we have a couple people on the infra/admin team who do enterprise SQL server tuning for their day job. We're very well aware that it will try to consume all available system memory for page cache.  😉

 

But the reason it's so critically important to run SQL on a dedicated host separate from dbserver and/or other game components is exactly so that it can make use of all that RAM without having to artificially limit it to a much lower cache size to accommodate a variable amount of memory demand from other processes.

 

That and dbserver is extremely sensitive to latency. It can't handle a stall if SQL is monopolizing resources and something gets paged out or has to wait on a kernel mutex.

Say no more.

I don't know enough about the way the machines are tuned to do more than just really basic advice.

[Hyperstrike]

38 minutes ago, Lockely said:

 

Sometimes games have to go down for extended maintenance, especially MMOs. It's cool. At least the beta servers are up!

Better extended downtime than half-ass "solutions".

[Mr. Infinitum]

Personally, I can't speak to actual hard cap numbers, but I only came across Forced to Queue as a new login the last night of CoX OG.

I was in a mothership raid and crashed, which would happen from shear effects overrun in memory, all the bla bla bla, differential coloring of costume bits and powers etc...

I specifically remember trying to log back in because, it would take me like 2 or 3 mins tops...and we were about 5 mins into the fight on top of the mother ship.

Playing Mr. Infinitum, my first very first character created, on Infinity, and when tried to log back in I was over 500k back in the Queue?!?!?! like 513thousandsemodd?!?!!?

I ended up going to the Incarnate Server for the rest of the evening I couldn't get back into Infinity... 

 

[Hyperstrike]

2 hours ago, Ridiculous Girl said:

can someone explain this to me like i am a five year old? 😄

Also known as "Throw hardware at it"

Like swapping in an LS7 engine into a Volkswagen Beetle.

[Mr. Infinitum]

5 minutes ago, Mr. Infinitum said:

Personally, I can't speak to actual hard cap numbers, but I only came across Forced to Queue as a new login the last night of CoX OG.

I was in a mothership raid and crashed, which would happen from shear effects overrun in memory, all the bla bla bla, differential coloring of costume bits and powers etc...

I specifically remember trying to log back in because, it would take me like 2 or 3 mins tops...and we were about 5 mins into the fight on top of the mother ship.

Playing Mr. Infinitum, my first very first character created, on Infinity, and when tried to log back in I was over 500k back in the Queue?!?!?! like 513thousandsemodd?!?!!?

I ended up going to the Incarnate Server for the rest of the evening I couldn't get back into Infinity... 

 

Maybe again this is flawed memory and it was mor like 5131 back in queue, but never seeing a queue there personally? =/

[TorontoSunshine]

Ex-celsior!!!!

(sure, like some of you haven't said that already!)

[TorontoSunshine]

24 minutes ago, victusfate said:

I'd be happy to contribute to a hardware fundraiser for whatever servers you folks think is best. I suspect I'm not alone.

 

I need to set an alarm for that  Saturday of the month when donations occur. I kinda wish we could donate anytime.

I have a reminder for that on the last Saturday of every month on my app. 

I wish we can donate any time as well. 

 

[ZacKing]

2 hours ago, Number Six said:

Bigger server go faster.

 

Internet win!

[Bionic_Flea]

2 hours ago, Number Six said:

 

Bigger server go faster.

Hey look!  A ball!   *rolls ball at Ridiculous Girl*

[Kaoticken]

33 minutes ago, Hyperstrike said:

Also known as "Throw hardware at it"

Like swapping in an LS7 engine into a Volkswagen Beetle.

Ooohh - Ork'n it up sum are we... gonna give it more Dakka?

[Hyperstrike]

22 minutes ago, Kaoticken said:

Ooohh - Ork'n it up sum are we... gonna give it more Dakka?

No, because that intimates we're blowing things up.
NOT something we want to do with an expensive server.

[elfeanor]

Thanks a lot for this detailed information.

[heyedwin]

I told ChatGPT (3.5 free version), "Explain SQL Queue in layman's terms as if I was a kid."

 

The following was the answer:

 

Alright, imagine playing with building blocks. The SQL Queue is like a line of tasks that our computer needs to do. Most tasks are quick and happen at the same time, like adding small blocks to our tower. But sometimes, there are big tasks that take a bit longer, like adding a really large block.

Now, imagine there's a rule that if someone says "wait," we all have to stop playing until they finish what they're doing. In the computer world, these "waits" are like barriers in the SQL Queue. When a big task happens, everything else has to pause until it's done.

Upgrading the computer's brain (hardware) should help it handle these big tasks better, like getting a bigger table for our blocks. But sometimes, if we have too many big tasks, we might need an even bigger table or find smarter ways to play with our blocks.
 

Is the explanation correct?

Edited February 1 by heyedwin
Spelling error(s).

[Focron]

That license deal popped the Genie out of the bottle. I'm pretty sure the flood will continue now that "ancient" COHers hear the news. I confess I am one of them. Running it like a dream in Linux too. You gurus do what you have to do to get it running smooth, we have faith in you. If I had more than two thumbs they would all be sticking up. 😉

 

 

[Zombra]

Forgive a ground level question, but does the server selection screen automatically update when server status changes, or will I need to refresh it?  I'm leaving the client open on a 2nd monitor so I can see when we're back online 🙂