Server & Timer Suggestions

[liberty 0 / 325]

I didn't want to bump a very old thread, so I formed a new one as I have a couple of suggestions, which should be easy enough.

 

1: Timelimit - the server-side mp_timelimit is set to 60 minutes whereas the TL SM plugin is set to 30 minutes per map. This is a conflict as you can't ever rely on the actual map time, and have to use timeleft in chat to get an accurate timing. A simple change to the server variables, setting them to 30, will fix this. Additionally, I'm in GFL quite often and surfers are generally seeking to extend a map -- I'd recommend possibly bumping the timelimit up to 45 minutes per map. 

2: TL plugin - previously stated 45 minutes per map, changing the map decision from 10 minutes, down to 5 minutes left on the map. This is a common practice among a majority of servers out there.

3: "SurfTimer" - the SurfTimer plugin is opensource, therefore, you're able to change and recompile anything about the timer. I recommend poking around in the "phrases" text file, and changing the SurfTimer | name to GFL, making it more customized to the server for your checkpoints, and general chat-based timer messages. Right now, it feels as if someone took the timer straight out of the box, slapped it on the server, and that's what we got.

4: Database - the server frequently has database issues because I'm imagining the amount of players on the server, this also causes it to become congested, and the server crashes soon after. Unclear as to the specifics right now with the database & server, but I'd recommend trying: getting a small VPS to act as the mysql server alone & transferring the entire database over to it. If that's the current setup, perhaps try utilizing a localized mysql server, on the same server, that it can read from locally. There's a few things to try, and I'm sure it can get resolved quickly to help stabilize the server.

5: Rank Coloring: Perhaps look into the coloring of ranks a bit more, and altering the colors. Currently, as of one I know of personally, the [Pro] rank coloring seems to be either a default color which defaults to the team you're on, or it's a bland orange color. Either way, the ranking colors need a bit of a refresher. I like putting in the time on a surf server to get a decent to higher rank, and one of the benefits of it is the cool ranks with a prominent color that lets others know. 

 

There's quite a bit more of suggestions I have, and I suppose it's from my years of running a surf server back in 2014-2016. I'd like to see more basic customization minimally on the servers if possible, it'll put some shine on things and give it a refresher, keep regulars around.

[Ash 177 / 3,624]

While I'm not a fan of raising any time limits, I'll comment on a few points:

 

3) Since starting work on the server some weeks ago, I've since forked the existing fluffyst fork of the timer and have been working on the more pressing back-end issues, to include more critical exploits. Cosmetic changes may happen eventually, but they're exceptionally low on the priority list, given other problems.

 

4) Unfortunately, a VPS isn't a fix in this situation, nor is it going to solve the root of the issue. While a MySQL environment running one of the 5.X.X builds would allow for a query cache, this would be more of a band-aid fix rather than solving the root of the problems - i.e. relying on a single connection to the database to deal with unoptimized queries. Not to mention, the query cache no longer exists past and including 8.X.X. I've altered some specific queries which resulted in a reduction of single-user load times from roughly 18-25 seconds to roughly 1-3 seconds on maps with high completions and high stages. There is a space to fix this here first and foremost. Crashes are unrelated to database performance in this case - The stock fluffyst has memory leak issues, which I'm also working to delve through. At this stage, I may ultimately re-write it for a V3.0. It's looking like this will be more and more necessary for the eventual upgrade to SM 1.11 anyhow.

 

Lastly, our Surf Timer suggestions thread is pinned close to the top of the list - feel free to make use of that! It helps keep things more organized so I don't miss any suggestions.

 

Edited March 6, 2020 by AshAkiri

[liberty 0 / 325]

@AshAkiri

 

Although styling and customization isn't a priority, it's the frontend of what the community is experiencing, and it takes no time at all to give it a refresher as I've done it with my server in the past many times -- shouldn't be no issue. It's similarly said if a website isn't using a styling sheet, but instead, the development team is working on things that we don't see (can be said vise-versa, but in this instance, the server is working for the most part).

 

It's appreciated that someone is taking on the task of the SurfTimer, or, a fork of it at least, to fix some common issues that have been around for a number of years since ckSurf, so I do commend anyone on that task.

 

In regards to the databasing issue, it's worth a shot to weigh in all options. I didn't have a community as large as GFL in Counter-Strike, however, my server "the cave" had seen a little over 40,000 records in the span of a couple months on surf_beginner (this was during the days of Opium Gaming, and beginnings of sneaK's community). With that being noted, I had a dedicated VPS handling the mysql database, and it worked absolutely flawlessly, so that was a reason I suggested it, perhaps it might be something to look into. 

 

In regards to the time limits -- I'd say do a community poll, and advertise it on Discord/in-game, to get a general discussion and decision about it from the community. I feel that as an actual individual that surfs in the server regularly, maps are generally vote extended whenever a user with access is on. 

 

Ultimately, I really want to see GFL's presence in the surfing community expand, I hope you don't take these suggestions the wrong way 🙂

[Ash 177 / 3,624]

I'll be honest - with the amount of work I've been doing on the back-end to optimize the timer, this does come across as patronizing. Especially all of the highlighted messages. It doesn't come across as being helpful. It did, however, put a bad taste in my mouth at the end of a long work day.

 

Without giving a full rundown of everything I've done to test the database (which I've been working on, for a number hours, from several different angles), I did spin up a DO droplet to see if it would make any meaningful impact on load performance. It did not. I did however identify all of the software issues with the timer that could be better ways to fix a problem for all folks using the timer that don't require spending resources.

 

For transparency's sake, I've got about 25 commits for fixes made to the existing timer codebase so far. Because the original developer has made it clear that he isn't processing merge requests, I'm working to get things cleaned to eventually publish this fork to the CS:GO surf community at large. With a significant amount of timer functionality currently using deprecated syntax which will cease to function when SM goes from 1.10 to 1.11, there is no shortage of work that I'm quickly trying to get done to ensure everything is compliant with the new syntax.

 

This, coupled with work I've done to compile approximately 600 maps not currently on the server and organize them for zoning so that our community's biggest complaint - a lack of maps - is resolved, takes up a majority of my afternoon. I work in development from 7-3, then come home and work on this.

 

I'm one person, just trying to tackle the fires from biggest to smallest.

 

That being said, my parting thoughts to this thread are - consider how things that are written come across. For someone who has no prior history in the forums to come in and make multiple assumptions about work being done, things being tested, etc. and then - in a way that comes across as extremely patronizing - to start providing advice on how to perform basic  debugging, you generally won't get a warm reception by folks working on it day in and out. It doesn't seem helpful at all. Most of the regulars on the server know me well enough to reach out and ask questions. Maybe do that next time. I like to think I'm a pretty chill person.

 

Edited March 6, 2020 by AshAkiri

[liberty 0 / 325]

They're merely suggestions, don't be too flustered over some simple suggestions, friend.

 

I don't doubt the work being put into the timer, nor making assumptions -- I'm just inquiring about it

 

However, from partaking in zipcore Timer, and ckSurf during 1.15-1.17, I know how simplistic certain modifications can be to truly give us, the frontend users, a little jive with some welcoming implementations. That isn't to be confused with backend work, as those are things we don't see, nor know about unless it's publicly released somewhere (github?). 

 

Are all of these commits and such you're doing public work? I don't see anything published on your github (github.com/AshAkiri), I'd be curious to know the details and perhaps help out in areas needed.