The thing is I know a decent bit about coding applications and games, but I know nothing about networking. But, from what I do know, I can’t think of what would cause the refund item to be executed more than once. What kind of “lag” would cause such a thing? It certainly must be an internal issue within their own code, as in how the have implemented their system/logging etc in their networking layer
They paid out the conditional award first before invalidating the pre condition.
The pre condition in this case would be the refund button press state? or having put the item into a TP offer?
With patch 1.8 they have reworked the system where it could come to canceled transactions /buy /sell. Presumably they have built in some kind of buffer or retry or something similar. I think that exactly the system under the “lag” conditions then acted incorrectly.
I’m not entirely certain what their pre condition is in this case.
Logging does not cause performance issues. Only question is storage, and storage is cheap. Usually 30 days of logs are kept
Sorry, lol.
Have you ever activated a debug logging on a software?
1 Like
I work on a distributed system with debug logs constantly on that generates millions of logs per minute
So yes. Perhaps you are not using the right software for handling your logs. If you ever feel you can’t write a lot statement because it will slow something down then you have a bad logging solution.
New Dupe of Trophy mats and high teir items from players cancelling buy orders in an error state
Disabled Gold Transfer allowed Town Owners to create unlimited gold at the moment out of thin air, same as last Christmas event! I love when they bring back old features!
I work for a large ERP company. Normal logging, in our case audit, is easily feasible without performance problems, even if several gigabytes of logs are generated per minute. But when I enable debug logging that records and logs absolutely every little thing, logs are written that sometimes write over 100GB per minute.
Since logging is not just queued up in the back, it can slow down the process.
Yes, modern storage systems can handle that, but AGS will never use such hardware for a game and therefore keep logging at a moderate level.
Interesting. Our system logging is in fact queued up in the back. Specifically the app only had to write to the console stream which is scraped by a side process, v aggregated into samples, forwarded to another cluster for processing, reduced for query where possible on several partitions.
Now on a desktop client I could see that being different, but in this case the living would be server side.
1 Like
Why wouldn’t Amazon use a decent system considering they’re world leaders when it comes to Cloud computing? You’re talking out of your rear end here… logs are not taxing, if you’re not dealing with the output throughput in an appropriate manner then sure, you’re getting yourself into bottlenecks, but that’s more of a reflection of the poor implementation you have rather than the simple fact “logging slows everything down”.
It happened day 1 on Fresh Start Servers. There was a player on my server selling stacks of 10.000 Eggs and everyone asked where he got them from, and he said he had a Chicken Farm, and everyone asked where it was.
Oh well, I guess there never was a chicken farm.
This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.