When I connect to the North America server using the same node, the ping can be either 189ms or 235ms. In order for me to get the lower ping, I have to leave and create a new game. Why is that so?
Weird… Could you please confirm that it’s also same with Full VPN mode enabled?
Do not use a version higher than 5.4.0. More details in this thread.
The new version 5.4.5 does not fix the current problem, it only fixes the ping display in the web interface, but the game still pings with a large ping. For me the most stable version is 5.4.0.
@buzarini Did you check ‘Status → Process’ menu to see “Core” log section when your issue happened? Please check there’s an error or warning message there.
There are no mistakes. The funny thing is that on version 2.1.26 for openwrt 19.07.7, I get a high ping, and on version 2.1.25 there are no problems with high ping.
Okay. It seems I found a reason. When I checked your account, it seems your data plan is Subscription. As you might guess, Subscription data plan has the speed cap (4 Mbps burst / 1 Mbps).
When I checked the code review, it seems we found a bug that the speed cap didn’t work properly in the previous version and patched it around several months ago. I think recent releases include that patch.
Please note that nowadays the speed cap is implemented at client side of mudfish side not to overload the mudfish nodes. So to avoid this issue, you need to use old versions of mudfish if you don’t want to encounter this issue. However please note that we’d implemented a server side speed cap feature now and we have a plan to move the speed cap implementation from client side to server side as soon as possible because this week point could be abused by other users. So at some days, using old mudfish versions would not work.
Please measure and compare the speed with the latest or the previous version. I think you can find that this issue is caused by a bug I mentioned above.
This is very bad. And then what to do? Switch to a tariff with traffic or look for a better alternative?
I hope you understand our decisions. Three workarounds at this moment in my opinion:
- Changes your data plan from Subscription to PPT (Pay-Per-Traffic).
- Or at end of this month or beginning of next month, our Subscription data plan will be changed. At that moment mudfish will support 4 different subscription types with different speed caps. For details, please check Policy change for Data Plan - 20210607 link.
- Or if you’re not happy with mudfish, you can look for other alternatives.
If I use the SOCKS proxy method, then I get much less latency than in the official mudfish client.
SOCKS - High ping drops after a minute. Subsequent downloads are five seconds.
mudfish 5.4.5 - High ping drops after five minutes or more. Subsequent downloads for five minutes or more.
Very big difference.
@buzarini This is a known issue because the speed cap was working based on the per-connection based rather than the per-user based. For example, if a user makes 8 connections to the mudfish node via SOCKS proxy, its limit of speed cap could be jumped 8 times.
Unfortunately this bug was also addressed with latest mudfish server package and deployed/tested with mudfish PBE (testing) nodes. As you might guess, the following mudfish nodes are patched so you can test against these nodes if you want:
- MD Europe (Chisinau - MivoCloud) - PBE
- RO Europe (Romania - Virtono) - PBE
- VN Asia (Vietnam - vHost 2) - PBE
So all speed cap issues will be addressed when next mudfish server version is upgraded. I expect it’ll happen at beginning of next month.