Anyone else experiencing long page load times?

[fasontales]

Specifically page load times, the website itself pops right up lightning quick. This has only started for me about a week ago and it's happening every day sporadically.

Just want to know if anyone else is experiencing this, thanks!

 

[piapple]

I've been having the same problem too! I thought it was just my bad wifi, which it still might be the case, but sometimes the site just gives me a 504 error and other times it takes awhile for a page or two to load. Now this also might just be to my terrible wifi at times, but does your account sometimes not mark chapters as read?

 

[53r4phPh4r35]

I thought it was my connection but it's been a few days since sometimes pages take a lot of time to charge when they are charging because sometimes it's just don't charge at all.

 

[Lynkaster]

feels like it's recurring after 8pm everyday on my side (paris time)... the home page and every chapters are having a hard time to load, requiring quite a lot of refresh to get a page or two correctly loaded (the @home function is already deactivated so it's something else)

 

[Holo]

Then it's a ddos-guard issue, which is out of our hands.

 

[gr8g]

I'm not able to read any manga older than a day 😥. I thought the recent announcement of MD@home would solve the issue, but no luck so far. Able to read only latest chapters of any manga, not able to binge read anything!

SEA country btw.

 

[gr8g]

@AbyssalMonkey
Woah, didnt realize I could switch it off, I see a significant improvement already. Although, now it loops back to the original issue of the thread, where most images are loaded, and few don't.

Yeah, I was avoiding aggregators, and going to the scanlators websites, thankfully many new Manga's scanlators have readers. But for many of the old mangas, which now I have the time to read, the group/site doesnt exist. I hope MD tries to solve these issues, and maintain their position as one of the best scanlation sites.

 

[Lynkaster]

reading on mangak*lot since the support won't be able to do a thing here 🙄

 

[Bogo]

god todays lag is crazy, cant even load main page without doing 5 refreshes

edit: its a bit better without MD@Home

 

[spectrewolf28]

Yeah i've been getting alot of issues too with loading one or two pages, which can sometimes be really annoying with long pages, sometimes even after finishing loading it only loads half the image.

 

[Bogo]

everything is back to normal with MD@Home turned on, tbh dont know where bottleneck was

 

[BraveDude8]

@AbyssalMonkey Requests are not routed to offline nodes, for obvious reasons, and the network currently successfully serves >99.9% of requests. Nodes with excessively slow loads have their requests decreased until they start performing correctly.

If you're failing half the time, something else is wrong. Are you browsing the site from a restrictive network, i.e. public wireless/workplace/school?

 

[BraveDude8]

@AbyssalMonkey An option to limit the nodes you use to ones running on port 443 is being added.

Again, the nodes are self-reporting and self-regulating. Bad nodes have their requests reduced until they approve, awful nodes have them paused entirely.

>If people are loading just fine without it

If the network wasn't working perfectly fine for the vast majority of people, you wouldn't be "just fine" without it, because our image servers would be maxed out the vast majority of the time. If you have further issues, grab the subdomain of the node you're connecting to and tell me what it is so I can see if it's slipping through the net.

 

[Holo]

Our 2 image archive cache servers are currently going at 25% capacity, mostly feeding the clients, whereas before MD@ H, they were going at something like 150% of capacity, so the network is definitely working for the vast majority. We just have to figure out why for a minority of people it doesn't work. Perhaps an option for a restriction to :443 ports will further reduce problems.

 

[BraveDude8]

@AbyssalMonkey As I said, the nodes self-report failures. They just don't do it in the same kind of detail a user can.

The system is "pushed on users" because it's currently maintaining a 99.95% success rate, compared to the vast majority of the chapters being effectively unusable.

If the node fails, it does failover to the image server. If that isn't working for you, it's the kind of edge case where someone being available to troubleshoot would be nice.

If you're failing after a reload, then you're failing to load images from multiple nodes and if the nodes themselves were having issues you wouldn't be requesting images from them in the first place.

Considering less than 1/2000 chapters are failing at the moment, it's far more likely there's an issue on your end. Hardly the rate you'd expect from "bad decision-making" or a system with "blatant issues ignored". The system is the default because it is far more reliable than not using it, and the system being the default gives us enough overhead for people to be able to opt out. It's still a work in progress, but it's providing the average user with a far better experience than the site would without it.