New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
★ VirMach ★ RYZEN ★ NVMe ★★ $8.88/YR- 384MB ★★ $21.85/YR- 2.5GB ★ Instant ★ Japan Pre-order ★ & More
This discussion has been closed.
Comments
I'm using my vps for openVPN it's located in LA USA, why is My county Germany when I'm connected to the VPN?
I want to cancel my service, but no refund because I have to use PayPal, and to be able to use PayPal, you have to add funds to store credit, which is non refundable. Not everyone uses credit cards. Wish I would have thought this through before hand. Data center advertise on LA USA, buy it's really not. Wasted money that is no use to me.
Ordered vps in LA USA, it's in Germany. Check the ip. 149.57.129.61 How can this be?
Update your IP data, this is la ..................
Learn how IP location software works, how IP registration works and how IP leasing works before accusing someone of "fraud".
Did you check routes and connections? What were the results? That's more of a sign of where the location of your VPS is, as looking up with tools these days. We're living in 2022, not in 1993 when there were plenty IPs...
Yeah, My mistake for jumping the gun. Still it's unusable to me if it truly reside in another country. Thanks for pointing it out.
Hey @melp57 how many tickets do you have open already? xD
--
None, because they never get answers.
Yeah, I understand you. I also get very pissed at the local lottery here since I never win a price. I also never participate, but let's forget that, it's the thought that counts.
There's always One in the crowd...
Why was my order cancelled and no ticket reply? I have waited since March 12 and this is the result? I used a combination payment, how will the account balance be refunded? @VirMach
Well you still believe clients abusing your network instead of side effect caused by UDP accelerate software. Do some research on KCPTUN and Hysteria, these are widely popular used in PRC to improve network quality. The side effect is every VPS in the VLAN can receive others' broadcast, the packets increase exponentially when the VLAN is big. A big VLAN works fine only if no one abuse or uses these kind of UDP accelerate software, this is impossible since these tools invented. The way to solve is isolate each trunk using PVLAN (Setup on switch/router) or isolate each VPS using OVS (Setup on Node system). Anyway glad to see you noticed and trying to solve this issue, good luck.
They wont give you a refund, I suppose.
They wont even give you a look because you bought the cheap servers. Good luck:)
P.S. I also bought cheap servers so I am also experiencing similar things
or instaban those MJJ abusers.
Tokyo and San Jose haven't been and will never be stable with the influx of MJJ abusers.
or isolate them on separate node. this should fix majority of the problems.
insert checkbox "are you going to use this for jumping firewall"
I feel extremely angry and disappointed for not addressing their huge problems and closing my tickets bluntly without any explanations.
My server is down over fourth consecutive weeks, and it’s still being locked, and no details was ever tried to be disclosed, and two promised timeline was overdue without any actions, and my one ticket created by yesterday was closed without any response!
If it's in Phoenix then it's already Ryzen. It should still work for now but try again tomorrow message just means whatever you're selecting is most likely full.
If there's any updates on anything moving forward I'll provide it here or somewhere else. > @shkong said:
Based on the time of your post I'm guessing you were one of the multi-account people or previous chargeback people (last one is more rare than the first) or the rarest potential option of recent suspicious activity on your account like a lot of failed login attempts (account considered potentially multi-user or compromised.)
If you have a ticket open about it or opened one about it then it will just be closed without a response if you receive a refund already but of course if it's about an issue with the refund then you can contact us.
Yeah, we also reached out to xTom to see if they have a network engineer that can assist us for those locations since they set up the initial switch configuration for us and seem to have talented people on the team.
I noticed the same thing you're describing after diving into it deeper and when we first had the problems I already began setting up newer nodes with divided subnets.
The main issue ends up being when we try to isolate them with ebtables, everything goes haywire and it causes network issues. But then without ebtables, there's more abuse we have to deal with manually. So yeah we need to either just split everything up as we had it in the past or deploy some other kind of solution, which we can do in some locations such as Tokyo that have newer switches with layer 2 capabilities, as long as I don't have to be the one that reads the 400 page manual.
Is this huge problem you are explaining the inability to migrate portion of it or the locked node? The migration doesn't work because it'll cause problems if you migrate from a broken node. As for the broken node, we're still working on it with the limited time we have but as long as people keep creating tickets about it we're going to focus on answering the tickets first right now or going through them and closing them because we don't want to keep falling behind on tickets.
If your ticket is closed without a response then that means time saved for actually fixing it.
Take it easy bro, Hysteria will increase 50% network traffic, it's still a reasonable usage. And I have to remind you it's highly possible that a increase usage in one node may lead to a discrease useage in another, so technically the network traffic will not significant increase in total number.
Super glad some measures are underway. Besides the udp multiplexing tools mentioned, there's also this so called bbr-plus that intended to boost tcp throughput for networks with high packet loss, only end up making the VM extremely unstable by replacing a stable kernel with a ancient, modified one. The sad part is it can't actually even compete with vannila bbr congestion control shipped with modern stable kernels.
I have no idea how you can isolate these but maybe keep an eye on VMs with some odd janky kernels would help.
Ticket #323941
When will it be solved?？？？？？？？？？？？？？？？？？？？？？？？
Hello, San Jose-002 has been offline for more than half a month. This has a great impact on me. Can you stop selling if it is not repaired? This is too bad！
WTF is BBR Plus? Is there a BBR Pro or BBR Max?
Did some searches and it seems that both bbr plus and kcptun mentioned above returns results exclusively in Chinese, barring irrelevant ones like Boston building resources.
Are these MJJ's new silver bullets to fuck with service stability? both UDP and TCP, whether via UDP abuse or extreme TCP congestion control?
Can Virmach simply instaban abusers and put all proxy users on a dedicated node so that if they want to disrupt service stability just let them have fun with each other.
edit:
It's even funnier that kcptun GitHub page has English, Japanese and Korean descriptions but no Chinese like who dafaq needs this in English/Japanese/Korean speaking countries not even NK cuz people there have access to their bright light intranet only, and Japanese description is understandable yet so unnatural and laughable.
The "fun with eachother" node idea backfired as you may remember and also assuming it doesn't just end up offending everyone, all it will do is increase ticket load as we constantly get complaints from probably everyone on that node in the form of 3 to 10 tickets per person per day. One way it could maybe work is if we exclusively migrated all the people making a lot of tickets to the same node and then restricted ticket acces
As for the level of "abuse" this could potentially cause, assuming these numbers are not inflated based on our sub-optimal setup when it comes to things like this, each user still only uses a non-abusive level of networking. It's only when it's combined with 100 people doing the same thing on the same node where it becomes collectively problematic. The nodes could probably handle maybe 30 people doing this, not 100. So 30 people can be using it for this while everyone else uses it for other purposes and it'd be fine, and this is what we're seeing in pretty much most locations that are not desirable for it, as well as nodes that didn't heavily fill with low cost specials.
Offline services that have been created improperly. I'm currently going through all of these, independent from tickets, and fixing/recreating them. Doing Tokyo first. Then Los Angeles, then NYC, then Amsterdam, then all other locations. During this I'll exclusively be going through those that don't have tickets open. If a ticket is open we'll process those afterward once we get to the ticket. I'd include those but then if we just close them to handle it there will be a lot of angry tickets asking why their ticket is closed, etc.
Tokyo recreations complete. Around 50-100 were not actually broken, they just had an old template or old non-existent ISO mounted, or still on CentOS 5.8, in rescue mode, or left in a state of disarray by the customer by sending like 30 button requests at once. 2 of these are broken still because too many buttons sent but the disk image exists, will have to fix and troubleshoot later.
20 were actually broken in Tokyo, no proper disk image.
Around 5 had tickets open, skipped. 15 didn't have active tickets.
5 of the 15 already recreated using Ryzen Migrate button.
10 recreated, and due date reset, welcome email re-sent.
How did you check the IP? If you try < https://www.iplocation.net/ip-lookup > you'll see that for most geolocation services it shows a US location, but the first entry incorrectly shows Germany, probably because it hasn't been updated for a month. So VirMach has done nothing wrong.
Perhaps next time, instead of crying on the forums that you've been defrauded, that you want you money back and embarrassing yourself, just ask the question of how it could be possible that your server is reported in Germany, and wait for help.
Is the upload speed of LAXA008 limited?
I did a few data(2G) transfer tests:
Transfer from LAXA008 to LAXA010, about 30MB/s
Transfer from LAXA010 to LAXA008, about 90MB/s
Los Angeles recreations complete. Most of these just needed ISO unmounted or booted. 2 had a port issue, fixed. 0 of them required recreation.