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.
Hosthatch http and https not work?
Hello Guys,
I have somehow since a few days problems on my VPS at Hosthatch in Amsterdam Netherlands from the blackfriday action on http or Https URLs to access.
It does not affect all URLs but only a few e.g. : wieistmeineip.de or Speedtest.net or iplocation.net
Can someone check if he also has the problem ?
E.g. when I do wget speedtest.net it should download a HTML page but it waits all the time for the https request. With other server providers it works. I use Firefox on my server and it loads and loads but nothing comes only a timeout.
Comments
Maybe reinstall? This sounds like a software problem not an issue particularly of your unmanaged hosthatch vps.
Could you share logs of your commands maybe? Without any information what is going wrong it's hard to tell..
And please use verbose logging!
No it isn't an software problem. Because wget is not an external software. Yes firefox of course but the problem appear since a few days I think
I just tried
wget speedtest.net
and can confirm that it runs into a timeout on my Hosthatch AMS VPS. A traceroute towards speedtest.net looks fine though, so I suspect that they're throttling/blocking certain IPs/Networks on their end.wget -v speedtest.net
--2021-04-20 11:59:42-- http://speedtest.net/
Resolving speedtest.net (speedtest.net)... 151.101.130.219, 151.101.66.219, 151.101.194.219, ...
Connecting to speedtest.net (speedtest.net)|151.101.130.219|:80... connected.
HTTP request sent, awaiting response... 301 Moved Permanently
Location: https://www.speedtest.net/ [following]
--2021-04-20 11:59:42-- https://www.speedtest.net/
Resolving www.speedtest.net (www.speedtest.net)... 151.101.130.219, 151.101.66.219, 151.101.2.219, ...
Connecting to www.speedtest.net (www.speedtest.net)|151.101.130.219|:443... connected.
@brueggus thank you very much for conformation. I wait and wait but connecting timeout
Before the Downtime yesterday everythink works. I think
@brueggus No it is on the site of hosthatch because I have as well an VPN to my server and there speedtest.net works.
Same error on my storage VM. However, nvme loads fine
@dosai I think it must be an firewall Issue on hosthatch server. That port 80 and 443 are blocked somehow.
But the strange think is why I can access to speedtest.net with mv phone over vpn to hosthatch ? Or with my notebook over vpn to hosthatch.
My guess is the MTU is too big due to some jumbo frames thing on the new router. Try reduce the MTU, e.g.
ip link set dev eth0 mtu 1400
, to see if it works.Can you explain that ?
Read previous message
On my AMS vps ever since last night they moved to a new rack had network issue, I think @tetech is correct with the mtu causing it, I changed the default mtu 1500 to 1450 all the network issue I noticed like random timeouts disappeared.
We need a no MTU=9000 hall of shame.
@HostHatch will not be on this list, but one of their upstream providers will be.
I want the entire Internet to support jumbo frames!
When you run TCP, the TCP sender should have Path MTU Discovery.
Your interface MTU could be set to a high value, but PMTU discovery is supposed to find out the proper packet size over the path, and then align TCP segment size accordingly.
If PMTU discovery isn't working properly, you should check your firewall to see if ICMP "fragmentation needed" messages are allowed.
This message type is needed for PMTU discovery to work.
As usually, tcpdump and Wireshark are your friends.
wget and ping logs are of limited usefulness.
@Hosthatch @Emil can you give any info about whether the changes in AMS would affect PMTU or discovery? Thanks.
I confirm the problem. Can't run bench.monster on AMS storage VPS anymore
We're making some changes in Amsterdam (we sent out the maintenance emails about this to everyone affected).
This should be fixed shortly.
I confirm I received one email per VPS, so the downtime should not be much of a surprise. Still down, though.
Amsterdam seems to be moved to a new upstream. Traffic to my home ISP is routed through Germany now . NFOrce had direct peering.
Thoughts @hosthatch ? Routing optimalisation in the making?
Please open a ticket in that case
The issue in question here (in the OP) was fixed a few hours ago.
Maybe they are busy with preparing more storage offers as promised last week.
Will do, but had to check here first.
Sounds like bandwidth costing a fortune and they needed to block speedtests until on the new provider?
No, we never 'blocked' anything in particular.
We are waiting for Core-backbone to accept some of our prefixes - but not sure what range you are on and whether that has anything to do with this. We can look into it once you have a ticket in and get you a resolution
Mentioned that in my ticket too. Ingress goes through Core Backbone and stays in NL.
Yes, the MTU thing does seem to be fixed - no need to manually lower MTU now.