[X] My Assistant

Hello,

I discovered the H@H today, just applied last hour, my applicationis in pending.
But I've started the client to see what it is an a Client ID is asked.
When I read the GUI, that say that I need to register it at hentaiathome. net but this website didn't exist:
Not Found
The requested URL was not found on this server.

I think this will become available after my application will be approuve but I prefer ask before. Am I missing something ?

This post has been edited by gouzi: Oct 3 2022, 23:01

Hello,

I discovered the H@H today, just applied last hour, my applicationis in pending.
But I've started the client to see what it is an a Client ID is asked.
When I read the GUI, that say that I need to register it at hentaiathome. net but this website didn't exist:
Not Found
The requested URL was not found on this server.

I think this will become available after my application will be approuve but I prefer ask before. Am I missing something ?

You aren't missing anything, you'll get your client ID once your application is accepted at https://e-hentai.org/hentaiathome.php

I think I've failed.
I use the ID of my client and a password that I've configured seems not work.
[i.imgur.com] https://i.imgur.com/Ruy67l7.png

On the settings page I can not change the port used that I configured on my router:
"Note: The port cannot be changed while the client is running." The client is not runniong and I'm not connected too.
and in the title bar I've:
Client Key: Not Yet Configured

This post has been edited by gouzi: Oct 4 2022, 21:30

I think I've failed.
I use the ID of my client and a password that I've configured seems not work.
[i.imgur.com] https://i.imgur.com/Ruy67l7.png

On the settings page I can not change the port used that I configured on my router:
"Note: The port cannot be changed while the client is running." The client is not runniong and I'm not connected too.
and in the title bar I've:
Client Key: Not Yet Configured

If I'm not mistaking, to get a "Client key" (and change the port), you need to enter these three values, then click on "Apply Settings", at the bottom of the page :
(IMG:[i.imgur.com] https://i.imgur.com/FXZxmY1.png)

Then, it will generate a "Client key".
When you open the program, it will ask you for the "Client ID", then for the "Client key" (when I launched it, the first time, it took a good minute to ask for the "Client ID").
Both of these values are at the top of the page:
(IMG:[i.imgur.com] https://i.imgur.com/d5yoiQr.png)

This post has been edited by m0t: Oct 4 2022, 22:28

Thank you.
That's works.

Can static Range cap becomes greater than 6,000 for single client?

This post has been edited by sl4yer: Oct 8 2022, 15:41

Can static Range cap becomes greater than 6,000 for single client?

Nope

Is it normal that my client is doing nothing?

Is it normal that my client is doing nothing?

How long have you been running it for in total?

Putting together some of what other people here were trying to do and my own pet project I present this:
[gitlab.com] https://gitlab.com/gruntledw/HentaiAtHome/#overview
It's a GraalVM native-image compiled and dockerized version of the H@H client with additional (command-line optional) features and instructions to allow nginx to terminate the TLS connections.
Why? All the advantages of offloading work to a modern widely used webserver:
- HTTP/2 support
- Range support
- Better generic logging/metrics support
- Ability to add extra TLS related items "for free", like HSTS headers and OCSP stapling
- Lower CPU usage
It also looked like an interesting and fun side project.

Command line features that were added:
- Ability to disable TLS (since nginx is handling it)
- Ability to get requester IP from an nginx supplied header (so that origin checks and client side logging still work)
- X-Accel-Redirect support: cached file service can be delegated to nginx (but is still controlled by the client)
- HTTP Keepalive support. (Note: this is only useful because of the nginx proxy, real end users would not benefit from connecting to a H@H client directly with keepalives)

Limitations:
- if the TLS disable option is supplied, bandwidth/connection/request-flood protection are disabled. This could be fixed, but nginx can be configured to handle it instead, so why bother.

Concretely, what I got:
- Substantial whole system CPU usage reduction (compared to an OpenJDK 8 docker image base). It's difficult to give precise numbers, but I got ~50% lower user CPU usage on a system with a bunch of lightly used other services
- Mild memory savings, including the cost of running a new web server. If you already have one, that could be more substantial.

I currently have 3 clients running this code, and it looks fairly stable

Other notes:
Using SNI to select which cert to send (and sending a cert other than the H@H cert by default) will absolutely trash your quality, despite your trust being pegged at +1000 (result so far with <100 static ranges, if that matters).

[gitlab.com] https://gitlab.com/gruntledw/HentaiAtHome/#overview

Got it work on my machine. (Although H@H client thinks every request is zero bytes.)

CODE

2022-11-14T09:42:49Z [info] {CENSORED} Code=200 Bytes=0  Finished processing request in 0.00s

One of my clients that was only down a few weeks is showing as revoked. Another one is down much longer, but I don't care as much about that one. I might as well get the good one back up though.

This post has been edited by yeller: Nov 23 2022, 07:49

The speedtest between H@H should not be performed between the same users, because multiple H@H clients may be in the same LAN, and the speed is unreliable.

The speedtest between H@H should not be performed between the same users, because multiple H@H clients may be in the same LAN, and the speed is unreliable.

Hmm. I thought it already avoided testing between clients from the same user, but apparently it didn't. It should be doing that now.

Hey! I just started running H@H on it's own VPS. My trust and quality have been great but I noticed a sudden quality drop; I want to check and see if it's an issue with my client or if it is an issue outside of my client.

The log messages I keep seeing are:

CODE

2022-12-12T03:11:46Z [debug] {4020/$ip}   The connection was interrupted or closed by the remote host.
2022-12-12T03:11:46Z [debug] Remote host terminated the handshake

Most of these IPs seem to belong to Cloudflare, with a few being from various other vendors. Is this to be expected or should I dig deeper on my end? I don't mind troubleshooting too much (my day job is a Devops Eng) but I don't want to chase an endless rabbit hole since this seems to be some sort of TLS/SSL negotiation issue. But it does seem to be causing a quality drop that's pretty massive (which I want to avoid!).

Tried searching the forums and Google but nothing super useful came up. Not sure if Java supports a more verbose logging option with JVMs since all of my professional work has been with other languages. Let me know if there's something I should look for or if I can ignore these.

Thanks!

Hey! I just started running H@H on it's own VPS. My trust and quality have been great but I noticed a sudden quality drop; I want to check and see if it's an issue with my client or if it is an issue outside of my client.

Most of these IPs seem to belong to Cloudflare, with a few being from various other vendors. Is this to be expected or should I dig deeper on my end? I don't mind troubleshooting too much (my day job is a Devops Eng) but I don't want to chase an endless rabbit hole since this seems to be some sort of TLS/SSL negotiation issue. But it does seem to be causing a quality drop that's pretty massive (which I want to avoid!).

Quality looks fine to me. H@H doesn't work over Cloudflare, so if you're getting Cloudflare IPs it might be someone with a misconfigured Cloudflare site pointing to your IP or something along those lines.

Quality looks fine to me. H@H doesn't work over Cloudflare, so if you're getting Cloudflare IPs it might be someone with a misconfigured Cloudflare site pointing to your IP or something along those lines.

Awesome! Yeah, quality went back up after dipping. Assuming that's normal from the wiki due to this being a newer client. Just wanted to make sure it was an issue with the connections to my server (probably scanning tbh) and not my server. Thanks!

Edit: Now that I think about it, it could also be Cloudflare WARP or their VPN scanning ranges (if clients don't make direct connections to the server, if they do, then it's probably just someone using their VPN.) No huge deal. Just glad nothing's misconfigured on my end

This post has been edited by viablerr: Dec 12 2022, 12:36

Hello,
My static ranges, hitrate and hathrate were steadily growing.
But it's been several days that the static range doesn't move, with hitrate and hathrate stabilizing (see image under).

My "maximum disk cache size" is set to 1550Go, my "maximum burst speed" is set to 43750 KBytes/s, and my "quality" is above 7000.
Looking at the wiki, with these values, I should be able to go to 6000 static ranges.

1375 static range corresponds to 343,75Go of cache.
I'm now at ~107Go of used cache.
Do I have to wait for my cache usage to go above 343,75Go, to get more static range ?


(IMG:[i.imgur.com] https://i.imgur.com/5OZ0bzz.png)

This post has been edited by m0t: Dec 18 2022, 01:00

My static ranges, hitrate and hathrate were steadily growing.
But it's been several days that the static range doesn't move, with hitrate and hathrate stabilizing (see image under).
Do I have to wait for my cache usage to go above 343,75Go, to get more static range ?

Static range assignments do not depend on cache usage. Your client is in a fairly saturated area, so most likely there are no unassigned ranges available.

Static range assignments do not depend on cache usage. Your client is in a fairly saturated area, so most likely there are no unassigned ranges available.

Thank you for the information.
I'm too impatient (IMG:[invalid] style_emoticons/default/smile.gif) Just need to let it on, and wait.