[X] My Assistant

Now, even though I've had three times the speed since January, 2022 and its stored on a pair of SSDs my cache won't generally show an average rate of more than about 0.9KB/s. I can understand slow growth when the client is absolutely brand new... but it's been over six months.

Some locations are rather over-saturated, and while clients in those locations will still grow over time, new ranges would only be assigned when old clients go away.

Hopefully getting a new PC soon and will try this out

I didn't even know there was a client. Link for more info please.

There is even a FAQ linked in the first post, but okay:

https://ehwiki.org/wiki/Hentai@Home
/index.php?showtopic=19795

Is there any way i can regain my id? it says it's been revoked due to inactivity...

inspired by LocalMaximumJo's nginx conf , i managed to run h@h along with my websites on port 443 with haproxy.

any suggestions would be appreciated since i'm a noob.

CODE

global
    maxconn 500000
    ca-base /data/ssl
    crt-base /data/ssl

    #only disable SSLv3 since tls1 and tls1.1 are still used by h@h client
    ssl-default-bind-options no-sslv3
    tune.ssl.default-dh-param 2048

defaults
    log    global
    mode    tcp
    option    dontlognull
    timeout connect 10000
    timeout client 100000
    timeout server 100000

frontend tls-in
    bind *:443
    tcp-request inspect-delay 5s
    tcp-request content accept if HTTP
    acl hath req.ssl_sni -m end .hath.network #match requests contain a sni with *.hath.network
    acl foobar req.ssl_sni -m end .foobar.com #match requests contain a sni with *.foobar.com
    use_backend hath if hath #redirect h@h traffic to h@h client
    use_backend foobar if foobar #redirect other web traffic to website frontend
    
backend hath
    server server1 127.0.0.1:8443 #running h@h with port 8443

backend foobar
    server server1 127.0.0.1:8444 #same as website frontend port
    
    
frontend foobar
    bind 127.0.0.1:8444 ssl crt /data/ssl/foobar.com
    tcp-request inspect-delay 5s
    tcp-request content accept if HTTP
    acl 1 hdr_beg(host) -i 1.foobar.com
    acl 2 hdr_beg(host) -i 2.foobar.com
    use_backend 1 if 1
    use_backend 2 if 2

backend 1
    server server1 127.0.0.1:8080

backend 2
    server server1 127.0.0.1:8081

frontend http-in
    bind *:80
    mode http
    redirect scheme https code 301

with this config i'm running h@h with

CODE

java -jar HentaiAtHome.jar --disable-ip-origin-check --port=8443 --silentstart

This post has been edited by wtmsbb: Jul 16 2022, 05:42

Lately I have problem with very slow speed test server although my network is fine.

Whenever the client doing speed test routine, the connection to the test server (94.100.24.69) is very slow, like less than 1MBps.


However I have confirmed that my network is fine, and connecting to other clients during normal operation yields my actual speed.

The slowness only happen during speedtest.

This false speedtest result is hurting my client stats causing its hit and hath rate to tank.


Is there anything I could do to fix this?

This post has been edited by rozos: Jul 24 2022, 21:15

Lately I have problem with very slow speed test server although my network is fine.

Whenever the client doing speed test routine, the connection to the test server (94.100.24.69) is very slow, like less than 1MBps.

The startup speed test only matters if you don't have good tests from other clients yet. Assuming that you're talking about client 42620, it's being throttled because it keeps getting overloaded, which means that either the network or the host machine is struggling at times.

The startup speed test only matters if you don't have good tests from other clients yet. Assuming that you're talking about client 42620, it's being throttled because it keeps getting overloaded, which means that either the network or the host machine is struggling at times.

Ah. makes sense. I've been running this particular client with its cache in NFS, and there were some internal networking issues going on previously.
The throttling is new to me. Should be updated in wiki.

edit
Btw how long the throttle will go on?
I think 230kbps is bit excessive tho.

This post has been edited by rozos: Jul 25 2022, 08:30

Btw how long the throttle will go on?
I think 230kbps is bit excessive tho.

It's dropped by 20% every time, so it's only 230 KB/s because it overloaded at 288 KB/s. It starts lifting the restriction after 10 days and removes it entirely after 30 days, unless it gets overloaded again.

When I try to initialize the client it gets stuck at...

2022-07-26T13:02:14Z [info] Finished applying settings
2022-07-26T13:02:14Z [info] CacheHandler: Attempting to load persistent cache data...
2022-07-26T13:02:15Z [info] CacheHandler: All persistent fields found, loading remaining objects

Taskmanager shows a very high Java usage with HaH going to a full core at multiple cores. Also it doesn't seem to load further than that.
I am using 1.6.1

Both logs also show 0 errors. Heck the error log only shows that the logging started at the same time.

This post has been edited by Hakker: Jul 26 2022, 16:50

When I try to initialize the client it gets stuck at...

2022-07-26T13:02:14Z [info] Finished applying settings
2022-07-26T13:02:14Z [info] CacheHandler: Attempting to load persistent cache data...
2022-07-26T13:02:15Z [info] CacheHandler: All persistent fields found, loading remaining objects

Taskmanager shows a very high Java usage with HaH going to a full core at multiple cores. Also it doesn't seem to load further than that.

Sounds like the persistent cache data got corrupted, which could be a hardware issue or a problem with Java itself. Shut it down and delete all the pcache files from the data directory, then start it back up.

Sounds like the persistent cache data got corrupted, which could be a hardware issue or a problem with Java itself. Shut it down and delete all the pcache files from the data directory, then start it back up.

That indeed solved it. Thank you.

Some locations are rather over-saturated, and while clients in those locations will still grow over time, new ranges would only be assigned when old clients go away.

It's been almost two more months stuck at a sub 1KB/s DL speed and I thought things might pick up, but am I stuck at roughly 2 hath per day for the next umpteen years?

It's been almost two more months stuck at a sub 1KB/s DL speed and I thought things might pick up, but am I stuck at roughly 2 hath per day for the next umpteen years?

For a new client located in the US, yeah, it will probably take ages. Not really anything that can be done about that.

The software defaults to single-file download, but in most cases it is far from reaching the upper limit. Can the download thread be increased in later versions?
Also, is this the formula for calculating the weekly download limit?
min(avghits,LOG(1.003 + (avghits / 200), 1.003))x10.94MB

I created this Docker image for my own personal use but for some reason it has been pulled 500K+ times now... It'll work on ARM based processors too. I try to keep it up to date using the latest H@H client. It is one of the few that actually lets you shutdown H@H cleanly when you stop the Docker container.

[hub.docker.com] https://hub.docker.com/r/frosty5689/hath/ta...ng=last_updated

Can you help me with the configuration in UnRaid?
I had already written to you privately on 23.08.22, but have not yet received a reply
I do not know exactly which parameters are specified where.
What I specify IN the parameters I already understand but how are the following command/variables to understand?
--name hath \
--user 99:100 \
-v
-e
-p

help

help

You must forward your chosen port in your router and firewall.
Port: 20108

This might help you if you are not familiar with port forwarding:
[portforward.com] https://portforward.com/how-to-port-forward/

This post has been edited by Spamfeind: Sep 21 2022, 22:39

I am merging my J4205 box to my server. The H@H will be run inside TrueNAS Jail.

In TrueNAS, I can set the Compression Level (e.g. LZ4) of the Dataset (which is like directory). I would like to know if there is any impact when the H@H cache is located in a LZ4 Dataset.