[X] My Assistant

After I updated my h@h client to 1.6.2, the Hitrate of my two clients in China continued to drop, and recently I had a problem with the upload speed of my h@h client, the upload speed of each client was only 500kB/s, and the upload speed was around 7000kB/s before that, what is the cause of this. (My English is not good, so I used a translation tool) (IMG:[invalid] style_emoticons/default/cry.gif)

I've made some further tweaks to the initial speed test for China and, to some extent, the other distant regions where it's not all that relevant. Should be much easier to get it running now.

i noticed that now the speedtest is only 25000 bytes long lol
furthermore, i think some of the ISPs supposed to have a bad connection (or simply block) with rpc.hentaiathome.net (and the 3 ips behind that domain).

CODE

This is from China Telecom, Home Network (106.4.*.*)

C:\Users\Administrator>tcping -t rpc.hentaiathome.net

** Pinging continuously.  Press control-c to stop **

Probing 94.100.24.69:80/tcp - Port is open - time=331.706ms
Probing 94.100.24.69:80/tcp - No response - time=2014.440ms
Probing 94.100.24.69:80/tcp - Port is open - time=321.356ms
Probing 94.100.24.69:80/tcp - Port is open - time=348.978ms
Probing 94.100.24.69:80/tcp - Port is open - time=311.453ms
Probing 94.100.24.69:80/tcp - No response - time=2011.889ms
Probing 94.100.24.69:80/tcp - Port is open - time=315.798ms
Probing 94.100.24.69:80/tcp - Port is open - time=320.881ms
Probing 94.100.24.69:80/tcp - Port is open - time=311.807ms
Probing 94.100.24.69:80/tcp - Port is open - time=314.420ms
Probing 94.100.24.69:80/tcp - Port is open - time=347.864ms
Probing 94.100.24.69:80/tcp - Port is open - time=328.334ms
Control-C

Ping statistics for 94.100.24.69:80
     12 probes sent.
     10 successful, 2 failed.  (16.67% fail)
Approximate trip times in milli-seconds (successful connections only):
     Minimum = 311.453ms, Maximum = 348.978ms, Average = 325.260ms

CODE

and this is from #46202 which hosted on a China Mobile Static IP network (they blocked all ips which h@h receives static ranges and images, but they didnt block rpc.hentaiathome.net)

C:\Users\Administrator>tcping -t rpc.hentaiathome.net

** Pinging continuously.  Press control-c to stop **

Probing 94.100.24.69:80/tcp - Port is open - time=1258.774ms
Probing 94.100.24.69:80/tcp - Port is open - time=261.412ms
Probing 94.100.24.69:80/tcp - Port is open - time=1281.638ms
Probing 94.100.24.69:80/tcp - Port is open - time=284.072ms
Probing 94.100.24.69:80/tcp - Port is open - time=256.988ms
Probing 94.100.24.69:80/tcp - Port is open - time=1261.639ms
Probing 94.100.24.69:80/tcp - Port is open - time=279.656ms
Probing 94.100.24.69:80/tcp - Port is open - time=278.749ms
Probing 94.100.24.69:80/tcp - Port is open - time=266.005ms
Probing 94.100.24.69:80/tcp - Port is open - time=275.744ms
Probing 94.100.24.69:80/tcp - Port is open - time=284.744ms
Probing 94.100.24.69:80/tcp - Port is open - time=257.901ms
Control-C

Ping statistics for 94.100.24.69:80
     12 probes sent.
     12 successful, 0 failed.  (0.00% fail)
Approximate trip times in milli-seconds:
     Minimum = 256.988ms, Maximum = 1281.638ms, Average = 520.610ms

i noticed that now the speedtest is only 25000 bytes long lol
furthermore, i think some of the ISPs supposed to have a bad connection (or simply block) with rpc.hentaiathome.net (and the 3 ips behind that domain).

This is most likely due to the maintenance of the SMW5 cable, the bandwidth between Asia and Europe is seriously insufficient.
[www.kentik.com] https://www.kentik.com/analysis/smw-5-maint...urope-and-asia/

When I tried to start , it show me an error message like this, how to solve this...


FAIL_CONNECT_TEST:error:1408F10B:SSL routines:ssl3_get_record:wrong version number

When I tried to start , it show me an error message like this, how to solve this...
FAIL_CONNECT_TEST:error:1408F10B:SSL routines:ssl3_get_record:wrong version number

Usually means that when it tries to connect, it's something else than H@H running on the port in question. You'd need to diagnose your network setup to figure out what, or try changing the port.

I saw that I had 622 hours of uptime (not as many java vm crash as in the past it seems).

It made me wonder if it could be recommended after a certain amount to restart the client?

It made me wonder if it could be recommended after a certain amount to restart the client?

Not really. Pretty sure I've run the client for like a year without restarting it, so unless it's having some sort of issue there should be no need to restart it.

After repairing my home network and updating H@H to 1.6.2, I noticed that my trust kept dropping to almost 0, and also noticed something strange in the log: the first TCP connection closed immediately after the handshake, and then the second TCP connection worked fine. What could be causing of this, and how do I fix it? Thanks.
log and wireshark screenshot //i.imgur.com/YbsqDC5.png

After repairing my home network and updating H@H to 1.6.2, I noticed that my trust kept dropping to almost 0, and also noticed something strange in the log: the first TCP connection closed immediately after the handshake, and then the second TCP connection worked fine. What could be causing of this, and how do I fix it? Thanks.
log and wireshark screenshot //i.imgur.com/YbsqDC5.png

Client looks fine to me. It's not unusual that connections to clients in that area randomly have connections get RST'ed due to government censorship.

sorry if its already been asked but I couldn't find the answer.

is any direct performance benefit of running hath at port 443, or just ease of setup?

is any direct performance benefit of running hath at port 443, or just ease of setup?

The direct benefits are a slight increase in traffic and a moderate increase in quality. But increased quality also makes the client score better in various algorithms, which has indirect benefits, like more ranges.

I think with the latest changes my server's hath/day has plummeted. IIRC I was getting a hitrate of 120 hits/min and around 80 hath/day and now it's down to 28 hits/min and 30 hath/day. Also my average upload is extremely slow at around 2 Mbps.

I've seen it recovering a bit when I updated to 1.6.2 from 1.6.0 but I think it stagnated since. I've been top 300 for years but I don't even appear any more in the last month/day rankings...

This post has been edited by EmuAGR: Nov 21 2023, 00:29

Most clients that had a large cache but comparatively slow connections relative to the cache size ended up with a drop in hitrate after the switchover to priority ranges. This was intended to some extent, but was harsher than expected.

The scoring algorithm has been tweaked significantly since the initial assignment in favor of clients with a large cache and high quality, and this is not fully reflected in the priority range allocation yet. HC ranges should also to some extent have made up for it, but while the hath rate is adjusted to compensate for the relatively large files, it's not currently reflected in the toplist score to the extent that it should be - I'll make a note to fix that.

There are concrete plans to offload more traffic onto H@H, specifically thumbnails, but this requires a fundamental rework of how thumbnails work. This is underway, but will take several months as it requires a redeployment of the relevant servers followed by a regeneration of all the thumbnails.

Since october updates my client(41453) cache hitrate has been decreasing to almost 50%.
Is this behavior to be expected?


last year| Total File Sent: 10,xxx,xxx Total Files Rcvd 4xx,xxx
After | Total File Sent: 100,xxx Total Files Rcvd 5x,xxx

This post has been edited by sl4yer: Nov 21 2023, 14:19

"Static Range allocation is limited to 1 per 5 KB/s", seems to be a recent update? Means no need for big hard disk anymore? I was able to fill up 6000 static range, 200Hath/day before, now after resetting the static range I can run 1620 static range only, and 90+Hath/day. that's bad. (IMG:[invalid] style_emoticons/default/unsure.gif)

"Static Range allocation is limited to 1 per 5 KB/s", seems to be a recent update? Means no need for big hard disk anymore? I was able to fill up 6000 static range, 200Hath/day before, now after resetting the static range I can run 1620 static range only, and 90+Hath/day. that's bad. (IMG:[invalid] style_emoticons/default/unsure.gif)

yes, its a recent update. i think the meaning of this is to disperse the static range to improve quality and speed.
(i use a 30mbps connection in few years ago and i received likely 4k static ranges at the peak, now im using 100mbps one but static ranges capped to 2785)

I want to move my h@h client from a windows pc to a macos one. Can I simply:

1. Shutdown client on PC
2. Copy entire folder, which should include the cache files, to mac
3. Start client on mac

I want to move my h@h client from a windows pc to a macos one. Can I simply:

1. Shutdown client on PC
2. Copy entire folder, which should include the cache files, to mac
3. Start client on mac

Yes

My client's Trust was suddenly negative but the webpage says it's Online. It wasn't clear what was wrong so I restarted the client to verify cache. Then it said that external access to my client's port wasn't working. I think I've fixed the port issue but is Trust the only webpage indicator of inbound network issues? I'm guessing the Online status just means the client was able to make an outgoing connecting to the server.

Would be cool if there were an app to show your trust level in the system tray/notification area.

This post has been edited by badbob001: Dec 14 2023, 21:14

why is my max speed stuck at 32000 on 1 gigabit?

[www.speedtest.net] https://www.speedtest.net/result/15634809772