[X] My Assistant

Is there a way to make the client log debug info to the console?

I ask because I like to disable logging to disk and do my logging with journald, but it does not seem to log debug to console.

Is there a way to make the client log debug info to the console?

There's no argument for it, but it's fairly trivial to change the OUTPUT bitmask at line 44 in Out.java to match the one for LOGOUT, then run the build script to make your own version.

thank you for your efforts

I have setup port forwarding at router 192.168.2.1
But the client seems to keep rejecting request.
Open port check tool tested SUCCESS
What could be the problem?

CODE

2020-04-11T18:57:08Z [info] Internal HTTP Server was successfully started, and is listening on port 443
2020-04-11T18:57:08Z [info] Notifying the server that we have finished starting up the client...
2020-04-11T18:57:08Z [debug] Connecting to 94.100.18.250...
2020-04-11T18:57:10Z [WARN] Rejecting connection request from 192.168.2.1 during startup.
2020-04-11T18:57:10Z [debug] Reading 612 bytes from h t t p://94.100.18.250/15/rpc?clientbuild=151&act=client_start&add=&cid=31135&acttime=1586631427&actkey=bcb5534b32682bb2cf2078c3d9b233e0c6d9af67
2020-04-11T18:57:10Z [debug] Finished in 0 ms, writeoff=612, success=yes
2020-04-11T18:57:10Z [debug] Received response: FAIL_CONNECT_TEST:SSL connect error

I have setup port forwarding at router 192.168.2.1
But the client seems to keep rejecting request.
Open port check tool tested SUCCESS
What could be the problem?

CODE

2020-04-11T18:57:08Z [info] Internal HTTP Server was successfully started, and is listening on port 443
2020-04-11T18:57:08Z [info] Notifying the server that we have finished starting up the client...
2020-04-11T18:57:08Z [debug] Connecting to 94.100.18.250...
2020-04-11T18:57:10Z [WARN] Rejecting connection request from 192.168.2.1 during startup.
2020-04-11T18:57:10Z [debug] Reading 612 bytes from h t t p://94.100.18.250/15/rpc?clientbuild=151&act=client_start&add=&cid=31135&acttime=1586631427&actkey=bcb5534b32682bb2cf2078c3d9b233e0c6d9af67
2020-04-11T18:57:10Z [debug] Finished in 0 ms, writeoff=612, success=yes
2020-04-11T18:57:10Z [debug] Received response: FAIL_CONNECT_TEST:SSL connect error

That's a non-transparent proxy. i.e. the router does not follow the IP from where the request came from. And therefore h@h cannot know where to send the image to. Dig deeper in the router config - h@h must receive the IP that connected to the WAN interface of the router (the one facing the internet, not the one facing your home).

I am currently running h@h 1.6 on raspberry pi 4, oracle jdk 8, bbr enabled. I see a significant max speed drop from 80000kbps to 25000kbps. Before the update i run the client on port 80 and now is 443, is that normal and will not affect the performance of my client?

my client got similar case to it

use raspberry 4, home connection, 20Mbit/s upload banwidth, H@H1.6, openJDK 1.8, BBR open.
I begain startup the H@H on Feb, at first I get correct Max Speed (2200kb-2500kb)
but at the end of Mar, it show my Max Speed stuck at 793kb, today at 634kb, no longer change.
(IMG:[i.loli.net] https://i.loli.net/2020/04/29/oAp7N1mStLghn32.png)

but I check the log, find it actually run in correct speed:
(IMG:[i.loli.net] https://i.loli.net/2020/04/29/tX5bd8mjSf3CWJq.png)

also check the speedtest-cli(version 2.1.2), correct:
(IMG:[i.loli.net] https://i.loli.net/2020/04/29/ITBSiqMctux2CVy.png)

it really a big hit to my client, the panel show that Trust and Quality is always max
but Hitrate 320->213 and Hathrate 70->40 during Apr, painful

dont konw what happen, and want someway to fix it, plz
forgive my bad english

PS:
I find that it will get less speed result in speedtest-cli 2.0.2:
(IMG:[i.loli.net] https://i.loli.net/2020/04/29/CdFjTcXapOoGS3B.png)

and this result looklike similar to my wrong speed in H@H

use raspberry 4, home connection, 20Mbit/s upload banwidth, H@H1.6, openJDK 1.8, BBR open.
I begain startup the H@H on Feb, at first I get correct Max Speed (2200kb-2500kb)
but at the end of Mar, it show my Max Speed stuck at 793kb, today at 634kb, no longer change.

If the client appears to be speed-limited that's because it was recently unable to handle a higher level of traffic. This could be either because of a hardware overload or bandwidth congestion. If the situation is resolved, the limit will be automatically lifted over time.

If the client appears to be speed-limited that's because it was recently unable to handle a higher level of traffic. This could be either because of a hardware overload or bandwidth congestion. If the situation is resolved, the limit will be automatically lifted over time.

got that，thank for solve my question, exactly my client appear serious trust problem before limited because of wrong network config, I have fixed that in this month, now I will wait for limited lifted.

CODE

top - 12:02:08 up  4:41,  3 users,  load average: 3.12, 3.05, 3.00
Tasks: 222 total,   1 running, 148 sleeping,   0 stopped,   0 zombie
%Cpu(s): 75.3 us,  0.2 sy,  0.0 ni, 24.3 id,  0.1 wa,  0.0 hi,  0.1 si,  0.0 st
KiB Mem : 16326244 total, 13047804 free,  2935044 used,   343396 buff/cache
KiB Swap:        0 total,        0 free,        0 used. 13108784 avail Mem

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
32313 hath      20   0 2435180 2.288g   1468 S 300.3 14.7 153:34.59 kswapd0
21645 hath      20   0 7679772 131076  29472 S   1.7  0.8   1:48.75 java
2039 hath      20   0   42956   4184   3420 R   0.3  0.0   0:00.19 top                                      

Recently, H@H seems to be the cause for incredibly high kswapd0 CPU consumption (3/4 cores capped), no matter if swap is enabled or disabled. I'm unsure if this is due to my system, but I haven't found any fix. Consulting the internet I have found none of the proposed fixes to help, including but not limited to echoing various numbers to /proc/sys/vm/drop_caches and /proc/sys/swappiness or to edit them at sysctl.conf.

CODE

top - 12:02:08 up  4:41,  3 users,  load average: 3.12, 3.05, 3.00
Tasks: 222 total,   1 running, 148 sleeping,   0 stopped,   0 zombie
%Cpu(s): 75.3 us,  0.2 sy,  0.0 ni, 24.3 id,  0.1 wa,  0.0 hi,  0.1 si,  0.0 st
KiB Mem : 16326244 total, 13047804 free,  2935044 used,   343396 buff/cache
KiB Swap:        0 total,        0 free,        0 used. 13108784 avail Mem

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
32313 hath      20   0 2435180 2.288g   1468 S 300.3 14.7 153:34.59 kswapd0
21645 hath      20   0 7679772 131076  29472 S   1.7  0.8   1:48.75 java
2039 hath      20   0   42956   4184   3420 R   0.3  0.0   0:00.19 top                                      

Recently, H@H seems to be the cause for incredibly high kswapd0 CPU consumption (3/4 cores capped), no matter if swap is enabled or disabled. I'm unsure if this is due to my system, but I haven't found any fix. Consulting the internet I have found none of the proposed fixes to help, including but not limited to echoing various numbers to /proc/sys/vm/drop_caches and /proc/sys/swappiness or to edit them at sysctl.conf.

Your kswapd runs as the hath user instead of root, and also uses 2.2 GB RAM by itself? That seems very odd and more than little suspicious, it's a system (root) process on every system I've ever used, nor does it list any used VSZ or RSS by itself; that kinda looks like something else pretending to be kswapd. What system are you running?

Try doing sudo ls -l /proc/32313/exe and see what it's actually running. Replace for actual PID if the system rebooted.

Edit: [yoroi.company] Possible relevant article. Your system is likely compromised and I would recommend you wipe and reinstall your OS.

cool im doing folding@home might add this to that

Your kswapd runs as the hath user instead of root, and also uses 2.2 GB RAM by itself? That seems very odd and more than little suspicious, it's a system (root) process on every system I've ever used, nor does it list any used VSZ or RSS by itself; that kinda looks like something else pretending to be kswapd. What system are you running?

Try doing sudo ls -l /proc/32313/exe and see what it's actually running. Replace for actual PID if the system rebooted.

Edit: [yoroi.company] Possible relevant article. Your system is likely compromised and I would recommend you wipe and reinstall your OS.

I guess I got pwned, though I could swear I had hath excluded from sshd_config. Thanks.

Edit: My IP has been reported on abuseipdb for SSH bruteforce, and looking into it everything else checks out as well. Thanks again, Tenboro.

This post has been edited by Red-Ass Panda: Apr 29 2020, 15:57

I liked this update so far.

Docker image for those that need it. Fully configurable with environment variables and volume mounts. Shuts down cleanly so no cache check every restart.

[hub.docker.com] https://hub.docker.com/r/frosty5689/hath/

For the local H@H, isn't it possible to swap out the H@H client domains of others with the local H@H instance's domain? Then it would hit local H@H for the images, which will result in a fetch if it isn't found. Kinda miss being able to load galleries instantly from H@H cache.

For the local H@H, isn't it possible to swap out the H@H client domains of others with the local H@H instance's domain? Then it would hit local H@H for the images, which will result in a fetch if it isn't found. Kinda miss being able to load galleries instantly from H@H cache.

H@H way back when (prior to 1.3) used to have a different way of mapping files to the cache, and while it would still be technically possible, you'd no longer get much of the caching/speed benefits, so it's not really worth the complexity.

H@H way back when (prior to 1.3) used to have a different way of mapping files to the cache, and while it would still be technically possible, you'd no longer get much of the caching/speed benefits, so it's not really worth the complexity.

Yeah, makes sense. Especially since there is a chance H@H is unresponsive/down and needs the viewer to be able to handle that without failing to load images entirely.

I have a strange issue where I have forwarded a port to a VM I am running on UnRAID (my server OS) but the external test fails.

I setup a port forward to my UnRAID IP address and it works fine. It was working fine earlier in the day on the VM but I extended the HDD Partition so I could increase the Cache on the machine which has then caused it to fail.

What logs would you require to possibly assist?

Regards

I just saw something on the options on the site of H@H:

This must be at least 10 GB. We recommend at least 1 GB disk per 25 KByte/s bandwidth, or 5 GB for every MBit/s, but more is always better.

The part in bold should be either Mbit/s or MByte/s. If it's a mix of the two, that doesn't help the fact that those 2 units are confusing.
Actually, it could even be 125 KByte/s, would be consistent with the previous value (IMG:[invalid] style_emoticons/default/wacko.gif)

This post has been edited by uareader: May 13 2020, 22:56

Actually, it could even be 125 KByte/s, would be consistent with the previous value (IMG:[invalid] style_emoticons/default/wacko.gif)

It's the same. 1GB / (25 Kbyte * 8) = 5GB / 1000 Kbit