Repeatedly no response from WU-Server

[KhaleedWalten01]

Hello there.
I started to contribute to F@H again yet I think F@H seems to have major problems with their WU servers.
My GPU finished multiple WUs but after uploading them, the server solely responds nothing:

This line repeatedly shows in my log:

Code: Select all

12:32:37:WARNING:WU04:FS01:Exception: Failed to send results to work server: 10002: Received short response, expected 512 bytes, got 0
12:32:37:WU04:FS01:Trying to send results to collection server
12:32:37:WU04:FS01:Uploading 51.45MiB to 128.252.203.13
12:32:37:WU04:FS01:Connecting to 128.252.203.13:8080

It is not once that this error occurs. Had to even delete WUs in order to unclog the FAHClient window as these WUs would have expired soon anyways. Even restarting F@HCLient doesn't help.

Interesting though that WUs completed by my CPU seem to upload fine and get credited. Any problems reported with servers 128.252.203.10 and 128.252.203.13?
Because these two appear to be the troublemakes with my completed WUs and it affects GPU only (the more powerful one usually granting more credits!)

This is highly frustrating as contributing to F@H makes no sense if the server solely gives you the middle finger salute after painstakingly completing a WU.
Also it makes no sense forking over 50MB of data again and again just to get a "Nah, am not willing to respond to you at all" from that server

[gunnarre]

There have been reports from some users who can't upload to those servers, but the servers work well for most other people. I wonder if we need to start finding out which ISPs the affected users are from or if there's some common denominator in the software running on the machine.

[Joe_H]

A common problem besides software for those in Europe is Fritz!Box routers. Default settings can block some of the packets involved in HTTP file transfers

[toTOW]

It looks like the same issue as thins one : viewtopic.php?f=18&t=37716

[jjmiller]

Hi KhaleedWalten01,

Would you be willing to provide the Project/Run/Clone/Gen codes for a few of the WUs that are stuck? In particular it would be good if there's a WU that is currently stuck. I've been working to with the lead FAH developer to try and track down what exactly is going awry here.
Thanks!

[toTOW]

There's a screenshot here : viewtopic.php?p=354933#p354933

p17257, Run 1411, Clone 0, Gen 20 and p17257, Run 1081, Clone 2, Gen 19

[appepi]

I am running new RTX2060 and some WUs never upload even after restart, while others did so after restart. I reactivate RTX 1060 that sent hundreds of WUs OK before, now doesn't. Both have Kaspersky, both reach ISP through Fritzbox 7490 set as IP relay, then main 7490.

Log from 1060 below:

00:31:44:WU00:FS01:0x22:Completed 2650000 out of 5000000 steps (53%)
00:31:47:WU01:FS01:Sending unit results: id:01 state:SEND error:NO_ERROR project:18036 run:3 clone:30 gen:94 core:0x22 unit:0x0000001e0000005e0000467400000003
00:31:47:WU01:FS01:Uploading 32.62MiB to 34.72.228.44
00:31:47:WU01:FS01:Connecting to 34.72.228.44:8080
00:31:53:WU01:FS01:Upload 87.37%
00:31:59:WU01:FS01:Upload 97.33%
00:32:30:WARNING:WU01:FS01:Exception: Failed to send results to work server: 10002: Received short response, expected 512 bytes, got 0
00:32:30:WU01:FS01:Trying to send results to collection server
00:32:30:WU01:FS01:Uploading 32.62MiB to 128.252.203.13
00:32:30:WU01:FS01:Connecting to 128.252.203.13:8080
00:32:36:WU01:FS01:Upload 47.52%
00:32:42:WU01:FS01:Upload 57.48%
00:32:48:WU01:FS01:Upload 67.63%
00:32:54:WU01:FS01:Upload 77.79%
00:33:00:WU01:FS01:Upload 87.75%
00:33:06:WU01:FS01:Upload 97.72%
00:33:37:ERROR:WU01:FS01:Exception: 10002: Received short response, expected 512 bytes, got 0

[aetch]

A few suggestions:-
1). Kaspersky is known to cause problems with FAH, I think the best advice is to remove it.
At the very least every internet security application product has a list of exclusions, add "c:\programdata\fahclient" to this list.

2). I have no idea about your fritzbox. I do know some routers also cause problems with FAH.

3). To clear your backlog something I have seen someone else suggest is to pause and disconnect all your problem machines, reconnect one at a time and allow the backlog to upload.
One thing I have to note is that for roughly the past 10 weeks uploads have been deathly slow and may take each machine a while. I personally went from a handful of failed uploads per month to multiple failed uploads per day. I tried highlighting it but no-one at FAH seems to care.

[Joe_H]

People have reported problems with Fritz!Box units. I forget all of the details, but at default settings they can block or drop some of the network packets involved in HTTP file transfers used by the F@h client. One setting is involved with deep packet inspection, transfers only are allowed for "known browsers", FAHClient needs to be added as an exception. The other I have heard people run into is dropping of network transfer ACK packets under load.

[appepi]

Thanks for suggestions aetch and Joe_H. The problem is partly resolved: with the 2060 in a Z440 (Donor Z442 in my solo "Team") : Not Fritzboxes, not Kaspersky but Acronis True Image 2019 scheduler seems to be the villain. Since I don't schedule backups but do occasional full backups, disabling scheduler isn't a problem, and it seems to unclog the queue (though was combined with restart so maybe susperstitious behaviour). However, the 1060 in a Z600 (Donor Z603) that ran happily through 2020 delivering 101 million points didn't benefit from shutting down ATI scheduler and helper, nor from restart, nor from uninstall/reinstall of FAHclient. Strangely, the points seem to be arriving (hard to tell, but the total for Z603 increases now and then) , but it is the final ACK that doesn't work. Not wanting to burn down the planet by repeated useless uploads, Plan B is to retire Z603, but I'm letting it run for a good while to see if it sorts itself out. Maybe I'll reactivate one of the other antiques in my HP workstation museum to see if it OK or not.

[appepi]

Good News, thanks to viewtopic.php?p=354937#p354937 by KhaleedWalten01. The post contains a long recipe for Kaspersky Internet Security (KIS) settings (German), to which I applied Google translate to find them in my English Kaspersky Total Security. Translated bits are in bold. I only did the first part of the first step: That is, I simply added FAHclient.exe (not FAHControl.exe, but followed all the other steps except I didn't see an option to name the rule. It looks like the Fritzboxes are OK, and probably the Acronis scheduler was OK too. Thx to all parties.

"1. Like toTow already mentioned, check that you add FAHClient to your Program Exclusions. (Will write in German here as I don't have the English KIS)
- Open your KIS and click the gear icon at the lower left
- Click on "Network Settings" on the left, then on the main screen seek for "Trusted Applications" and click it.
- Add FAHClient.exe and FAHControl.exe to this list and check every item in the following screen. (2 rules)
- Note the arrow to the right at the "Do not scan encrypted traffic" and click it to change this entry to "Do not scan all traffic"
- Name the rule to your needs
- Make sure the status is "Active"
- Click on OK to add the program to the exclusion list
- Okay, this should help the Applications not to be stopped to freely transfer data to and from the web.

Then restart. Instant fix on Z442 with the 2060 and Z603 with the 1060

[appepi]

This is just an update to my previous post, needed because Kaspersky Total Security has been replaced by Kaspersky Premium and the behaviour is different and so is the interface. The good news seems to be that there is no need for any special settings and you can just treat FAHclient.exe as a regular Trusted Application. However, some of my devices ended up with odd "Exclusions" in the automated changeover, so this post shows how to reset the ones I found to the usual values for a Trusted application.

- Open your KP and click on the gear icon at lower left
- Click on "Security settings" in the Settings panel, then scroll down to the "Intrusion Prevention" box and click on the "Manage applications" [White text in green box]
- Scroll down the Trusted applications until you find FAHClient.exe. If it shows with a blue box (rather than the regular green) then double-click the box and it will bring up the "Application rules" box.
- Click on "Exclusions". You will then see a collection of tick-boxes and all of them should be unticked for regular use.

[I have only checked this with two devices so far: Z800/GTX1080 and a Z440/RTX2060. If you think you may still need to exclude scanning then scroll down to the option in green where you select "Do not scan all traffic" rather than "Do not scan encrypted traffic" and tick the box.]

- Click "Save", and confirm the change when asked. This will take you back to the "Manage applications" panel and the "Restrictions" box should be green (or blue if you have applied a rule somewhere).