fah-client 8.4.9: HTTP_SERVICE_UNAVAILABLE ("No appropriate assignment")?

[joekienzle]

Hello,

I recently installed fah again and set it up sucessfully in the past, but now I am getting this log:

Code: Select all

cat /var/log/fah-client/log.txt 
*********************** Log Started 2025-07-06T16:03:07Z ***********************
16:03:07:I1:*********************** Folding@home Client ***********************
16:03:07:I1:    Version: 8.4.9
16:03:07:I1:     Author: Joseph Coffland <joseph@cauldrondevelopment.com>
16:03:07:I1:        Org: foldingathome.org
16:03:07:I1:  Copyright: 2023-2024, foldingathome.org
16:03:07:I1:   Homepage: https://foldingathome.org/
16:03:07:I1:    License: GPL-3.0-or-later
16:03:07:I1:        URL: https://v8-4.foldingathome.org/
16:03:07:I1:       Date: Nov 20 2024
16:03:07:I1:       Time: 14:47:19
16:03:07:I1:   Revision: 360fe71b1bd05bb89814bfb97b73a5bda84802d6
16:03:07:I1:     Branch: master
16:03:07:I1:   Compiler: GNU 8.3.0
16:03:07:I1:    Options: -Wsuggest-override -faligned-new -std=c++17 -fsigned-char
16:03:07:I1:             -ffunction-sections -fdata-sections -O3 -funroll-loops -fno-pie
16:03:07:I1:   Platform: linux 4.19.0-27-cloud-amd64
16:03:07:I1:       Bits: 64
16:03:07:I1:       Mode: Release
16:03:07:I1:       Args: --config=/etc/fah-client/config.xml
16:03:07:I1:             --log=/var/log/fah-client/log.txt
16:03:07:I1:             --log-rotate-dir=/var/log/fah-client/
16:03:07:I1:     Config: /etc/fah-client/config.xml
16:03:07:I1:****************************** CBang ******************************
16:03:07:I1:    Version: 1.7.2
16:03:07:I1:     Author: Joseph Coffland <joseph@cauldrondevelopment.com>
16:03:07:I1:        Org: Cauldron Development
16:03:07:I1:  Copyright: Cauldron Development, 2003-2024
16:03:07:I1:   Homepage: https://cauldrondevelopment.com/
16:03:07:I1:    License: LGPL-2.1-or-later
16:03:07:I1:       Date: Nov 19 2024
16:03:07:I1:       Time: 21:54:38
16:03:07:I1:   Revision: 443c54e909eb8d8994405a18fb328b5b05a623a5
16:03:07:I1:     Branch: master
16:03:07:I1:   Compiler: GNU 8.3.0
16:03:07:I1:    Options: -Wsuggest-override -faligned-new -std=c++17 -fsigned-char
16:03:07:I1:             -ffunction-sections -fdata-sections -O3 -funroll-loops -fno-pie
16:03:07:I1:             -fPIC
16:03:07:I1:   Platform: linux 4.19.0-27-cloud-amd64
16:03:07:I1:       Bits: 64
16:03:07:I1:       Mode: Release
16:03:07:I1:***************************** System ******************************
16:03:07:I1:        CPU: Intel(R) Celeron(R) N5105 @ 2.00GHz
16:03:07:I1:     CPU ID: GenuineIntel Family 6 Model 156 Stepping 0
16:03:07:I1:       CPUs: 2
16:03:07:I1:     Memory: 31.19GiB
16:03:07:I1:Free Memory: 2.77GiB
16:03:07:I1: OS Version: 6.8
16:03:07:I1:Has Battery: false
16:03:07:I1: On Battery: false
16:03:07:I1:   Hostname: folding
16:03:07:I1: UTC Offset: 2
16:03:07:I1:        PID: 125
16:03:07:I1:        CWD: /var/lib/fah-client
16:03:07:I1:       Exec: /usr/bin/fah-client
16:03:07:I1:*******************************************************************
16:03:07:I2:<config>
16:03:07:I2:  <!-- Account -->
16:03:07:I2:  <account-token v='XXXXXX'/>
16:03:07:I2:  <machine-name v='XXXXXX'/>
16:03:07:I2:
16:03:07:I2:  <!-- User Information -->
16:03:07:I2:  <passkey v='*****'/>
16:03:07:I2:  <user v='XXXXXX'/>
16:03:07:I2:</config>
16:03:07:I1:Opening Database
16:03:07:I1:F@H ID = k9wrqpmLZnb1YCIOSqxmJcaUOBLbG4ab81cp_--w8p4
16:03:07:I3:Loading default resource group
16:03:07:I1:Listening for HTTP on 127.0.0.1:7396
16:03:07:I3:Loaded 0 wus.
16:03:07:I1:Default:Added new work unit: cpus:2 gpus:
16:03:07:I1:WU7:Requesting WU assignment for user XXXXXX team 0
16:03:08:I1:OUT1:> POST https://assign1.foldingathome.org/api/assign HTTP/1.1
16:03:08:I1:OUT1:< HTTP/1.1 503 HTTP_SERVICE_UNAVAILABLE
16:03:08:E :OUT1:HTTP_SERVICE_UNAVAILABLE: {"error":"No appropriate assignment"}
16:03:08:I1:WU7:Retry #1 in 2 secs
16:03:10:I1:WU7:Requesting WU assignment for user XXXXXX team 0
16:03:11:I1:OUT2:> POST https://assign2.foldingathome.org/api/assign HTTP/1.1
16:03:11:I1:OUT2:< HTTP/1.1 503 HTTP_SERVICE_UNAVAILABLE
16:03:11:E :OUT2:HTTP_SERVICE_UNAVAILABLE: {"error":"No appropriate assignment"}
16:03:11:I1:WU7:Retry #2 in 4 secs
16:03:15:I1:WU7:Requesting WU assignment for user XXXXXX team 0
16:03:15:I1:OUT3:> POST https://assign3.foldingathome.org/api/assign HTTP/1.1
16:03:16:I1:OUT3:< HTTP/1.1 503 HTTP_SERVICE_UNAVAILABLE
16:03:16:E :OUT3:HTTP_SERVICE_UNAVAILABLE: {"error":"No appropriate assignment"}
16:03:16:I1:WU7:Retry #3 in 8 secs
16:03:24:I1:WU7:Requesting WU assignment for user XXXXXX team 0
16:03:24:I1:OUT4:> POST https://assign4.foldingathome.org/api/assign HTTP/1.1
16:03:25:I1:OUT4:< HTTP/1.1 503 HTTP_SERVICE_UNAVAILABLE
16:03:25:E :OUT4:HTTP_SERVICE_UNAVAILABLE: {"error":"No appropriate assignment"}
16:03:25:I1:WU7:Retry #4 in 16 secs

Firewall is not blocking. I do not get what I am doing wrong. Any advice?

[Nicolas_orleans]

If you are trying to fold with 2 CPUs I think the minimum is 3 CPUs these days

[joekienzle]

Thanks for the reply. I tried with 4 cpus and got the same result sadly

[Nicolas_orleans]

Mmm based on a quick internet search it would appear this Celeron N5105 laptop CPU would not support AVX instructions that are required for recent cores, and is limited to SSE 4.2

This looks similar to this thread for me:
viewtopic.php?p=370319&hilit=Sse+4.2#p370319

If one of the mods could confirm?

[toTOW]

I have an old i7 920 that is still happily folding with the SSE2 version of the folding cores, so it is not the issue. The SSE 4.2 requirement is for GPU cores.

16:03:07:I1:***************************** System ******************************
16:03:07:I1: CPU: Intel(R) Celeron(R) N5105 @ 2.00GHz
16:03:07:I1: CPU ID: GenuineIntel Family 6 Model 156 Stepping 0
16:03:07:I1: CPUs: 2

That's weird because this CPU is supposed to have 4 cores .... are you running in a VM that doesn't have access to all available CPU cores ?

[joekienzle]

Yes I'm not giving the vm all cores. But I suppose I could and then control the cpu usage with the web interface. Still even with 4 cores the issues persists.

[muziqaz]

Is there a reason why you are running fah in VM instead of natively?

[joekienzle]

Yes. Because I run all kinds of other services on this host and would like to give a percentage of the unused cpu power to fah and proxmox can prioritize cpu cost at the vm level, so I can make sure if my other services need more cpu time proxmox can handle that.

[muziqaz]

Seems like a lot of unnecessary wasted power introducing another layer with VM, instead of just trusting linux scheduler to do its work

[joekienzle]

It's not a vm directly but an lxc container. The vm layer also wouldn't waste cpu power. How would it do that? If you have a lot of vms you add the oses resources that could be considered wasted but since its an lxc container. Instead of the Linux scheduler I just use proxmox to do the scheduling.

[muziqaz]

It's not a vm directly but an lxc container. The vm layer also wouldn't waste cpu power. How would it do that? If you have a lot of vms you add the oses resources that could be considered wasted but since its an lxc container. Instead of the Linux scheduler I just use proxmox to do the scheduling.

FAH requires minimum 3CPU cores for folding

[joekienzle]

I tried it with 4 cpus as well and it didn't work (see previous post). I've given up now, since it won't work.

[Blue_Bubble]

I've been having the same problem, no work units since:

Code: Select all

Code	User	Team	Credit	Assigned	Credited
Ok	Blue_Bubble	13232	6,882.97	2025-07-08 06:44:52	2025-07-09 03:13:02

This system is using 4 CPUs, and has been folding successfully for many years:

Code: Select all

Description
Intel(R) Core(TM) i7-4500U CPU @ 1.80GHz
Cores  4    Type   amd64

Should I just give up folding with this system?


ETA: I rebooted the system and it immediately picked up a new work unit

[muziqaz]

Create a separate resource group just for the CPU. There seems to be a bug in fah client where if you use default client settings and have all the resources jumbled up together in default group, when GPU is paused or finished, CPU will finish its work, and will not be able to pick any other work until the client restart.

[nada391]

like in default group "gpu only + zero cpu" and cpu group "desired available cpu cores, minimum 3 cores"?