v8.1.18 - frequent re-installs required (Ubuntu 22.04)

[DocJonz]

I have been running the beta on Ubuntu 22.04.2 LTS, via Firefox, for some time now. Currently using the beta v8.1.18.

What I find is that, if I restart the PC for any reason (I had some electric works done recently that cut all power), or I do a general software update, or I refresh 'Snap' (as Ubuntu Firefox is now done via Snap), the web client forgets the GPU details - the icons in the web-control just grey out.
So I have to re-install the v8.1.18 client everytime, and then it picks up fine where it left off.

Reading around, are others seeing this? and is it an issue with Firefox Snap?
If so, any remedies?

[prcowley]

I don't have any experience with v8 client but I had the same issue with V7 for quite some time, so I offer this in the hopes it might help you.

It seems that the config file got stomped on and the default one used on startup. I think this was related to the fact that FAH would auto start and sometimes also try and restart the unit it was working on when the power went off of the PC rebooted

To get around this I did:
1- Take a copy of the working xml config file as save it under a different name. If you get the blank clint on start-up, you can then stop FAH, copy the saved xml config file over the default one and restart FAH - all should be OK although you might loose the work unit that was being processed.

2 - Because this happened to me a lot. If I was going to do a shutdown I would stop the FAH service, do the reboot and when it started up again and FAH auto started, it was usually OK

Somewhere along the line this problem went away for reasons unknown

I hope this is useful because the problem drove me crazy for quite a while.

Regards
Pete Cowley

[calxalot]

Yea, pretty much what Pete said.

I think the problem is that the client doesn't have (over)write permission for config.xml.
By default /var/lib/fah-client/config.xml is a symbolic link to /etc/fahclient/config.xml

Something like this should work (untested):

Stop the client

Code: Select all

sudo systemctl stop fah-client

Make a copy of config.xml, delete the (possibly) symlink config, and rename your copy to config.xml.

Code: Select all

sudo su fah-client
cd /var/lib/fah-client
cp config.xml config-orig.xml
rm config.xml
mv config-orig.xml config.xml
exit

Start the client.

Code: Select all

sudo systemctl start fah-client

[calxalot]

If this fixes your problem, there should probably be a change made to the client install script.

[calxalot]

Thinking about it, I think something else is going on with v8.

The config file is not altered by v8.

[calxalot]

Is there any error message in the log about writing to client.db ?

[DocJonz]

@calxalot - I tried your approach, but when it came to 'cp config.xml config-orig.xml', although the file is listed in the directory, it says it doesn't exist - so I'm assuming that's the symlink (as its only 25 bytes)? I couldn't find a way to delete that.

I can check the log next time, but I don't expect it will say anything, as the GPU client doesn't start - in the GPU resource area of the FAH web interface, it doesn't detect CUDA or OpenCL, so it doesn't enable the GPU.

[calxalot]

Is there a gpus.json file?

[DocJonz]

Is there a gpus.json file?

Yes there is.

[DocJonz]

Is there any error message in the log about writing to client.db ?

So, following a general software update, I ran the webclient directly - this is the result (see below) - it doesn't detect the GPU and reverts to a single CPU operation. Fix - stop the client, re-install it, and it picks up Folding happily on the GPU where it left off previously.

Code: Select all

********************** Log Started 2023-07-05T06:54:44Z ***********************
06:54:44:I1:*********************** Folding@home Client ***********************
06:54:44:I1: Version: 8.1.18
06:54:44:I1: Author: Joseph Coffland
06:54:44:I1: Org: foldingathome.org
06:54:44:I1: Copyright: 2023 foldingathome.org
06:54:44:I1: Homepage: https://foldingathome.org/
06:54:44:I1: License: https://www.gnu.org/licenses/gpl-3.0.txt
06:54:44:I1: Date: Apr 18 2023
06:54:44:I1: Time: 12:09:09
06:54:44:I1: Revision: 80a3d5eb8f60f7833de2954087682958b511895c
06:54:44:I1: Branch: master
06:54:44:I1: Compiler: GNU 10.2.1 20210110
06:54:44:I1: Options: -faligned-new -std=c++17 -fsigned-char -ffunction-sections
06:54:44:I1: -fdata-sections -O3 -funroll-loops -fno-pie
06:54:44:I1: Platform: linux 5.10.0-16-cloud-amd64
06:54:44:I1: Bits: 64
06:54:44:I1: Mode: Release
06:54:44:I1: Args: --log=/var/log/fah-client/log.txt
06:54:44:I1: --log-rotate-dir=/var/log/fah-client/
06:54:44:I1:****************************** CBang ******************************
06:54:44:I1: Version: 1.7.2
06:54:44:I1: Author: Joseph Coffland
06:54:44:I1: Org: Cauldron Development LLC
06:54:44:I1: Copyright: Cauldron Development LLC, 2003-2023
06:54:44:I1: Homepage: https://cauldrondevelopment.com/
06:54:44:I1: License: GPL 2+
06:54:44:I1: Date: Apr 14 2023
06:54:44:I1: Time: 16:26:30
06:54:44:I1: Revision: ac8bbdd5bb93c01679a881f5962fed800bf29e58
06:54:44:I1: Branch: master
06:54:44:I1: Compiler: GNU 10.2.1 20210110
06:54:44:I1: Options: -faligned-new -std=c++17 -fsigned-char -ffunction-sections
06:54:44:I1: -fdata-sections -O3 -funroll-loops -fno-pie -fPIC
06:54:44:I1: Platform: linux 5.10.0-16-cloud-amd64
06:54:44:I1: Bits: 64
06:54:44:I1: Mode: Release
06:54:44:I1:***************************** System ******************************
06:54:44:I1: CPU: Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
06:54:44:I1: CPU ID: GenuineIntel Family 6 Model 60 Stepping 3
06:54:44:I1: CPUs: 8
06:54:44:I1: Memory: 7.69GiB
06:54:44:I1:Free Memory: 7.12GiB
06:54:44:I1: Threads: POSIX_THREADS
06:54:44:I1: OS Version: 5.15
06:54:44:I1:Has Battery: false
06:54:44:I1: On Battery: false
06:54:44:I1: UTC Offset: 1
06:54:44:I1: PID: 749
06:54:44:I1: CWD: /var/lib/fah-client
06:54:44:I1: Exec: /usr/bin/fah-client
06:54:44:I1:*******************************************************************
06:54:44:I2:
06:54:44:I1:Opening Database
06:54:44:I1:Listening for HTTP on 127.0.0.1:7396
06:54:44:I3:id = JlfY57OvY9Z3CHog6y6eBm7Kok3dnJJ/FwnX9heeRwE=
06:54:44:I3:Loading work unit 1478 to group '' with ID vmAxjtAFa6-NP9z9o75-2rNY58xV9tOWCY3QgTB5bOk
06:54:44:I3:Loaded 1 wus.
06:54:44:E :Exception: clGetPlatformIDs() returned -1001
06:54:44:E :Exception: cuInit() returned 100
06:54:44:I3:gpus = {
06:54:44:I3: "gpu:01:00:00": {"vendor": 4318, "device": 9352, "type": "nvidia", "supported": false}
06:54:44:I3:}
06:54:44:I1:OUT1:> GET https://api.foldingathome.org/gpus HTTP/1.1
06:54:44:E :Exception: Could not get IP address for api.foldingathome.org: Temporary failure in name resolution
06:55:37:I1::Added new work unit: cpus:1 gpus:
06:55:37:I1::WU1479:Requesting WU assignment for user DocJonz team 35947
06:55:37:I1:OUT3:> POST https://assign1.foldingathome.org/api/assign HTTP/1.1
06:55:37:I3:Connecting to assign1.foldingathome.org:443
06:55:38:I1:OUT3:< assign1.foldingathome.org:443 HTTP/1.1 200 HTTP_OK
06:55:38:I1::WU1479:Received WU assignment m2FBWVelhW3-aE0kZgRZhAZuKVOnAz2qGeySCBvMnSk
06:55:38:I1::WU1479:Downloading WU
06:55:38:I1:OUT4:> POST https://fah1.innovatr.ca/api/assign HTTP/1.1
06:55:38:I3:Connecting to fah1.innovatr.ca:443
06:55:39:I1::WU1479:DOWNLOAD 94% 2.85MiB of 3.03MiB
06:55:39:I1::WU1479:DOWNLOAD 100% 3.03MiB of 3.03MiB
06:55:39:I1:OUT4:< fah1.innovatr.ca:443 HTTP/1.1 200 HTTP_OK
06:55:39:I1::WU1479:Received WU
06:55:39:I1:Loaded cores/fahcore-a8-lin-64bit-avx2_256-0.0.12/FahCore_a8
06:55:39:I3::WU1479:Running FahCore: /var/lib/fah-client/cores/fahcore-a8-lin-64bit-avx2_256-0.0.12/FahCore_a8 -dir m2FBWVelhW3-aE0kZgRZhAZuKVOnAz2qGeySCBvMnSk -suffix 01 -version 8.1.18 -lifeline 749 -np 1
06:55:39:I3::WU1479:Started FahCore on PID 3877
06:55:40:I1::WU1479:*********************** Log Started 2023-07-05T06:55:39Z ***********************
06:55:40:I1::WU1479:************************** Gromacs Folding@home Core ***************************
06:55:40:I1::WU1479: Core: Gromacs
06:55:40:I1::WU1479: Type: 0xa8
06:55:40:I1::WU1479: Version: 0.0.12
06:55:40:I1::WU1479: Author: Joseph Coffland
06:55:40:I1::WU1479: Copyright: 2020 foldingathome.org
06:55:40:I1::WU1479: Homepage: https://foldingathome.org/
06:55:40:I1::WU1479: Date: Jan 16 2021
06:55:40:I1::WU1479: Time: 19:24:44
06:55:40:I1::WU1479: Compiler: GNU 8.3.0
06:55:40:I1::WU1479: Options: -faligned-new -std=c++14 -fsigned-char -ffunction-sections
06:55:40:I1::WU1479: -fdata-sections -O3 -funroll-loops -fno-pie
06:55:40:I1::WU1479: Platform: linux2 4.15.0-128-generic
06:55:40:I1::WU1479: Bits: 64
06:55:40:I1::WU1479: Mode: Release
06:55:40:I1::WU1479: SIMD: avx2_256
06:55:40:I1::WU1479: OpenMP: ON
06:55:40:I1::WU1479: CUDA: OFF
06:55:40:I1::WU1479: Args: -dir m2FBWVelhW3-aE0kZgRZhAZuKVOnAz2qGeySCBvMnSk -suffix 01
06:55:40:I1::WU1479: -version 8.1.18 -lifeline 749 -np 1
06:55:40:I1::WU1479:************************************ libFAH ************************************
06:55:40:I1::WU1479: Date: Jan 16 2021
06:55:40:I1::WU1479: Time: 19:21:38
06:55:40:I1::WU1479: Compiler: GNU 8.3.0
06:55:40:I1::WU1479: Options: -faligned-new -std=c++14 -fsigned-char -ffunction-sections
06:55:40:I1::WU1479: -fdata-sections -O3 -funroll-loops -fno-pie
06:55:40:I1::WU1479: Platform: linux2 4.15.0-128-generic
06:55:40:I1::WU1479: Bits: 64
06:55:40:I1::WU1479: Mode: Release
06:55:40:I1::WU1479:************************************ CBang *************************************
06:55:40:I1::WU1479: Date: Jan 16 2021
06:55:40:I1::WU1479: Time: 19:21:24
06:55:40:I1::WU1479: Compiler: GNU 8.3.0
06:55:40:I1::WU1479: Options: -faligned-new -std=c++14 -fsigned-char -ffunction-sections
06:55:40:I1::WU1479: -fdata-sections -O3 -funroll-loops -fno-pie -fPIC
06:55:40:I1::WU1479: Platform: linux2 4.15.0-128-generic
06:55:40:I1::WU1479: Bits: 64
06:55:40:I1::WU1479: Mode: Release
06:55:40:I1::WU1479:************************************ System ************************************
06:55:40:I1::WU1479: CPU: Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
06:55:40:I1::WU1479: CPU ID: GenuineIntel Family 6 Model 60 Stepping 3
06:55:40:I1::WU1479: CPUs: 8
06:55:40:I1::WU1479: Memory: 7.69GiB
06:55:40:I1::WU1479:Free Memory: 4.34GiB
06:55:40:I1::WU1479: Threads: POSIX_THREADS
06:55:40:I1::WU1479: OS Version: 5.15
06:55:40:I1::WU1479:Has Battery: false
06:55:40:I1::WU1479: On Battery: false
06:55:40:I1::WU1479: UTC Offset: 1
06:55:40:I1::WU1479: PID: 3877
06:55:40:I1::WU1479: CWD: /var/lib/fah-client/work
06:55:40:I1::WU1479:********************************************************************************
06:55:40:I1::WU1479:Project: 19421 (Run 1, Clone 70, Gen 32)
06:55:40:I1::WU1479:Unit: 0x00000000000000000000000000000000
06:55:40:I1::WU1479:Reading tar file core.xml
06:55:40:I1::WU1479:Reading tar file frame32.tpr
06:55:40:I1::WU1479:Digital signatures verified
06:55:40:I1::WU1479:Calling: mdrun -c frame32.gro -s frame32.tpr -x frame32.xtc -cpt 5 -nt 1 -ntmpi 1
06:55:40:I1::WU1479:Steps: first=40000000 total=41250000
06:55:41:I1::WU1479:Completed 1 out of 1250000 steps (0%)

[toTOW]

Something is wrong with your permissions :

06:54:44:E :Exception: clGetPlatformIDs() returned -1001
06:54:44:E :Exception: cuInit() returned 100

The client can't access OpenCL and CUDA ... which is why it won't use the GPU.

Some distributions requires that you add the user that the client will run under to be added to some specific groups (video, render, ...) that have rights to access the GPU.

[DocJonz]

@toTOW - thanks for the info. I've added 'video' and 'render' as groups, but (following a restart) it has not overcome this issue.

[toTOW]

I'm not sure that Ubuntu was subject to the issue, but there are so many different distributions that I don't remember all cases.

Does the client start automatically with the system or do you start it manually from terminal ?

I know nothing about v8 (except that I would call it alpha and not beta), so if you can't figure it how to make it work, I suggest that you open a ticket here : https://github.com/FoldingAtHome/fah-cl ... tet/issues and roll back to v7 client.

[calxalot]

This seems unusual:

Code: Select all

06:54:44:E :Exception: Could not get IP address for api.foldingathome.org: Temporary failure in name resolution

Are you using a Fritz box?

[DocJonz]

@toTOW - I start it manually when I turn the machine, and use a Firefox browser (as v8 is a webclient). Unfortunately, I can't 'roll back' as the the last few Ubuntu releases don't support Python 2.7 (and it is a real faff to get v7 working), hence using v8, which seems fine when its running.

@ calxalot - no, I'm not using a Fritz box (I had to search for that term!)