F@Hclient not folding after adding SSD

[Zac67]

Yesterday, I added another NVMe SSD to my machine (Win 10 22H2) and today it's not folding any more.

https://v8-4.foldingathome.org/ Web Control shows "No resources". Pausing and resuming the machine changes the play/pause button, but nothing else happens. The log shows normal startup, play/pause activity but nothing else.

Comparing with previous states, the GPU appears to have moved from 9:0:0 to 10:0:0:

Code: Select all

11:41:46:I3:gpus = {
11:41:46:I3: "gpu:10:00:00": {
11:41:46:I3: "vendor": 4318,
11:41:46:I3: "type": "nvidia",
11:41:46:I3: "description": "NVIDIA GeForce RTX 4070 Ti SUPER",
11:41:46:I3: "uuid": "xxxxxxxxxxxxxxxxxxxxx",
11:41:46:I3: "opencl": {"platform": 0, "device": 0, "compute": "3.0", "driver": "566.36"},
11:41:46:I3: "cuda": {"platform": 0, "device": 0, "compute": "8.9", "driver": "12.7"},
11:41:46:I3: "device": 9989,
11:41:46:I3: "supported": true
11:41:46:I3: } 

Entering configuration of the default resource group, the GPU is displayed, but the Enabled box is empty. I can tick it but the Save button on top stays greyed out. Even when I increase CPU cores the Save button stays inactive. Even adding yet another resource group and trying to assign resource there doesn't change that.

This actually happened with 8.3, upgrading to 8.4 didn't help either. I've even edited client.db/groups and changed 9:0:0 to 10:0:0 which didn't change a thing. lufah . groups shows [""].

[calxalot]

Not your issue, but please upgrade lufah.

Peer argument has become an address option. Misc bugs were fixed.

[calxalot]

Are you using https://v8-4.foldingathome.org/ ?

What OS and version?

Folks that can help will want to see more of the log, including the first 200 lines and any error messages.

If you are running the debug build, please redact any instance of your passkey and anything else you consider confidential.

[calxalot]

Also look for errors in your browser JavaScript console.

[Zac67]

Added missing info to the post above. The browser console shows only errors related to the team logo that can't be found.

lufah upgraded to 0.8.2 - lufah state shows

Code: Select all

{
  "info": {
    "version": "8.4.9",
    "url": "https://v8-4.foldingathome.org/",
    "os": "win32",
    "os_version": "6.2",
    "cpu": "amd64",
    "cpu_brand": "AMD Ryzen 7 5800X 8-Core Processor",
    "cpus": 16,
    "gpus": {
      "gpu:10:00:00": {
        "vendor": 4318,
        "type": "nvidia",
        "description": "NVIDIA GeForce RTX 4070 Ti SUPER",
        "uuid": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
        "opencl": {
          "platform": 0,
          "device": 0,
          "compute": "3.0",
          "driver": "566.36"
        },
        "cuda": {
          "platform": 0,
          "device": 0,
          "compute": "8.9",
          "driver": "12.7"
        },
        "device": 9989,
        "supported": true
      }
    },
    "mach_name": "xxxxxxxxxxxxxxxx",
    "hostname": "xxxxxxxxxxxx",
    "mode": "Release",
    "revision": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
    "has_battery": false,
    "on_battery": false,
    "id": xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
    "account": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
  },
  "config": {
    "user": "NilsP",
    "team": 45456,
    "passkey": "xxxxxxxxxxxxxxxxxxxxx",
    "cause": "alzheimers"
  },
  "groups": {
    "": {
      "config": {
        "on_idle": false,
        "on_battery": true,
        "keep_awake": true,
        "paused": false,
        "finish": false,
        "beta": false,
        "cuda": true,
        "key": 0,
        "cpus": 0,
        "gpus": {
          "gpu:09:00:00": {
            "enabled": true
          }
        }
      }
    }
  },
  "units": []
}

[calxalot]

I see the default group has the old gpu listed.

Please try
lufah enable-all-gpus

Then reload browser page

[Zac67]

Code: Select all

lufah enable-all-gpus
Error: an existing group must be specified for enable-all-gpus

I'm not sure how to specify the default group... lufah groups still comes up emtpy.

[calxalot]

I’d there is only one group, you shouldn’t need to specify.
Otherwise, the default group is specified by slash (group name is “”)

lufah -a / enable-all-gpus

[calxalot]

Maybe I have a bug

[Zac67]

OK, lufah -a / enable-all-gpus did seem to work. Status shows the default (only) group with another gpu:10:00:00.

Aaaand it's folding again! The gpu:09:00:00 is still present in the default group but apparently that doesn't hurt. Web Control only shows one GPU, enabled.

Thank you!

[calxalot]

Is your ssd using a pci slot, or was it an open M.2 slot?

Did you physically move the GPU to another pci slot?

[muziqaz]

Hm, I'll have to play around with my GPUs to try replicate the issue.

[Zac67]

The new SSD is an M.2 NVMe. After all, M.2 is just another PCIe slot form factor and strangely, it seems to be enumerated before the PEG slot on this board (Asus TUF Gaming X570-Plus). The new SSD went to the CPU-connected slot, the chipset-connected slot was already populated before.

Since fitting another SSD is not really a rare use case, it might be smart to stop differentiating between GPUs and slots if there's just a single one.

[muziqaz]

MY PCs have m.2 nvme ssds connected to CPU slots. All of them. And I had no problems whatsoever. have you tried reinstalling fahclient? Since in my case, PCs were already assembled before fahclient was installed.