Time for me to give up? 8 days with no assignment

[malkie]

I think this is related to a recent topic: fah-client 8.4.9: HTTP_SERVICE_UNAVAILABLE ("No appropriate assignment")

I don't recall how long I've been a FAH donor, but, for reference, my stats are:

• 12,054,905 points earned
  426 WUs completed
  Active Clients
  0 active clients within 7 days
  1 active clients within 50 days

A week ago I had to restart my Win 11, AMD Ryzen 5 3400G with gfx902 - only the GPU available to FAH. As usual after a restart, had to reinstall the FAH client. In the past it has taken somewhere between 10 and 25 retries for me to get working on a work unit. However, I'm now experiencing over 100 retries with no assignment.

Today, in an attempt to get something going, I dedicated 1 CPU in addition to the GPU. In the past I've found that more than 1 CPU assigned to FAH causes a noticeable slowdown on my machine, which is why I've been giving FAH only the GPU - for a long time now, probably over a year.

Here's where I am now:

• 17:54:04:I1:WU497:Requesting WU assignment for user Malkie team 0
  17:54:05:I1:WU497:Retry #20 in 8 mins 32 secs

Is it possible that there are simply no work units that can use my machine? Is this a sign that I should find a different project to volunteer my machine time to?

Sorry if I'm just not savvy enough to be sure if the above linked thread contains the answer to my question.

[Joe_H]

There are plenty of CPU WUs that your machine would be able to process. You just need to set it to use 3 or more CPU threads. As for your iGPU, the included RX Vega 11 is borderline at this point for processing the available GPU projects.

The GPU processing is assumed to use 1 CPU thread by the client. You do not need to configure a CPU thread for it, and in the case of the v8 client doing so can trigger an assignment bug if there are no GPU WUs available.

My recommendation would be to disable the GPU in the client and fold on your CPU. As I mentioned earlier, it will need at least 3 threads assigned. The client is set to run CPU folding at the lowest priority, it shouldn't slow down your machine. What may actually have been slowing your machine down was doing both CPU and GPU processing. The iGPU and CPU share the same thermal environment in the processor packaging, both generating heat can cause down clocking to remain within the TDP set for the processor. That is either 65 W or 35 W depending on settings for the hardware the 3400G is installed in.

As for having to reinstall every time after a restart, that is not normal. I don't work with Windows these days, so someone else may have some idea on what you need to do to stop reinstalling every time.

[malkie]

Thanks, Joe - I appreciate your help.

I've done as you suggested: disabled the GPU, and assigned 3 CPUs. A WU downloaded almost immediately, and is working away.

If there's anything further that I think is significant, I'll comment again. But for the moment all seems to be good.

Any idea, though, why I was previously able to complete hundreds of WU on the GPU alone, but not at all since about a week ago?

[Joe_H]

My best guess would be that projects that could use the Vega 11 finished or their servers are offline for one reason or another. I don't follow what projects come out or finish closely. Possibly some projects will start up in the future that can use the Vega 11, hard to tell at this point.

[malkie]

Got it! Thanks again, Joe.

[malkie]

After a couple of days and a couple of WUs, I'm seeing
OUT117:HTTP_SERVICE_UNAVAILABLE: {"error":"No appropriate assignment"}, and
OUT119:HTTP_INTERNAL_SERVER_ERROR: {"error":{"message":"DB ERROR: IO error: log-v2.leveldb/068411.log: No space left on device"}}

In addition I'm finding that since assigning 3 of my 8 Ryzen 3400 cores to FAH, there is a noticeable slowdown on almost everything I do.

My PC needs to restart for some Win11 updates, so I'll see what happens after I reinstall the FAH client once again.

[arisu]

After a couple of days and a couple of WUs, I'm seeing
OUT117:HTTP_SERVICE_UNAVAILABLE: {"error":"No appropriate assignment"}, and
OUT119:HTTP_INTERNAL_SERVER_ERROR: {"error":{"message":"DB ERROR: IO error: log-v2.leveldb/068411.log: No space left on device"}}

In addition I'm finding that since assigning 3 of my 8 Ryzen 3400 cores to FAH, there is a noticeable slowdown on almost everything I do.

My PC needs to restart for some Win11 updates, so I'll see what happens after I reinstall the FAH client once again.

There are currently three ongoing issues, all of which are being worked on:

1. A GPU WU shortage - This is because we've temporarily run out of big GPU projects. Some new ones are in the pipeline and will be ready soon.
2. One of the assignment servers running out of space - The person in charge of the assignment servers has been notified and will fix this.
3. A problem with assigning CPU WUs, despite there being no CPU WU shortage - No idea what's causing this. I've reported it internally.

We don't know if #2 and #3 are related. But there's no reason to give up. It's just a hiccup! There are still multiple researchers using and benefiting from FAH, and they're being kept in the loop as well.

[malkie]

After a couple of days and a couple of WUs, I'm seeing
OUT117:HTTP_SERVICE_UNAVAILABLE: {"error":"No appropriate assignment"}, and
OUT119:HTTP_INTERNAL_SERVER_ERROR: {"error":{"message":"DB ERROR: IO error: log-v2.leveldb/068411.log: No space left on device"}}

In addition I'm finding that since assigning 3 of my 8 Ryzen 3400 cores to FAH, there is a noticeable slowdown on almost everything I do.

My PC needs to restart for some Win11 updates, so I'll see what happens after I reinstall the FAH client once again.

There are currently three ongoing issues, all of which are being worked on:

1. A GPU WU shortage - This is because we've temporarily run out of big GPU projects. Some new ones are in the pipeline and will be ready soon.
2. One of the assignment servers running out of space - The person in charge of the assignment servers has been notified and will fix this.
3. A problem with assigning CPU WUs, despite there being no CPU WU shortage - No idea what's causing this. I've reported it internally.

We don't know if #2 and #3 are related. But there's no reason to give up. It's just a hiccup! There are still multiple researchers using and benefiting from FAH, and they're being kept in the loop as well.

Thanks, arisu.

I'll keep an eye on the forum for news about fixes, but in the meantime I'm pausing my participation due to the performance hit I'm seeing. I should probably replace my machine - it's more than 5 years old - but the budget for elective updates is a bit low.

[muziqaz]

There are no new projects coming out which support Vega11.
Maybe when HIP hits the street

[arisu]

Thanks, arisu.

I'll keep an eye on the forum for news about fixes, but in the meantime I'm pausing my participation due to the performance hit I'm seeing. I should probably replace my machine - it's more than 5 years old - but the budget for elective updates is a bit low.

Issue #2 has been fixed and the assignment server has more space available now. #1 will be resolved soon as well.