Hi,
I have a problem where new clients are starved of work to do. My boinc WU are taking a very short time, less than 30seconds.

Each time the client requests new work the server only seemed to give it 6 new tasks at a time, then eventually stopped giving it anything at all.

max_result for the host is sset to zero.

I'm thinking I may have to wait for some perodic job to happen in the server to update some stats to allow it more work. Is this correct?

My other clients that are very old and have a lot of work to do have no isssues getting as much work as they want.

Any help appreciated.

Cheers
Seth

What project are you talking about?
(BOINC is an environment, not a computational project)

My own project, I dun built one myself ya see!

Yes I understand what BOINC is.

In Any case I fixed the issue with the client not getting enough work initially by tweaking cc_config.xml settings.

However eventually the scheduler stops giving it new tasks to perform. Is it hitting some daily limit?

It's not a quota error I'm getting. simply: "Scheduler request completed: got 0 new tasks"

in my project config.xml I have:
<daily_result_quota>10000</daily_result_quota>

and also:
<max_wus_to_send>200</max_wus_to_send>

Could either of these be impacting a new client?

Cheers
Seth

My own project, I dun built one myself ya see!

Yes I understand what BOINC is.

In Any case I fixed the issue with the client not getting enough work initially by tweaking cc_config.xml settings.

However eventually the scheduler stops giving it new tasks to perform. Is it hitting some daily limit?

It's not a quota error I'm getting. simply: "Scheduler request completed: got 0 new tasks"

in my project config.xml I have:
<daily_result_quota>10000</daily_result_quota>

and also:
<max_wus_to_send>200</max_wus_to_send>

Could either of these be impacting a new client?

Cheers
Seth

You should debug the server to see whats wrong with the client request.

There's nothing wrong with the client requests.

There is documentation in the form of https://boinc.berkeley.edu/trac/wiki/ProjectOptions that explains all that you find in config.xml, you don't have to guess at its meaning.

If still in complete doubt, why not post the contents of your config.xml file?

---

Thanks Jord!

I seem to have overcome the limitations I was experiencing but it's unclear exactly what change I made to the config had an impact. Certainly changing:

<shmem_work_items>1000</shmem_work_items>
<feeder_query_size>2000</feeder_query_size>

Has made a difference plus there was an error in my cc_config.xml that I fixed.

My AWS C4.8xlarge BOINC client instance that was exhibiting the problems previously has been running fine now for over 2 hours so fingers crossed there.

Cheers
Seth

Thanks Jord!

I seem to have overcome the limitations I was experiencing but it's unclear exactly what change I made to the config had an impact. Certainly changing:

<shmem_work_items>1000</shmem_work_items>
<feeder_query_size>2000</feeder_query_size>

Has made a difference plus there was an error in my cc_config.xml that I fixed.

My AWS C4.8xlarge BOINC client instance that was exhibiting the problems previously has been running fine now for over 2 hours so fingers crossed there.

Cheers
Seth

The <feeder_query_size>2000</feeder_query_size> is too high, mine is crashing at 1000

Hi Rebirther,
No crashes for me at 2000.

Cheers
Seth