Facing similar error issues. Around 90% of the results are showing up as errors for me. I have a total of 47errors for African Rainfall Project in the past few days. Previously when IBM hosted the project I faced may be 1 out of 50 error results but never on this scale.

What sort of errors? And how long does it usually take to run an ARP1 task on your system(s)?

if you are suffering file download errors (file size error or checksum error, not transient errors) there's not a lot that can be done about that, but it should be possible to evade some other errors...

If you are getting lots of errors where the only text in the report sent to WCG is a line describing the client in use, that probably means that the task hadn't started before the deadline so it cancels it and sends a "Not started by deadline" message to the server. As there might be [long] delays downloading work at present and the deadline is based on request time, not file receipt time, it may be a good idea to shrink your work cache size to reduce the likelihood of having jobs unable to start in time, or alter your device profile(s) on the server to reduce the total number of Arp1 tasks it will send to a system.

If you are getting application failures (SIGSEGV or SIGILL on Linux; Access violation or Illegal instruction [I think] on Windows) that's a completely different matter, and if you give more details someone might be able to help...

Cheers - Al.

Each task takes a little over a day to complete. I'm running the tasks on an i7 3610qm,it has 4 cores and 8 threads and 8GB RAM. It completes 8 tasks per day and I'm on ubuntu 20.04 lts.

Its the task that has been downloaded and run for a day and completed successfully and uploaded to WCG is showing up as "Error" under status of WCG website's results page.So no credit is provided for the completed task.

Well, one can't possibly run 8 ARP1's at a time on a 8 GB RAM computer, without getting into serious resource trouble. Check the System Requirements on the following link. Each ARP1, can use as much as 1 GB of Memory. The rest of the system needs memory too, of course.
With 8 ARP1 at a time on a 8 GB computer, there's got to be some serious swapping going on all the time, and errors will occur, I'm sure.
https://www.worldcommunitygrid.org/help/topic.s?shortName=minimumreq

Thanks for the reply.I checked now and it says download error,might be bandwidth issues and the website is also having a few bugs as valid tasks seems to disappear from the list in results page after a few days.The error message was for the tasks that weren't downloaded properly and not for the completed tasks I guess.Completed tasks might have become valid and disappeared from results page.
Here is the report-
<core_client_version>7.16.6</core_client_version>
<![CDATA[
<message>
WU download error: couldn't get input files:
<file_xfer_error>
<file_name>ARP1_0004208_134_ARP1_0004208_input_d02</file_name>
<error_code>-200 (wrong size)</error_code>
</file_xfer_error>
</message>
]]>

I'm running 8 tasks and my RAM usage is around 6.2GB out of 7.7GB available and there seems to be no issue with L2 cache usage as well( checked it using cachestat). So I thought there is no issue in running 8 tasks.I have been running for years like this and I have not faced any issues.I also run climateprediction projects as well.When the tasks are of lower resolution I run 8 of those tasks as well and when the resolution of the tasks are higher I run 4 tasks.