LHC@home

Every Atlas-Task in Win11pro have 6.09 GByte file!

I got a few of those as well. None of them running yet.
Most of the tasks are between 400 - 500 MB though.

---

They have all been failing for me. All of them are returning EXIT_DISK_LIMIT_EXCEEDED.

They have all been failing for me. All of them are returning EXIT_DISK_LIMIT_EXCEEDED.

Failing here also with same error.

---

100 Tasks with this Error's overnight.
Every download 6 GByte => 600 GByte for nothing, OMG.

They have all been failing for me. All of them are returning EXIT_DISK_LIMIT_EXCEEDED.

Failing here also with same error.

Linux native, idem.

Bye.

All those monster tasks will fail by the start, cause the files needed are exceeding the 9.3GB allowed to be in one slot directory.
Fortunately they do not notice this disk exceeding at the end of the task.
I will give 4 tasks (two by two) a try with 10 threads after one 'normal' task now running has finished and allowing those monsters more diskspace.

Same here since 12 hours. I run Atlas native with linux mint.
The big ones crashing with this error message:
LHC@home 26.09.2024 13:43:44 Aborting task U85LDmlK0D6nsSi4ap6QjLDmwznN0nGgGQJmDC4LDmbblKDme3lgMm_0: exceeded disk limit: 12452.66MB > 9536.74MB

There are some normal with about 500MB an then these mega-Workunits:

~$ sudo du /var/lib/boinc-client/slots/0 -hs
455M /var/lib/boinc-client/slots/0

:~$ sudo du /var/lib/boinc-client/slots/2 -hs
6,1G /var/lib/boinc-client/slots/2

Only the big ones crashing, I understand i can't do anything about it, right?

I will pause Atlas on my client until this is solved.

These 6 GB-WUs have crashed 4 Hosts of mine ==> ATLAS is paused on all machines until this is solved

---

Supporting BOINC, a great concept !

All those monster tasks will fail by the start, cause the files needed are exceeding the 9.3GB allowed to be in one slot directory.
Fortunately they do not notice this disk exceeding at the end of the task.
I will give 4 tasks (two by two) a try with 10 threads after one 'normal' task now running has finished and allowing those monsters more diskspace.

Just for testing, I started 2 of these tasks with the 6.1GB root-file.
After the input files are copied to the slot-directory and the VM is created, the slot contains 13.490.000.000 bytes and increasing, where max 10.000.000.000 is allowed.
The tasks have the same number of events to process: 400 and processing now after I made an adjustment to BOINC.
Those tasks should not be sent to volunteers.

Those tasks should not be sent to volunteers.

We hoping Cern-IT have someone to stop this 6 GByte Atlas-Tasks.
We need an answer from there, to do the work without this Tasks!

What adjustment?

What adjustment?

rsc_disk_bound in client_state.xml for every (new) workunit.

First two tasks returned successfully:

https://lhcathome.cern.ch/lhcathome/result.php?resultid=414531235
https://lhcathome.cern.ch/lhcathome/result.php?resultid=414531094

Look for the size of ATLAS.root_0 in the results.
Next two are running.

Crystal,
it's ok to find this <rsc_disk_bound>xxxx</rsc_disk_bound>,
but, this is no solution, but a problem for all other user.

Crystal,
it's ok to find this <rsc_disk_bound>xxxx</rsc_disk_bound>,
but, this is no solution, but a problem for all other user.

As I wrote: "Just for testing"