LHC@home

Since I have been running ATLAS on a 32GB Ramdisk, I have made the experience that download of tasks beyond a certain number is not possible; the BOINC event log always tells that so and so much more disk space is needed.

Example: free space on the Ramdisk: 9.7GB; when pushing the "update" botton for retrieving another task, BOINC says that it needs 383.85 more disk space. This seems totally unlogical to me, since the vm_image.vdi is about 3.2GB. So, how come that 9.7GB free disk space is not enough?
Does anyone have an explanation why this is happening?

You can search it in the forums, it's a old discussion.

You should check what values you have set in Boinc Manager how much disk Boinc is allowed to use and how much it should leave free.

---

You should check what values you have set in Boinc Manager how much disk Boinc is allowed to use and how much it should leave free.

thanks for the hint. The only box in which I set a check from beginning on is: "don't use more than ...% of total". And the value I chose is 98%.
So, in theory, I could increase it to 100%, which in a few cases could help.

So, in theory, I could increase it to 100%, which in a few cases could help.

I did this now. In one case, it helped. In the next case, BOINC tells me that 2,52MB (!) more disk space is needed. How crazy :-(

Had the same message two days ago in a CentOS7 with native.
Made a cvmfs_config wipecache. One time it helped in the other time had to reboot the native-VM.
Since this time it running well.
Nothing changed in Boinc prefs.

So, in theory, I could increase it to 100%, which in a few cases could help.

I did this now. In one case, it helped. In the next case, BOINC tells me that 2,52MB (!) more disk space is needed. How crazy :-(

I suppose, you only have BOINC's data directory on the ramdisk?

I suppose, you only have BOINC's data directory on the ramdisk?

yes, your assumption is correct :-)