Bogus error: Not enough Disk space?

Stick
Stick
Joined: 24 Feb 05
Posts: 790
Credit: 4567876
RAC: 2237

RE: I solved the problem

Message 20149 in response to message 20148

Quote:
I solved the problem :)
You won't believe it: http://einsteinathome.org/goto/comment/20156

Uli,

Glad to hear it!! But, it makes me wonder if the fix will last. That is, as you know doubt read, several people reported "fixing" similar problems by just "reloading" their existing preferences. This had us speculating the cause could be either a "bug" in the BOINC client or (as Gary Roberts thinks) just random interactions between BOINC checkpointing and other (non-BOINC) computer disk operations (probably involving large temp files). We also believe, that if Gary is correct, the problem is likely to surface again for those who just "reloaded existing prefs". Since you didn't change your prefs by a significant amount, you may also be vulnerable.

As it stands now, I subscribe to Gary's theory (I didn't initially) - but we still don't have enough good evidence to prove it. I think you may be the first to report the problem while running an older BOINC version (not BOINC 5.2.X). That tends to discount the new BOINC "bug" theory. So you may have helped out!

Again, I apologize for my earlier antagonistic post. I hope we can put this episode behind us and work amicably together in the future.

Thank you and good luck on your "fix"!

Stick

Bruce Allen
Bruce Allen
Moderator
Joined: 15 Oct 04
Posts: 1113
Credit: 172127663
RAC: 0

RE: What ist this? I'm

Quote:
What ist this? I'm running BOINC (SETI, LHC and Einstein) on a 160 MByte ramdisk since more than a year. Suddenly it complains about not enough disk space? What have you changed on your end, cause i didn't do anything?

Ulrich, you have myself (and Walt Gribben) to blame for your problems! A few days ago Walt pointed out a bug in the scheduler code that checks disk allocations, and suggested a fix. i applied his fix to Einstein@Home right away, but with the following comment boinc checkin_notes.

This is completely consistent with what you observed. Set 'leave X free' to 0 and you get minimum 100 free. Set it to 0.001 and you get something smaller.

Sigh.

Bruce

Director, Einstein@Home

Walt Gribben
Walt Gribben
Joined: 20 Feb 05
Posts: 219
Credit: 1645393
RAC: 0

RE: RE: What ist this?

Message 20151 in response to message 20150

Quote:
Quote:
What ist this? I'm running BOINC (SETI, LHC and Einstein) on a 160 MByte ramdisk since more than a year. Suddenly it complains about not enough disk space? What have you changed on your end, cause i didn't do anything?

Ulrich, you have myself (and Walt Gribben) to blame for your problems! A few days ago Walt pointed out a bug in the scheduler code that checks disk allocations, and suggested a fix. i applied his fix to Einstein@Home right away, but with the following comment boinc checkin_notes.

This is completely consistent with what you observed. Set 'leave X free' to 0 and you get minimum 100 free. Set it to 0.001 and you get something smaller.

Sigh.

Bruce

Its my fault more than yours, Bruce, you questioned that line but I didn't get back with a reply until just recently.

Very sorry Ulrich, I try to not make destrutive changes but that one slipped by.

Walt

Ulrich Metzner
Ulrich Metzner
Joined: 22 Jan 05
Posts: 113
Credit: 963370
RAC: 0

Thank you sirs, no damage

Thank you sirs, no damage taken :)

Aloha, Uli

Stick
Stick
Joined: 24 Feb 05
Posts: 790
Credit: 4567876
RAC: 2237

RE: As it stands now, I

Message 20153 in response to message 20149

Quote:
As it stands now, I subscribe to Gary's theory (I didn't initially) - but we still don't have enough good evidence to prove it. I think you may be the first to report the problem while running an older BOINC version (not BOINC 5.2.X). That tends to discount the new BOINC "bug" theory. So you may have helped out!

On the other hand, the "bug" theory still looks pretty good. ;-)

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.