Message boards :
Number crunching :
Does VM task postponed situation also prevent new task download?
Message board moderation
| Author | Message |
|---|---|
 Send message Joined: 28 Sep 04 Posts: 675 Credit: 43,551,717 RAC: 15,544   
|
I have a Theory VM task in a postponed status with message: "task postponed 86400.000000 sec: VM environment needed to be cleaned up." This was due to error 197 situations I had previously which seems to be fixed now by changing the number of max CPUs in my web preferences. Does this somehow also prevent Boinc to requesting work for LHC same way as manually suspending a task does? Boinc says that CPU cache is full but there are several CPU cores idle doing nothing. It doesn't request even sixtrack work (Theory and sixtrack are the only allowed subprojects currently on this host). I only run LHC and CPDN on the CPU, CPDN is set for "No new tasks" and running currently 5 tasks (I normally allow only 1 CPDN task at the time but because of the idle CPUs, I let it run now all that I have). The situation doesn't change if I suspend CPDN completely and force LHC to update. Suspending a project (or individual tasks) removes those tasks from the cache calculation. It will take about 5 hours before the postpone expires so I'll see in the morning what is the situation. [edit]This seems to be the case, I aborted the postponed task and immediately Boinc downloaded new Theory and sixtrack tasks.[/edit] 
|
 Send message Joined: 15 Jun 08 Posts: 2401 Credit: 225,574,165 RAC: 121,000 |
... VM environment needed to be cleaned up. You may also check your VirtualBox manager for suspect VM entries that point to non existing VMs. |
|
Send message Joined: 28 Sep 04 Posts: 675 Credit: 43,551,717 RAC: 15,544
|
... VM environment needed to be cleaned up. That is what I did and removed those vmi files that were aborted because of the EXIT_TIME_LIMIT_EXCEEDED error. Some of them get aborted gracefully and these show "Aborted" and 0 seceonds runtime here on the website even when they actually did run about 14.5 hours in my case, these are removed from the VM manager automatically. But if they show on the website "Error while computing" and the error is EXIT_TIME_LIMIT_EXCEEDED, then they show the actual runtime on the website and these tasks are not automatically removed from VM manager. You have to remove them manually to get things going again.
|
©2024 CERN