Message boards : Theory Application : New version 263.90
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5

AuthorMessage
Toby Broom
Volunteer moderator

Send message
Joined: 27 Sep 08
Posts: 810
Credit: 655,511,100
RAC: 208,123
Message 39898 - Posted: 10 Sep 2019, 19:57:37 UTC

I just make them 8 core WU for the time being.
ID: 39898 · Report as offensive     Reply Quote
Luigi R.
Avatar

Send message
Joined: 7 Feb 14
Posts: 99
Credit: 5,180,005
RAC: 0
Message 39968 - Posted: 19 Sep 2019, 6:50:24 UTC

https://lhcathome.cern.ch/lhcathome/result.php?resultid=245995393

Why did this wu fail, why did VM prematurely shut down and why not getting credits for this reason?
ID: 39968 · Report as offensive     Reply Quote
Luigi R.
Avatar

Send message
Joined: 7 Feb 14
Posts: 99
Credit: 5,180,005
RAC: 0
Message 39969 - Posted: 19 Sep 2019, 8:58:44 UTC

I guess maybe Virtualbox crashed.
ID: 39969 · Report as offensive     Reply Quote
maeax

Send message
Joined: 2 May 07
Posts: 2125
Credit: 159,968,505
RAC: 38,628
Message 39970 - Posted: 19 Sep 2019, 9:27:28 UTC

When you take a deeper look to this task, slot1 stopped Hours earlier than slot2.
Would change the preferences to 1 Cpu instead of 2.
ID: 39970 · Report as offensive     Reply Quote
Luigi R.
Avatar

Send message
Joined: 7 Feb 14
Posts: 99
Credit: 5,180,005
RAC: 0
Message 39971 - Posted: 19 Sep 2019, 9:58:33 UTC - in response to Message 39970.  

It's not a real problem. That host has got 4 cores (8 threads) and 8GB RAM. It can run successfully 5, maybe 6, 1-cpu tasks.
So I set it to run 4x2-cpus VMs to use less RAM.
The best case: all the VMs run at 200% and it uses 8 threads and 4 cores
The worst case: all the VMs run at 100% and it uses 4 threads and 4 cores.
It's working good.
ID: 39971 · Report as offensive     Reply Quote
Erich56

Send message
Joined: 18 Dec 15
Posts: 1693
Credit: 104,864,927
RAC: 69,983
Message 39972 - Posted: 19 Sep 2019, 18:29:07 UTC - in response to Message 39971.  

for several hours now, all Theory tasks on several of my machines (1-core and 2-core) are running idle. No CPU usage at all. What kind of problem do we have here?
ID: 39972 · Report as offensive     Reply Quote
Erich56

Send message
Joined: 18 Dec 15
Posts: 1693
Credit: 104,864,927
RAC: 69,983
Message 39973 - Posted: 19 Sep 2019, 19:25:23 UTC - in response to Message 39972.  

for several hours now, all Theory tasks on several of my machines (1-core and 2-core) are running idle. No CPU usage at all. What kind of problem do we have here?
one of these tasks has now been runnig for about 28 hours (and idle for the last 6-8 hours), another one for 25 hours. I guess they will fail at some point, won't they?
ID: 39973 · Report as offensive     Reply Quote
Harri Liljeroos
Avatar

Send message
Joined: 28 Sep 04
Posts: 682
Credit: 43,874,454
RAC: 16,510
Message 39974 - Posted: 19 Sep 2019, 19:45:15 UTC - in response to Message 39973.  

I think that the cut-off time is nowadays 36 hours.
ID: 39974 · Report as offensive     Reply Quote
Luigi R.
Avatar

Send message
Joined: 7 Feb 14
Posts: 99
Credit: 5,180,005
RAC: 0
Message 39975 - Posted: 19 Sep 2019, 19:51:55 UTC - in response to Message 39973.  
Last modified: 19 Sep 2019, 20:09:02 UTC

Some theory tasks (idle or not) automatically end after 36 hours.

Otherwise, try this:
1) Set Leave applications in memory while suspended unckecked

2) Suspend your idle tasks, check task slot number (N=0,1,2, etc...)

3a) Open your_boinc_data_directory/slots/N/vbox_checkpoint.xml, replace elapsed_time value with "129570.000000", then save
3b) Open your_boinc_data_directory/slots/N/boinc_task_state.xml, replace checkpoint_elapsed_time value with "129570.000000", then save


4) Resume your idle tasks

They should end within 30 seconds.

P.S. I don't know if you can skip some steps of my procedure. It always worked, so I didn't modify it.
ID: 39975 · Report as offensive     Reply Quote
Erich56

Send message
Joined: 18 Dec 15
Posts: 1693
Credit: 104,864,927
RAC: 69,983
Message 39977 - Posted: 20 Sep 2019, 4:51:07 UTC - in response to Message 39975.  

Luigi, thanks for the instructions - they worked well :-)
ID: 39977 · Report as offensive     Reply Quote
Previous · 1 · 2 · 3 · 4 · 5

Message boards : Theory Application : New version 263.90


©2024 CERN