Hi,

Has anyone enountered a climate model which Boinc thinks is running, but no CPU time is going to it? In TOP, it shows as status \'D\' (uninterruptible sleep).

The model froze late last night, and I noticed it about 8 hours later. Rebooted the PC, but the shutdown halted as well (so I power cycled it). The model resumed OK after the reboot.


openSUSE 10.2, 64-bit, Intel Q6600, 4GB, 32-bit Boinc Manager V5.10.8.

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

Looks like one of the unexplainable Linux glitches. (Much as it pains me to admit it, since WinXP SP2, I\'ve had more such grief in Linux than XP.) Of course, boinc could also have tied itself into some sort of knot.

I wonder whether ZENworks stuck a toe into the pie... (openSuSE is dumping ZENworks, to which I say \'Yea!\')
[/conjecture]

To answer your question, I\'ve had a Run or three stop processing for unknown reasons, none recently, but don\'t recall seeing status \'D\'.

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

What with my insomnia, I sometimes feel like that I could do with a few nights of status D...

I had a search around with Google, and it seems to be when something calls a Kernal device driver which never responds. Kernal bugs and hardware issues are mentioned as the main causes. Both are bad...

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

Has anyone enountered a climate model which Boinc thinks is running, but no CPU time is going to it? In TOP, it shows as status \'D\' (uninterruptible sleep).

The model froze late last night, and I noticed it about 8 hours later. Rebooted the PC, but the shutdown halted as well (so I power cycled it). The model resumed OK after the reboot.

openSUSE 10.2, 64-bit, Intel Q6600, 4GB, 32-bit Boinc Manager V5.10.8.

If you computer became active or BOINC tried to suspend the workunit at one point with \"leave applications in memory\" = no, then the Defunct or Zombie status you saw was CPDN trying to exit itself. Unfortunately, it got stuck. Best option is to suspend BOINC, then exit BOINC completely.

If the model doesn\'t die after 1 minute, then you can kill -9 <top_pid> where <top_pid> is the process id of the model. In the below example, 6211 is the correct pid:
6211 ? RNl 1:54 hadcm3trans_5.41_i686-pc-linux-gnu
6219 ? TNl 2157:22 hadcm3transum_5.41_i686-pc-linux-gnu

This can corrupt your model, so only do this as a last resort to recovering your system.