Re: [Dmtcp-forum] Killing node causes major problems

[<Dar...@gm...>]

Thanks for the response Kapil. I tried installing the deb in an Ubuntu  
10.10 virtual machine and reproduce but I too was unable to do it in the  
virtual machine. So I tried again on my actual Ubuntu install and it still  
is having the same problem. I'm not sure what's different but I can give  
you as much detail as you need and hopefully you can see where the problem  
might be occurring. I can do some tests for you if you want.

Here is a little more detail from the latest test. I ran dmtcp_coordinator  
again, then ran "dmtcp_checkpoint vi asdf" in another window. Then I went  
back to the coordinator and did checkpoint. It printed:

[6590] NOTE at dmtcp_coordinator.cpp:985 in startCheckpoint;  
REASON='starting checkpoint, suspending all nodes'
s.numPeers = 1
[6590] NOTE at dmtcp_coordinator.cpp:477 in onData; REASON='locking all  
nodes'
[6590] NOTE at dmtcp_coordinator.cpp:509 in onData; REASON='draining all  
nodes'

and that's all the coordinator printed.

In the main VI window I got:

[6697] WARNING at kernelbufferdrainer.cpp:100 in onTimeoutInterval;  
REASON='JWARNING(false) failed'
_dataSockets[i]->socket().sockfd() = 5
buffer.size() = 24
WARN_INTERVAL_SEC = 10
Message: Still draining socket... perhaps remote host is not running under  
DMTCP?

and this just keeps repeating over and over. Like I said before, the system  
then gets into a weird state where I can't launch applications from Gnome,  
and even running vi from the command line hesitates a lot longer than  
normal before continuing. I ended up just rebooting.

There's not really anything unusual about the system that I know of. Let me  
know if you need any other data or have any tests for me to run. Thanks,

Nick



On Apr 11, 2011 11:03am, Kapil Arya <ka...@cc...> wrote:
> Hello Nick,

> I tried but failed to reproduce the problem. I used Ubuntu 10.10 and  
> tried using the svn version as well as the .deb package, but none of them  
> caused the problems that you are talking about. Is there any specific  
> setup that you are using? We haven't seen dmtcp exhibiting such behavior  
> in past. If you can tell us more about your setup, it can help us  
> reproduce the bug.




> We can also try some online collaboration tool, or us getting a guest  
> account on the machine you are trying to checkpoint to expedite the  
> process. Also, if it works, we can setup some time and talk over phone  
> and/or some messenger/irc.




> We definitely want to find and solve this issue and thanks again for  
> reporting this to us.


> --Kapil

> On Fri, Apr 8, 2011 at 4:27 PM, Kapil Arya ka...@cc...> wrote:


> Hello Nick,

> Thanks for reporting the bug. It looks pretty bad. I am looking into this  
> issue right now and will write back as soon as there is more.




> Thanks,
> --Kapil



> <> div>On Thu, Apr 7, 2011 at 4:32 PM, Dar...@gm...> wrote:






> I just installed DMTCP for the first time and am evaluating it. I'm using  
> version 1.2.1-1 X86_64 on Ubuntu 10.10.

> As a trial, I ran "dmtcp_coordinator" in one window and then in another  
> window ran "dmtcp_checkpoint vi asdf" and typed some data into it. I then  
> went back to the coordinator window and pressed "c" to checkpoint, and  
> then pressed "k" to kill the vi session (I intended that I would load the  
> vi state from the checkpoint).




> However, at this point things went screwy. The vi window displayed the  
> following:

> WARNING at kernelbufferdrainer.cpp:100 in onTimeoutInterval;  
> REASON='JWARNING(false) failed Message: Still draining socket... perhaps  
> remote host is not running under DMTCP




> and repeated this periodically. After doing this I became unable to  
> launch any new processes from Gnome, and when I tried to start vi from an  
> existing terminal it sat for several seconds and then gave an unusual  
> error message.




> I rebooted and everything was back to normal. I didn't not try logging  
> off and back on -- perhaps that would have fixed it as well.

> After I rebooted I tried the same steps again and it failed in the same  
> way again, prompting me to reboot. So its reproducible. Does anyone know  
> what is going on? Thanks,




> Nick


> ------------------------------------------------------------------------------

> Xperia(TM) PLAY

> It's a major breakthrough. An authentic gaming

> smartphone on the nation's most reliable network.

> And it wants your games.

> http://p.sf.net/sfu/verizon-sfdev
> _______________________________________________

> Dmtcp-forum mailing list

> Dmt...@li...

> https://lists.sourceforge.net/lists/listinfo/dmtcp-forum