the shared folder named Myapplication on the node centralStore is accessible from my node which we call node1 for example , i just miss some details about how to run the dmtcp_coordinator on node1 to store ./dmtcp_restart_script.sh file on central store on the folder Myapplication Basma From: [email protected] Date: Sun, 26 Oct 2014 01:15:29 -0700 Subject: Re: [Dmtcp-forum] checkpoint to a remote location To: [email protected] CC: [email protected]
On Sun, Oct 26, 2014 at 1:05 AM, basma a.azeem <[email protected]> wrote: Hi Kapil, thank you for your reply so i can put the path to shared folder in the checkpoint using the flag to dmtcp_launch like that:if i have a node named centralStore and shared folder on it named Myapplication --ckptdir //centralStore/Myapplicationthis will put the checkpoint files on the shared folder named Myapplication on a node named centralStore Correct. the second part concerning The restart script , i need to run an application on one node lets call it node1 and i need to checkpoint that application but store the checkpoint files and restart script on a second node lets call it central store . so to do this you said that i need to "cd <dirname>; /path/to/dmtcp_coordinator" but the directory here is not on the same machine If the shared folder is not accessible on the node where the dmtcp_coordinator is running, then unfortunately, there is nothing we can do except for copying the restart script by hand. An alternative is to run the dmtcp_coordinator on one of the other nodes where the shared folder is accessible. Basma From: [email protected] Date: Sun, 26 Oct 2014 00:33:16 -0700 Subject: Re: [Dmtcp-forum] checkpoint to a remote location To: [email protected] CC: [email protected] Hi Basma, You can use the "--ckptdir <dirname>" flag to dmtcp_launch to put ckpt images in a certain directory. The restart script is a little trickier. But the simplest way is to do "cd <dirname>; /path/to/dmtcp_coordinator". Dmtcp coordinator put the restart script in the same directory where it was started from, so if you change to your shared directory before running dmtcp_coordinator, everything should work just fine. Kapil On Sat, Oct 25, 2014 at 3:23 PM, basma a.azeem <[email protected]> wrote: Hi How can i checkpoint to a certain shared folder on a certain node on my network i mean to save the checkpoint files and the ./dmtcp_restart_script.sh file on a remote central node during the checkpoint process as if the checkpoint directory is located at a shared folder on a pc on a network thank you ------------------------------------------------------------------------------ _______________________________________________ Dmtcp-forum mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/dmtcp-forum
------------------------------------------------------------------------------
_______________________________________________ Dmtcp-forum mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/dmtcp-forum
