--- Neal Becker <ndbecker2(a)gmail.com> wrote:
One of the features I'd really like to see more widely used in
the
linux
world is checkpoint/restart. I want it for long-running simulations,
although there may be other uses.
If you are willing to live with virtualization, you can launch your
processes under qemu, then use its monitor commands to hibernate the
system to a ram-image file. If the target system is running from ram
(e.g. a livecd), then it is even fairly trivial to migrate the process
to another heterogeneous host. I.e. scp the hibernated qemu ram image
to another host (and the livecd iso) and resume there.
-dmc/jdog
Basic requirement:
1) checkpoint a running process and be able to restart on the same
machine.
Restore open files, including mappings of shared memory.
More advanced:
2) Migrate a process to another machine (probably with same
libraries).
3) Handle multiple threads/processes?
I am pleased to find that blcr
http://ftg.lbl.gov/CheckpointRestart/CheckpointRestart.shtml
seems to do just what I need. It has an srpm and (with minor change)
builds/runs fine of fc6 x86_64.
I have been following this area for last several years. There have
been
several starts at other checkpoint projects, but all except blcr seem
to
have died.
I think this is really exciting technology and I'm betting others
will want
to see this added to Fedora.
--
fedora-devel-list mailing list
fedora-devel-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
____________________________________________________________________________________
Be a PS3 game guru.
Get your game face on with the latest PS3 news and previews at Yahoo! Games.
http://videogames.yahoo.com/platform?platform=120121