Quoting Paul Menage ([EMAIL PROTECTED]): > On 7/11/07, Takenori Nagano <[EMAIL PROTECTED]> wrote: > >Hi, > > > >I think Balbir's idea is very simple and reasonable way to develop per > >container > >swapping. Because kernel needs the information that a target page belongs > >to > >which container. Fortunately, we already had page based memory management > >system > >which included in RSS controller. I think it is appropriate that we > >develop per > >container swapping on page based memory management system. > > > > So what about people who want to do per-container swapping without > using a page-based memory controller?
Hmm, yes, I'm a few steps behind, but will what Balbir is suggesting prevent me from just using a per-container swapfile for checkpointing an application? (where the checkpoint consists of kicking out the dirty pages to swap and making a new copy-on-write version of the swapfile) thanks, -serge _______________________________________________ Devel mailing list [email protected] https://openvz.org/mailman/listinfo/devel
