Messages in this thread | | | Date | Tue, 21 Oct 2008 22:55:13 -0400 | From | Daniel Jacobowitz <> | Subject | Re: [RFC v7][PATCH 2/9] General infrastructure for checkpoint restart |
| |
On Tue, Oct 21, 2008 at 09:33:19PM -0400, Oren Laadan wrote: > >> What happens if I pass it a pid of a process which I _do_ own, but it > >> does not refer to a container's init process? > > > > I would assume that do_checkpoint() would return -EINVAL, but it's a > > great question: Oren, did you have another plan? > > Since we intentional provide minimal functionality to keep the patchset > simple and allow easy review - we only checkpoint one task; it doesn't > really matter because we don't deal with the entire container. > > With the ability to checkpoint multiple process we will have to ensure > that we checkpoint an entire container. I planned to return -EINVAL if > the target task isn't a container init(1). Another option, if people > prefer, is to use any task in a container to "represent" the entire > container.
I haven't been following - but why this whole container restriction? Checkpoint/restart of individual processes is very useful too. There are issues with e.g. IPC, but I'm not convinced they're substantially different than the issues already present for a container.
-- Daniel Jacobowitz CodeSourcery
| |