Another approach would be to create a bind mount from the host to each container (say `/var/cache/data`). The first executor can copy data there and the subsequent executors can check for it's presence and re-use that data.
On Fri, Mar 3, 2017 at 8:32 AM, tommy xiao <xia...@gmail.com> wrote: > create a persistent disk to store the distribution and run executor > isolation. > > 2017-03-03 21:40 GMT+08:00 Egor Ryashin <egor.ryas...@metamarkets.com>: > >> Hi All, >> >> I'm writing custom scheduler which will be sending short-running tasks. >> I need those tasks to be properly isolated, that means I shouldn't run >> them in the same executor. Those tasks require large binary distributives >> and I suppose each run in a separate executor will spawn large sandboxes >> with copies of those distributives on disk. Is there an easy way to >> maintain isolation for those tasks meanwhile sharing a distributive between >> them? >> >> Thanks, >> Egor >> >> > > > -- > Deshi Xiao > Twitter: xds2000 > E-mail: xiaods(AT)gmail.com > > -- > Zameer Manji > <http://gmail.com> >
