Both rkt and docker run commands get very long very quickly. That's normal (and why such things should go in a definition file... such as systemd units). To answer your question, though, if you want multiple containers in the same pod, they all need to be listed in the same run command.
On Mon, Jul 18, 2016 at 12:52 PM Derek Mahar <[email protected]> wrote: > All of them? Wouldn't this create a very long and complicated 'rkt run' > command and systemd unit? How could I split the single 'rkt run' command > into multiple commands? At the moment, I think the answer to this question > is, "You can't." > > > On Monday, 18 July 2016 12:22:52 UTC-4, Florian Koch wrote: > >> why not run these apps in a pod? >> >> 2016-07-18 17:00 GMT+02:00 Derek Mahar <[email protected]>: >> > On Friday, 4 March 2016 14:18:59 UTC-5, Alex Crawford wrote: >> >> >> >> On 03/04, [email protected] wrote: >> >> > 1) Does either or both etcd and/or fleet actually need to be running >> on >> >> > a >> >> > standalone host? Are they so integral to CoreOS that they must be >> used? >> >> >> >> Nope. I have a standalone host that runs a few services. I don't >> actually >> >> use >> >> etcd or fleet since there is no cluster. >> > >> > >> > Do you run your services in Docker or rkt containers? If using rkt, >> what >> > kind of networking do you use and how do you configure your containers >> to >> > talk to each other? >> > >> > I wish to run rkt containers for vpnc, ActiveMQ, PostgreSQL server, >> Tomcat, >> > and a standalone Java application on a single CoreOS host, but am >> unsure how >> > the Tomcat and the Java application containers can address the shared >> > ActiveMQ and PostgreSQL server containers. A user on #coreos has told >> me >> > that I must use service discovery, but this strikes me as overkill for >> a few >> > containers that run on a single host. >> > >> > Any suggestions? >> > >> > Derek >> > -- Seán C McCord CyCore Systems, Inc +1 888 240 0308 PGP/GPG: http://cycoresys.com/scm.asc
