On Fri, Jun 19, 2015 at 07:15:26PM +0300, Cyrill Gorcunov wrote: > On Fri, Jun 19, 2015 at 06:01:11PM +0300, Cyrill Gorcunov wrote: > > Currently in dummy console (84364c191199c50f1bc6fbf7d5f121750c5a1f66) > > there is a single character device for all containers, and due to > > simplicity sake the driver doesn't support multiple VE working at > > same time, which lead that agetty service from container poke same > > console from various VE making it stuck eventually. > > > > So I've had to enhance console to generate one device per VE > > so that every container grab own instance of a virtual console. > > > > Note the console is still dump and doesn't provide a way to > > fetch data written yet, this gonna be addressed in next > > patches. Meanwhile this should prevent containers from > > stucks. > > > > https://jira.sw.ru/browse/PSBM-34357 > > Don't apply it please, seems there is a leak in a patch...
With two containers been running long enough (~40 minutes) I've got a memory leak from ve_legacy_pty_init, so the leak I thought been introduced by this patch is not that, it was there already, so I think we might grab this patch for testing. _______________________________________________ Devel mailing list [email protected] https://lists.openvz.org/mailman/listinfo/devel
