В Пт, 19/06/2015 в 14:39 +0300, Kirill Tkhai пишет: > В Пт, 19/06/2015 в 14:26 +0300, Vladimir Davydov пишет: > > On Fri, Jun 19, 2015 at 02:19:38PM +0300, Kirill Tkhai wrote: > > > В Пт, 19/06/2015 в 14:15 +0300, Vladimir Davydov пишет: > > > > On Fri, Jun 19, 2015 at 02:07:32PM +0300, Kirill Tkhai wrote: > > > > > New vzctl works with UUID, not ID. We should show it. > > > > > > > > Why? AFAIU all the info can be obtained right from the devices cgroup > > > > (please correct me if I'm wrong), so why do we need to duplicate it > > > > here? > > > > > > > > I think the situation with devperms is similar to what we have with > > > > /proc/fairsched - we don't need to show UUID based containers there, > > > > only legacy ones. > > > > > > Isn't UUID the main representation manner? I thought we are going to > > > completely drop legacy IDs from everywhere. > > > > Current plan: > > > > 1) If there is already upstream API for somehing duplicating our > > internal API - we only leave our internal API for legacy containers. > > For UUID-based containers one should use upstream API. This is the case > > in case of /proc/fairsched and /proc/vz/devperms. > > > > 2) If there is no upstream API for something we need, we continue to use > > our legacy API wherever possible and avoid introducing new API unless we > > can push it upstream. This is the case in case of all the legacy ioctls. > > CC: Dmitry Mishin >
So, I leave https://jira.sw.ru/browse/PSBM-34317 as INVALID. _______________________________________________ Devel mailing list [email protected] https://lists.openvz.org/mailman/listinfo/devel
