[lxc-devel] LXC system UUID

2012-06-05 Thread akrus
Hi, Currently we’re creating a new module for FusionInventory project to support LXC containers. It seems the problem is LXC doesn’t really have a system UUID. Is there any option to set container’s UUID? Or are there any plans to support such option? I’ve found somewhere libvirt creates a new

[lxc-devel] Does lxc support unionfs ?

2012-06-05 Thread Jin Chen
Currently, I am trying to implement full container clone (with all processes running in container). For performance, copy-on-write filesystem will be a good choice. Does lxc container support fs like unionfs-fuse? Any reply will be appreciated. Thanks. Jin -- 陈进 电话:+86-21-51355360 复旦大学并行处理研究所 上海

Re: [lxc-devel] Does lxc support unionfs ?

2012-06-05 Thread Stéphane Graber
On 06/05/2012 04:01 PM, Jin Chen wrote: > Currently, I am trying to implement full container clone (with all > processes running in container). For performance, copy-on-write > filesystem will be a good choice. Does lxc container support fs like > unionfs-fuse? Any reply will be appreciated. > >

Re: [lxc-devel] Does lxc support unionfs ?

2012-06-05 Thread Jin Chen
Thank you, Stéphane. It seems that I need to talk with my lab members about which cow fs is more proper. Wishes, Jin 2012/6/5 Stéphane Graber > On 06/05/2012 04:01 PM, Jin Chen wrote: > > Currently, I am trying to implement full container clone (with all > > processes running in container). For

Re: [lxc-devel] Does lxc support unionfs ?

2012-06-05 Thread Jin Chen
Another question. If I start a ubuntu template lxc container tm1 and do "lxc-start-ephemeral -o tm1", overlayfs mount will not include those ramfs (e.g. /proc, /sys, /run) in original container root directory, right ? Actually, I think this script run outside container tm1 and it even can't see t

Re: [lxc-devel] Does lxc support unionfs ?

2012-06-05 Thread Serge Hallyn
Quoting Jin Chen (ciccin...@gmail.com): > Another question. > > If I start a ubuntu template lxc container tm1 and do "lxc-start-ephemeral > -o tm1", overlayfs mount will not include those ramfs (e.g. /proc, /sys, > /run) in original container root directory, right ? Actually, I think this > scri