=== dendro-afk is now known as dendrobates
[08:09] <ayambit> Hello, anybody alive?
[08:14] <flaccid> no
[08:27] <ayambit> Ok dear undeads, help me then.
[08:27] <ayambit> I need to make about 50 vms on 2 physocal servers.
[08:29] <ayambit> Servers will be powerful enough, 2 8-core xeons each with hyperthreading and from 48 to 96 GB ram, 12 to 16 sas in raid 5 or like this.
[08:29] <ayambit> I want to understand, if it's a good idea to use ubuntu-cloud (or another cloud system like OpenNebula, Nimbus or OpenStack) for this. Or should I look to something more sikmple.
=== daker_ is now known as daker
[11:48] <kim0> ayambit: if you're the only user .. perhaps direct kvm would make more sense
[11:53] <ayambit> kim0: there is a little less than 20 developers to use this stuff.
[11:54] <kim0> hmm
[11:55] <kim0> well you probably want a smaller machine to act as a the cloud head node .. and use those two big servers as node controllers
[11:55] <ayambit> kim0: how do you think I should deal with vm's disk image storing?
[11:56] <ayambit> kim0: I don't think I'll be able to use dedicated SAN.
[11:56] <kim0> ayambit: UEC doesn't need that
[11:56] <kim0> images are copied and local disk is used
[11:56] <ayambit> kim0: is such configuration will be any kind of migration possible?
[11:57] <kim0> not really
[12:00] <ayambit> But can uec be configured to use live (or at least offline) migration? If I'll say take a smaller machine which will act as cloud head node and build a large raid array there, then I'll mount it's dir via nfs to node controllers, will it work then? Of I'll be able to connect servers to each other with say 1Gbit/sec link.
=== dendrobates is now known as dendro-afk
=== hallyn is now known as hallyn_afk
[15:16] <ayambit> ./join #debian
=== hallyn_afk is now known as hallyn
=== JGJones_ is now known as JGJones
=== lborda is now known as lborda_afk
=== dendro-afk is now known as dendrobates
=== daker__ is now known as daker_