cloudstack-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Clayton Weise <cwe...@iswest.net>
Subject RE: Advanced Network Clarification
Date Thu, 22 Nov 2012 02:48:11 GMT
Ivan, this may be of some help to you:

https://cwiki.apache.org/confluence/display/CLOUDSTACK/CloudStack+Example+Configurations

You don't have to specify a dedicated storage network, if you don't it will just use the management
network for storage as well.  You can also setup your storage outside of CloudStack using
something like ShareMountPoint in which case CloudStack will just look for that mount point
and not attempt to configure storage on your hosts.

As far as the cloudbrX interfaces go, there isn't a specific number that is required (e.g.
cloudbr0, cloudbr1, etc).  you could theoretically do it all on one interface as long as CloudStack
is aware of the names.  So when configuring your physical interfaces for your zone, the tag
of the interface should match the bridge name on your KVM hosts.

Finally, the guest VLAN range should be a range of VLANs that CloudStack can assign at will.
 So if you gave it a range of 300-350 then CloudStack would assign a VLAN from within that
range to guest networks as you create them.  Generally speaking, you can expect to have at
least one VLAN for each account and in some cases accounts can create more than one network
but each network will have its own VLAN.  In advanced networking, the VLAN range is used by
CloudStack to pick a VLAN to assign to a guest network.  Each guest network has a separate
VLAN to ensure that they are completely separate and VMs from separate accounts cannot see
each other.

-----Original Message-----
From: Ivan Rodriguez [mailto:ivanoch@gmail.com] 
Sent: Wednesday, November 21, 2012 5:32 PM
To: cloudstack-users@incubator.apache.org
Subject: Advanced Network Clarification

Dear Fellows,

We are trying to finish our cloudstack setup using advanced zone, and I'm
having some doubts about our network design, initially we allocate vlans
for hypervisor management, for primary storage and for secondary storage,
however we are using blades, and on the blade we only have one interface,
we need to tag our vlans on the interface and then build bridges on top of
that this increase the complexity of the configuration in the host, so my
first question is if we are using only 1 interface
in our host does it make sense to have all this vlan + aliases + bridges? I
was thinking to simplify the design just having 2 vlans the public and the
hypervisor + storage.

This bring my second question, when we add storage network, it will ask for
the the range lets say we allocate 10.1.1.0/24 for storage shall we fill
the whole range into the storage or just a part how many ip addresses
cloudstack needs ??? I haven't found a reference to the doco, it seems to
me that we only need 2 ip addresses for that subnet that will be assigned
to the storage system vm, anyone can explain what this ranges are for ???

And Finally on the documentation the example is using 3 vlans what is the
300 vlan for ?? when we use advanced zone we define a range for vlan ids
that we want cloudstack to use

   1. VLAN 100 for management of the hypervisor
   2. VLAN 200 for public network of the instances (cloudbr0)
   3. VLAN 300 for private network of the instances (cloudbr1)

http://incubator.apache.org/cloudstack/docs/en-US/Apache_CloudStack/4.0.0-incubating/html/Installation_Guide/hypervisor-kvm-install-flow.html

So far in our setup we don't use cloudbr1 and its been working fine, do we
need cloudbr1 on advanced network configuration ? is it recommended ??

Thanks in advance
Ivan

Mime
View raw message