NC not available when auto registrating with second cloud present

[Torsten Spindler]

One cloud is already running. When installing a new cloud in the same subnet, auto registration of nodes to the second cloud does not seem to work.

From the CLC registration.log:
2010-04-30 10:41:13-04:00 | 3882 -> Calling node rockstars-cluster node 10.153.108.113
2010-04-30 10:41:18-04:00 | 3882 -> euca_conf --register-nodes returned 0

Output of
$ euca-describe-availability-zones verbose
AVAILABILITYZONE rockstars-cluster 10.153.108.210
AVAILABILITYZONE |- vm types free / max cpu ram disk
AVAILABILITYZONE |- m1.small 0000 / 0000 1 192 2
AVAILABILITYZONE |- c1.medium 0000 / 0000 1 256 5
AVAILABILITYZONE |- m1.large 0000 / 0000 2 512 10
AVAILABILITYZONE |- m1.xlarge 0000 / 0000 2 1024 20
AVAILABILITYZONE |- c1.xlarge 0000 / 0000 4 2048 20

eucalyptus@rockstars-clc:~$ ssh 10.153.108.113 succeeds.

[Torsten Spindler (tspindler)]

After stopping the other cloud and reinstalling it worked.

[Mathias Gug (mathiaz)]

Could you outline the network topology used by both clouds?

Could you also outline what was failing?

[Torsten Spindler (tspindler)]

The first existing cloud was using the simple topology: one front-end, one node controller. The cloud we installed first was made up of 3 machines: 1 CLC+Walrus, 1 CC+SC, 2 NC.

After installing the second cloud, both clouds had two clusters registered against it. However, the output of euca-describe-availability-zones verbose showed that the nodes were not registered correctly against the CC of the second cloud. So we could never start an instance in the second cloud.

[brian mullan (bmullan)]

Are both "clouds" on the same network segment?
If so is there any chance that the NC of the 2nd cloud confused as to which
CLC it belongs to?
Is DHCP configured in both CLC's?

On Wed, May 19, 2010 at 3:52 AM, Torsten Spindler <email address hidden>wrote:

> The first existing cloud was using the simple topology: one front-end,
> one node controller. The cloud we installed first was made up of 3
> machines: 1 CLC+Walrus, 1 CC+SC, 2 NC.
>
> After installing the second cloud, both clouds had two clusters
> registered against it. However, the output of euca-describe-
> availability-zones verbose showed that the nodes were not registered
> correctly against the CC of the second cloud. So we could never start an
> instance in the second cloud.
>
> --
> NC not available when auto registrating with second cloud present
> https://bugs.launchpad.net/bugs/572388
> You received this bug notification because you are subscribed to
> eucalyptus in ubuntu.
>

[Torsten Spindler (tspindler)]

On Wed, 2010-05-19 at 11:35 +0000, brian mullan wrote:
> Are both "clouds" on the same network segment?

Yes.

> If so is there any chance that the NC of the 2nd cloud confused as to which
> CLC it belongs to?

Quite possible.

> Is DHCP configured in both CLC's?

It's not configured on either one, though the binaries are installed on
the CC.

[Launchpad Janitor (janitor)]

[Expired for eucalyptus (Ubuntu) because there has been no activity for 60 days.]