We were finally able to get around this using XenServer 6. I believe the fix was to obtain the most recent supplement package also. In the CloudStack 3.0 Install Guide, there is a link that I think will take you to the right location for this. I'm going off of memory at the moment, so I find out otherwise I'll remember to update this post.
Read More...

Hello. I've been trying to figure out why I might get the following error in the management log on the management server:

2012-02-27 21:05:42,009 INFO [cloud.resource.ResourceManagerImpl] (catalina-exec-4:null) Trying to add a new host at nfs://192.168.0.253/Open-Share-5G/new/share/secondary in data center 1
2012-02-27 21:05:42,011 DEBUG [cloud.resource.ServerResourceBase] (catalina-exec-4:null) Retrieving network interface: xenbr0
2012-02-27 21:05:42,011 DEBUG [cloud.resource.ServerResourceBase] (catalina-exec-4:null) Unable to get network interface for xenbr0
2012-02-27 21:05:42,011 DEBUG [cloud.resource.ServerResourceBase] (catalina-exec-4:null) Retrieving network interface: xenbr1
2012-02-27 21:05:42,011 DEBUG [cloud.resource.ServerResourceBase] (catalina-exec-4:null) Unable to get network interface for xenbr1
2012-02-27 21:05:42,011 DEBUG [cloud.resource.ServerResourceBase] (catalina-exec-4:null) Retrieving network interface: null
2012-02-27 21:05:42,011 DEBUG [cloud.resource.ServerResourceBase] (catalina-exec-4:null) Retrieving network interface: null
2012-02-27 21:05:42,011 ERROR [cloud.resource.ServerResourceBase] (catalina-exec-4:null) Nics are not configured!
2012-02-27 21:05:42,013 INFO [cloud.resource.ServerResourceBase] (catalina-exec-4:null) Designating private to be nic eth1
2012-02-27 21:05:42,015 DEBUG [cloud.resource.ServerResourceBase] (catalina-exec-4:null) Parameters for private nic: <PUBLIC IP> - <MAC ADDRESS>-255.255.255.224

The nics are configured, and it ends up designating the wrong nic as the private nic. Instead of eth1, it should be eth0. The log makes it look like the server is searching for xenbr0 and xenbr1 which do not exist on the management server. Should they?

I have no problem mounting secondary storage when using the 192.168 address from anywhere. This includes the management server, the host, and the secondary storage VM. When running the ssvm check from the ssvm, I get:

root@s-3-VM:/usr/local/cloud/systemvm# ./ssvm-check.sh
================================================
First DNS server is 192.168.0.250
PING 192.168.0.250 (192.168.0.250): 56 data bytes
64 bytes from 192.168.0.250: icmp_seq=0 ttl=64 time=2.955 ms
64 bytes from 192.168.0.250: icmp_seq=1 ttl=64 time=0.162 ms
--- 192.168.0.250 ping statistics ---
2 packets transmitted, 2 packets received, 0% packet loss
round-trip min/avg/max/stddev = 0.162/1.558/2.955/1.397 ms
Good: Can ping DNS server
================================================
Good: DNS resolves download.cloud.com
================================================
NFS is currently mounted
================================================
Management server is <PUBLIC IP>. Checking connectivity.
ERROR: Cannot connect to <PUBLIC IP> port 8250
2012/02/28 08:50:42 socat[3500] E connecting to AF=2 <PUBLIC IP>:8250: Connection timed out

(I pulled out the actual public IP and MAC above.)

I think if it would just see the management server as 192.168.0.5, all might be well. The only other thing special about eth0 (192.168.0.5) is that there are two alias interfaces eth0:0 and eth0:1 with 192.168.0.250 and .251 used for DNS since the management server is also running bind.

This is basically the point I've gotten to after several days of troubleshooting why an ISO always says READY: NO after creating one. The status and size never update and I can't create an instance with that ISO since it never shows up in the ISO selections.

Does anyone have any thoughts or suggestions?

Zach
Read More...

Just an update: I re-installaed with XenServer 5.6SP2 instead of 6.0 and then installed the CloudStack Xenserver Support Package according to the CloudStack 3.0 Install Guide. This time, there was no issue.

Perhaps the xenserver-cloud-supp.tgz file is only intended to be used with XenServer 5.6 and not XenServer 6.0.
Read More...

Hello. Does anyone happen to know why this might occur? All is well with the XenServer 6.0 install until adding the CloudStack supplemental pack according to the 3.0 install guide (yeah, I know it's beta). After rebooting, the mgmt interface in the XenConsole comes back with a netmask, gateway, but no ip address. I go to add one and after submiting, it says someting like "Cannot add ip address because the management interface has no ip address." I'm using Intel 82541GI nics with the e1000 driver.

If I enter the command xe-switch-network-backend bridge, then reboot, the interface is there with it's ip address and there is no problem. But, if I switch it back with xe-switch-network-backend openvswitch, the issue is back and I can't add an ip address. Any thoughts on if I just have an un-supported nic, or if this is an issue with openvswitch? I'm actually doing an install with a co-worker on another machine and we are getting the same error. I'd really be surprised if the nics are identical so I'm hoping someone might have some input here from having to deal with this before.

Zach
Read More...