Thanks for the update, I'm glad to hear it is resolved. Best regards, Kirk
On 08/14/2013 05:57 PM, Mir Islam wrote: > Thanks Kirk for the detailed instructions. Yes I did misread what Marty > mentioned before. Now the router vm is up and I can restart my old vms. > > Best regards > Mir > On Aug 14, 2013, at 4:44 PM, Kirk Kosinski wrote: > >> CloudStack won't automatically create a new virtual router to replace a >> destroyed one. To force CS to recreate one you need to start a new or >> existing VM in the network. This is what Marty meant by "start a user >> VM" in his steps. >> >> The next time you want CS to destroy and recreate a virtual router, you >> can use the network restart option with force enabled instead of >> manually destroying the virtual router. I believe this can be done via >> the UI, otherwise use the API: >> >> http://cloudstack.apache.org/docs/api/apidocs-4.1/root_admin/restartNetwork.html >> >> Lastly, to log on to a system VM you normally can SSH to its link-local >> IP shown in the UI. For system VMs running on XS/KVM hosts, do this >> from the host it is running on: >> >> ssh -p 3922 -i /root/.ssh/id_rsa.cloud root@<ip> >> >> For vSphere, from the CS management server: >> >> ssh -p 3922 -i /var/lib/cloud/management/.ssh/id_rsa root@<ip> >> >> If you see a system VM stuck in single-user mode it is best to destroy >> and recreate it. If you really want to log on to the console you can >> use the default root password of "6m1ll10n". >> >> Best regards, >> Kirk >> >> On 08/14/2013 03:59 PM, Mir Islam wrote: >>> Thanks I am running 4.1 and whenever I went to Virtual router page it >>> showed as "starting" so I could not destroy it like I did with the other >>> system VMs. But it finally got into "stopped" state and I was able to >>> destroy it. However, I do not see it getting recreated. And in the >>> infrastructure I see 0 Virtual Routers. >>> >>> >>> Is the router vm going to recreated by the management server or some other >>> steps needed? It has been about 10 mins but do not see it getting recreated. >>> >>> On Aug 14, 2013, at 3:40 PM, Marty Sweet <msweet....@gmail.com> wrote: >>> >>>> Hi Mir, >>>> >>>> If you are running 4.1.1 go to Infrastructure > Virtual Routers and simply >>>> destroy it, start a user VM and a new one will be created. >>>> I occasionally have this problem when an mishap like yours happens. >>>> >>>> Guest VMs will only start if the management server can communicate with the >>>> virtual router. >>>> >>>> Marty >>>> >>>> On Wednesday, August 14, 2013, Mir Islam wrote: >>>> >>>>> ok here is my problem. I have a setup with two hosts. Which also serve as >>>>> NFS file storage system for primary and secondary. A day ago someone >>>>> accidentally rebooted one of the hosts. Since then I am having all sorts >>>>> of >>>>> issues starting up the system VMs. After destroying the existing ssvm and >>>>> cpvm they were recreated and seems to be working fine. However the virtual >>>>> router is not coming up. I connected to it directly via VNC and it is in >>>>> fsck prompt because of inconsistencies in disk and asking for root >>>>> password. >>>>> >>>>> Is there a default password set for the default router VM ? >>>>> If not, how can I recreate it? >>>>> Also, would lack of router VM cause other VM not to start? I can't restart >>>>> any of the pre existing guest VMs. >>>>> >>>>> Thanks >>>>> Mir >>> >
