Re: Upgrade from 4.1.1 to 4.2 failed. Could not restart system VMs. Roll back to 4.1.1 backup initially failed: could not start management service. Finally got it to work. I hope this helps someone el
On Oct 16, 2013, at 10:48 AM, Milamber milam...@apache.org wrote: Hello, curl: (7) couldn't connect to host Your integration.api.port isn't define in your global settings. See below. Current instructions to upgrade from 4.1.1 to 4.2 are not complete on official docs (missing 2 important steps before upgrade). https://cloudstack.apache.org/docs/en-US/Apache_CloudStack/4.2.0/html/Release_Notes/upgrade-instructions.html#upgrade-from-4.0-to-4.1 (4.2 in reality) Please note too, in 4.2.0, currently the HA with CLVM (iscsi) don't works for Guest VM (works for VR/SystemVM). This have been fixed for 4.2.1 See : https://issues.apache.org/jira/browse/CLOUDSTACK-4627 https://issues.apache.org/jira/browse/CLOUDSTACK-4777 You can cherry-pick the commit to fix on 4.2.0 tag. Here my process for Centos 6.4, KVM, CS 4.1.1 to CS 4.2.0: BEFORE UPGRADE 1/ With the Web UI, add a new template with the name systemvm-kvm-4.2 (exactly) from this link: http://download.cloud.com/templates/4.2/systemvmtemplate-2013-06-12-master-kvm.qcow2.bz2 Don't start upgrade until the status change to Download complete 2/ With the Web UI, go to Global settings, check if the integration.api.port is sets to 8096. If not, set up. Restart Mngt service. (you can revert this change after upgrade for security reasons) For point 1/ ant 2/, See in the docs, the start of process 3.0.x to 4.2 https://cloudstack.apache.org/docs/en-US/Apache_CloudStack/4.2.0/html/Release_Notes/upgrade-instructions.html#upgrade-from-3.0.x-to-4.0 START OF UPGRADE 3/ Stop CS services (all mgnt, all nodes) /etc/init.d/cloudstack-management stop /etc/init.d/cloudstack-usage stop /etc/init.d/cloudstack-agent stop mysqldump -u root -p cloud cloudstack-cloud-backup.sql mysqldump -u root -p cloud_usage cloudstack-cloud_usage-backup.sql mysqldump -u root -p cloudbridge cloudstack-cloudbridge-backup.sql 4/ Modify /etc/yum.repos.d/cloudstack.repo to change to 4.2 repo 5/ Upgrade (all mgnt, all nodes) yum clean all yum update (full update CS + CentOS) OR for only CS update yum upgrade cloudstack-management yum upgrade cloudstack-agent See step 10.a-e in docs 6/ Start agent (all nodes) /etc/init.d/cloudstack-agent start 7/ Start Mngt (first mgnt) /etc/init.d/cloudstack-management start Check log tail -f /var/log/cloudstack/management | grep -v DEBUG 8/ Restart (recreate) the SSVM, Console proxy and all VR cloudstack-sysvmadm -d ipOfDatabase -u cloud -pyourDBpassword -a (long process, you can follow on console and see in Web UI the status of system vms/VR) 9/ Start Usage service /etc/init.d/cloudstack-usage start (and start all mgnt service if needs) Milamber Milamber, thanks for writing this down, I am copying Kelcey who hit this issue and Travis who has been working on docs. Le 15/10/2013 22:10, Adam a ecrit : I upgraded my 4 host private dev cloud running on CentOS 6.4 x86_64 from version 4.1.1 to 4.2 per upgrade instructions. I'm just using 4 HP Z 600 workstations (box 1 has two nics, the cloudstack-management server, the mysql db, primary and secondary storage and a cloudstack-agent, and boxes 2 - 4 each just have primary storage and a cloudstack-agent). It's nothing fancy, but it's been working perfectly now for months. All seemed to go very smoothly except for the very last step: {code} nohup cloudstack-sysvmadm -d cs-east-dev1 -u root -p support -a sysvm.log 21 {code} It could not restart the system vms for some reason. (I do not have the original sysvm.log as I've now been playing with this failed upgrade for two days). However, here is the sysvm.log from the very last attempt: {code} nohup: ignoring input Stopping and starting 1 secondary storage vm(s)... curl: (7) couldn't connect to host ERROR: Failed to stop secondary storage vm with id 14 Done stopping and starting secondary storage vm(s) Stopping and starting 0 console proxy vm(s)... No running console proxy vms found Stopping and starting 1 running routing vm(s)... curl: (7) couldn't connect to host 2 Done restarting router(s). {code} As I mentioned above, I've been playing around with this for 2 days now and actually got the 4.2 management server to finally start, but none of the System VMs worked. I even restarted all of the CentOS hosts (which was a huge hassle), but that didn't seem to help at all. I eventually found this bug: https://issues.apache.org/jira/browse/CLOUDSTACK-4826 which seemed similar to my issues. I was also experiencing a strange issue where all 10 of my private Management IP Addresses were used for some reason. Every time I restarted the cloudstack-management service, 2 more IPs were taken up, but none ever got released. Also since the System VMs would not start, my secondary storage wouldn't start up either. About an hour ago I gave up on 4.2 and I decided to roll back to 4.1.1 on all 4 workstations. First I
Re: [DOC]:baremetal kickstart document was old
jijunix is added On Thu, Oct 17, 2013 at 8:24 AM, Jijun jiju...@gmail.com wrote: recently i have test the CS + baremetal kickstart, found the baremetal kickstart document was old , https://cwiki.apache.org/confluence/display/CLOUDSTACK/Baremetal+KickStart i want to have the right to update it , can someone give me the right to do that? My account name is my email address. -- Thanks, Jijun
Re: cloudstack4.2.0 network configuration - ssvm has no IP
thank you. figured out. it's because of using wrong template. On Fri, Oct 18, 2013 at 10:45 AM, Sanjeev Neelarapu sanjeev.neelar...@citrix.com wrote: Hi, While creating a zone you must specify guest and management traffic with ip ranges from your primary network. If you want to use dedicated storage network then add the storage traffic to physical network (during zone creation) and specify ip range from your second network. This should work find. Thanks, Sanjeev -Original Message- From: raj kumar [mailto:rajkumar600...@gmail.com] Sent: Thursday, October 17, 2013 3:39 PM To: users@cloudstack.apache.org Subject: cloudstack4.2.0 network configuration - ssvm has no IP Hi, I'd like to use basic network configuration. No vlans in the network. I've two networks. one is primary used for all cloudstack servers(guest, system vms, management server, hypervisor, etc). another one is connected to storage. only hypervisors(kvm) have access to this network. my questions. 1. what should be my network configuration in hypervisors. I've used eth0, and cloudstack automatically created cloud0, cloudbr0, etc. 2. Even though ssvm is up, it didn't get any IP. Hence I could not connect the console from web or can't add any iso. please let me know what should be my configuration and how to fix the ssvm issue.
Re: XenServer 6.0.2 won't stay connected after upgrading to CS 4.2
Ryan, A null pointer exception is most definitely a bug. Can you file a jira ticket? thanks, Daan On Thu, Oct 10, 2013 at 5:01 PM, Ryan James r...@colocateusa.net wrote: We just upgraded to CloudStack 4.2 (from 4.0.2) and now our Xen Cluster will not stay connected and the host are in alert states. Here is a snip it out of the management-server.log 2013-10-10 09:22:13,995 DEBUG [cloud.capacity.CapacityManagerImpl] (AgentTaskPool-1:null) Found 6 VMs on host 4 2013-10-10 09:22:14,003 ERROR [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Monitor ComputeCapacityListener says there is an error in the connect process for 4 due to null java.lang.NullPointerException at com.cloud.capacity.CapacityManagerImpl.updateCapacityForHost(CapacityManagerImpl.java:543) at com.cloud.utils.component.ComponentInstantiationPostProcessor$InterceptorDispatcher.intercept(ComponentInstantiationPostProcessor.java:125) at com.cloud.capacity.ComputeCapacityListener.processConnect(ComputeCapacityListener.java:78) at com.cloud.agent.manager.AgentManagerImpl.notifyMonitorsOfConnection(AgentManagerImpl.java:587) at com.cloud.agent.manager.AgentManagerImpl.handleDirectConnectAgent(AgentManagerImpl.java:1479) at com.cloud.resource.ResourceManagerImpl.createHostAndAgent(ResourceManagerImpl.java:1762) at com.cloud.resource.ResourceManagerImpl.createHostAndAgent(ResourceManagerImpl.java:1924) at com.cloud.agent.manager.AgentManagerImpl$SimulateStartTask.run(AgentManagerImpl.java:1130) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1146) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) at java.lang.Thread.run(Thread.java:679) 2013-10-10 09:22:14,004 INFO [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Host 4 is disconnecting with event AgentDisconnected 2013-10-10 09:22:14,008 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) The next status of agent 4is Alert, current status is Connecting 2013-10-10 09:22:14,008 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Deregistering link for 4 with state Alert 2013-10-10 09:22:14,008 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Remove Agent : 4 2013-10-10 09:22:14,009 DEBUG [agent.manager.DirectAgentAttache] (AgentTaskPool-1:null) Processing disconnect 4 2013-10-10 09:22:14,009 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Sending Disconnect to listener: com.cloud.hypervisor.xen.discoverer.XcpServerDiscoverer_EnhancerByCloudStack_434ade97 2013-10-10 09:22:14,009 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Sending Disconnect to listener: com.cloud.deploy.DeploymentPlanningManagerImpl_EnhancerByCloudStack_a0f690d 2013-10-10 09:22:14,009 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Sending Disconnect to listener: com.cloud.network.NetworkManagerImpl_EnhancerByCloudStack_1ba07aa0 2013-10-10 09:22:14,009 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Sending Disconnect to listener: com.cloud.storage.secondary.SecondaryStorageListener 2013-10-10 09:22:14,009 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Sending Disconnect to listener: com.cloud.hypervisor.vmware.manager.VmwareManagerImpl_EnhancerByCloudStack_b315799a 2013-10-10 09:22:14,009 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Sending Disconnect to listener: com.cloud.network.security.SecurityGroupListener 2013-10-10 09:22:14,009 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Sending Disconnect to listener: com.cloud.storage.listener.StoragePoolMonitor 2013-10-10 09:22:14,009 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Sending Disconnect to listener: com.cloud.vm.ClusteredVirtualMachineManagerImpl_EnhancerByCloudStack_48612ba4 2013-10-10 09:22:14,009 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Sending Disconnect to listener: com.cloud.storage.LocalStoragePoolListener 2013-10-10 09:22:14,009 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Sending Disconnect to listener: com.cloud.network.SshKeysDistriMonitor 2013-10-10 09:22:14,009 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Sending Disconnect to listener: com.cloud.network.router.VirtualNetworkApplianceManagerImpl_EnhancerByCloudStack_e1d29845 2013-10-10 09:22:14,009 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Sending Disconnect to listener: com.cloud.network.SshKeysDistriMonitor 2013-10-10 09:22:14,009 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Sending Disconnect to listener: com.cloud.network.router.VpcVirtualNetworkApplianceManagerImpl_EnhancerByCloudStack_5cb66068 2013-10-10 09:22:14,009 DEBUG [agent.manager.AgentManagerImpl] (AgentTaskPool-1:null) Sending Disconnect to listener:
Re: Cloudstack 4.2 manaagement server not starting
I'd check the process with ps. if it's really not there remove the pid and lock files and try again. On Wed, Oct 9, 2013 at 6:26 PM, Support Staff serversupportst...@gmail.com wrote: Hi, I am trying to install cloudstack 4.2 on centos 6.4. I have three separate servers for management server and database and storage. I completed all configurations in all servers with no issues and the command cloudstack-setup-management finished with no errors. But when i tried to access using serverip:8080/client i am getting blank page. The status of cloudstack says. service cloudstack-management status cloudstack-management dead but pid file exists The pid file locates at /var/run/cloudstack-management.pid and lock file at /var/lock/subsys/cloudstack-management. Starting cloudstack-management will take care of them or you can manually clean up. I have tried changing the permission of catilina.out to 777. But it didn't fixed the issue. Please advice us on the issue. Also i am able to download the template of xenserver.
RE: Cloudstack 4.2 manaagement server not starting
How much RAM does your CloudStack Management machine have? Regards Geoff Higginbottom D: +44 20 3603 0542 | S: +44 20 3603 0540 | M: +447968161581 geoff.higginbot...@shapeblue.com -Original Message- From: Daan Hoogland [mailto:daan.hoogl...@gmail.com] Sent: 18 October 2013 11:59 To: users@cloudstack.apache.org Subject: Re: Cloudstack 4.2 manaagement server not starting I'd check the process with ps. if it's really not there remove the pid and lock files and try again. On Wed, Oct 9, 2013 at 6:26 PM, Support Staff serversupportst...@gmail.com wrote: Hi, I am trying to install cloudstack 4.2 on centos 6.4. I have three separate servers for management server and database and storage. I completed all configurations in all servers with no issues and the command cloudstack-setup-management finished with no errors. But when i tried to access using serverip:8080/client i am getting blank page. The status of cloudstack says. service cloudstack-management status cloudstack-management dead but pid file exists The pid file locates at /var/run/cloudstack-management.pid and lock file at /var/lock/subsys/cloudstack-management. Starting cloudstack-management will take care of them or you can manually clean up. I have tried changing the permission of catilina.out to 777. But it didn't fixed the issue. Please advice us on the issue. Also i am able to download the template of xenserver. This email and any attachments to it may be confidential and are intended solely for the use of the individual to whom it is addressed. Any views or opinions expressed are solely those of the author and do not necessarily represent those of Shape Blue Ltd or related companies. If you are not the intended recipient of this email, you must neither take any action based upon its contents, nor copy or show it to anyone. Please contact the sender if you believe you have received this email in error. Shape Blue Ltd is a company incorporated in England Wales. ShapeBlue Services India LLP is a company incorporated in India and is operated under license from Shape Blue Ltd. Shape Blue Brasil Consultoria Ltda is a company incorporated in Brasil and is operated under license from Shape Blue Ltd. ShapeBlue is a registered trademark.
SSVM Public IP, NAT and L2/L3 connectivity
Hi All, Still in the planning stages of deploying my first CS install and I have a question about the SSVM and the apparent requirement for a public routable IP address. From what I can gather, the only interaction that the SSVM has with the internet is for downloading files from user or admin supplied URLs. Does this mean that nothing actually makes inbound connections directly to it from the internet and that it only ever makes outbound connections? If so, why does it need a public IP? Would it be possible to simply have a route to the internet via the management network or to give it an IP from private network pool that has access to the internet via some other NAT device? Secondly, I’ve seen some excellent slides from Geoff Higginbottom but I’m still not quite sure whether the SSVM actually has four vNICs or whether it simply needs access to four networks via a lesser number of vNICs. Can anyone clarify how many vNICs each SSVM has and what the routing table looks like on the VM itself, especially where the management server and secondary storage server are on subnets accessible via a L3 hop and not directly attached to either the hypervisor mgmt or SSVM vNICs (and these subnets aren’t accessible by a default route on the SSVM)? My main concern is the public IP requirement as I’m finding it very difficult to get enough public IPs from my DC, especially where there doesn’t appear to be a reason for it. I can’t simply get a nice big block of IPv4 like some other CS users may be used to! I’m considering joining RIPE but this is not especially cheap and doesn’t seem to guarantee that I’d get an assignment of IPv4 addresses anyway. Confused, Adrian --- Alsi Consulting Ltd www.alsiconsulting.co.uk T: 0845 8676586 M: 07961 127738
Re: Domain Admins
Rafael, Please see http://cloudstack.apache.org/docs/api/apidocs-4.2/TOC_Domain_Admin.html You will find that domain admins can only en/disable users not create them. You need the cloudoperator / rootadmin for that On Mon, Oct 7, 2013 at 5:33 PM, Rafael Weingartner rafaelweingart...@gmail.com wrote: Hi all, Can domain-admins create new user? I thought they would be able to create user at list on the domain they are admins or subdomains. -- Rafael Weingartner
RE: Cloudstack 4.2 manaagement server not starting
The message: service cloudstack-management status cloudstack-management dead but pid file exists appears only if the service is stopped. Try service cloudstack-management start and check ps. -Original Message- From: Daan Hoogland [mailto:daan.hoogl...@gmail.com] Sent: Friday, October 18, 2013 4:29 PM To: users@cloudstack.apache.org Subject: Re: Cloudstack 4.2 manaagement server not starting I'd check the process with ps. if it's really not there remove the pid and lock files and try again. On Wed, Oct 9, 2013 at 6:26 PM, Support Staff serversupportst...@gmail.com wrote: Hi, I am trying to install cloudstack 4.2 on centos 6.4. I have three separate servers for management server and database and storage. I completed all configurations in all servers with no issues and the command cloudstack-setup-management finished with no errors. But when i tried to access using serverip:8080/client i am getting blank page. The status of cloudstack says. service cloudstack-management status cloudstack-management dead but pid file exists The pid file locates at /var/run/cloudstack-management.pid and lock file at /var/lock/subsys/cloudstack-management. Starting cloudstack-management will take care of them or you can manually clean up. I have tried changing the permission of catilina.out to 777. But it didn't fixed the issue. Please advice us on the issue. Also i am able to download the template of xenserver.
Re: Domain Admins
I have one more question about Domain Admin and rights, what is the reason domain admin can not crate the sub-domains? The domain level is the central level to differentiation rights in cloudstack_ the domain admin have to get more right to mange own domain. May be there is a reason for it? Rafael, Please see http://cloudstack.apache.org/docs/api/apidocs-4.2/TOC_Domain_Admin.html You will find that domain admins can only en/disable users not create them. You need the cloudoperator / rootadmin for that On Mon, Oct 7, 2013 at 5:33 PM, Rafael Weingartner rafaelweingart...@gmail.com wrote: Hi all, Can domain-admins create new user? I thought they would be able to create user at list on the domain they are admins or subdomains. -- Rafael Weingartner
RE: SSVM Public IP, NAT and L2/L3 connectivity
Hi Adrian, Please find responses inline -Original Message- From: Adrian Lewis [mailto:adr...@alsiconsulting.co.uk] Sent: Friday, October 18, 2013 4:41 PM To: users@cloudstack.apache.org Subject: SSVM Public IP, NAT and L2/L3 connectivity Hi All, Still in the planning stages of deploying my first CS install and I have a question about the SSVM and the apparent requirement for a public routable IP address. From what I can gather, the only interaction that the SSVM has with the internet is for downloading files from user or admin supplied URLs. Does this mean that nothing actually makes inbound connections directly to it from the internet and that it only ever makes outbound connections? If so, why does it need a public IP? [Sanjeev]: I could think of one scenario which is copying templates from one zone to another zone where zones are at different geographical locations(i.e. two zones are connected using internet). Would it be possible to simply have a route to the internet via the management network or to give it an IP from private network pool that has access to the internet via some other NAT device? [Sanjeev] As per the cloudstack terminology public ip does not mean that they are real public IPs. They can be private IPs from which internet can be accessed using other NAT device. Secondly, I've seen some excellent slides from Geoff Higginbottom but I'm still not quite sure whether the SSVM actually has four vNICs or whether it simply needs access to four networks via a lesser number of vNICs. Can anyone clarify how many vNICs each SSVM has and what the routing table looks like on the VM itself, especially where the management server and secondary storage server are on subnets accessible via a L3 hop and not directly attached to either the hypervisor mgmt or SSVM vNICs (and these subnets aren't accessible by a default route on the SSVM)? [Sanjeev] It is necessary that there should be four vNICs on SSVM , each belonging to only one network. Management and Storage servers need not be on the same subnet. They should be accessible via a L3 hop. My main concern is the public IP requirement as I'm finding it very difficult to get enough public IPs from my DC, especially where there doesn't appear to be a reason for it. I can't simply get a nice big block of IPv4 like some other CS users may be used to! I'm considering joining RIPE but this is not especially cheap and doesn't seem to guarantee that I'd get an assignment of IPv4 addresses anyway. Confused, Adrian --- Alsi Consulting Ltd www.alsiconsulting.co.uk T: 0845 8676586 M: 07961 127738
Re: Domain Admins
Hi Daan, Thanks for the reply. I also found this: http://markmail.org/message/ucxsapyannat2z6b https://issues.apache.org/jira/browse/CLOUDSTACK-4830 It seems that on CS 4.2.1 it will be possible. I will also work on a workaround here to make it possible since we need to enable domain admins to create users at their domains. At least for me it does not make much sense that domain admins cannot create users at domain level. 2013/10/18 Daan Hoogland daan.hoogl...@gmail.com Rafael, Please see http://cloudstack.apache.org/docs/api/apidocs-4.2/TOC_Domain_Admin.html You will find that domain admins can only en/disable users not create them. You need the cloudoperator / rootadmin for that On Mon, Oct 7, 2013 at 5:33 PM, Rafael Weingartner rafaelweingart...@gmail.com wrote: Hi all, Can domain-admins create new user? I thought they would be able to create user at list on the domain they are admins or subdomains. -- Rafael Weingartner -- Rafael Weingartner
Re: Domain Admins
Ivan, My colleagues and I had the same question. I think the domain system is just not crystallized completely yet. Whether it is a design or a technical limitation, it needs more work obviously. If you have specific requirements for your domain admins, please create tickets for them in jira. If there are users wanting to work on requirements for that please add a proposal for a hackathon at ccc europe at https://cwiki.apache.org/confluence/display/CLOUDSTACK/CCC+Europe or start a [DISCUSS] thread here and in dev@cloudst regards, Daan On Fri, Oct 18, 2013 at 1:52 PM, Ivan Gladenko webmas...@colobridge.net wrote: I have one more question about Domain Admin and rights, what is the reason domain admin can not crate the sub-domains? The domain level is the central level to differentiation rights in cloudstack_ the domain admin have to get more right to mange own domain. May be there is a reason for it? Rafael, Please see http://cloudstack.apache.org/docs/api/apidocs-4.2/TOC_Domain_Admin.html You will find that domain admins can only en/disable users not create them. You need the cloudoperator / rootadmin for that On Mon, Oct 7, 2013 at 5:33 PM, Rafael Weingartner rafaelweingart...@gmail.com wrote: Hi all, Can domain-admins create new user? I thought they would be able to create user at list on the domain they are admins or subdomains. -- Rafael Weingartner
Limits on tags per resource in CS
Hi all, Is there any limit on the number of tags I can use for a resource in CloudStack? I use some tags for the templates and I am concerned whether I can use as many tags as I want. Thank you! Krasimir
Re: Limits on tags per resource in CS
Wonderful! Each tag value cannot be longer than 255 characters. Is there any mechanism for simulating lists with tags then? Let me give an example. Assume that I want to associate my template with some other resources (other templates let's say). I want to enumerate their IDs as tags. Apparently, I cannot enumerate the IDs as I have a limitation of 255 characters. Additionally, I want to have those tags grouped somehow:) Krasi On Fri, Oct 18, 2013 at 3:47 PM, Stanislav Petrov devilfighter1...@gmail.com wrote: Hi Krasimir, The official documentation ( https://cloudstack.apache.org/docs/en-US/Apache_CloudStack/4.2.0/html/Admin_Guide/tagging-resources.html) says There is no limit on the number of tags you can define. Regards, Stanislav Petrov On Fri, Oct 18, 2013 at 3:15 PM, Krasimir Baylov kbai...@gmail.comwrote: Hi all, Is there any limit on the number of tags I can use for a resource in CloudStack? I use some tags for the templates and I am concerned whether I can use as many tags as I want. Thank you! Krasimir -- Best regards, Stanislav Petrov
Re: Limits on tags per resource in CS
On Oct 18, 2013, at 8:55 AM, Krasimir Baylov kbai...@gmail.com wrote: Wonderful! Each tag value cannot be longer than 255 characters. Is there any mechanism for simulating lists with tags then? Let me give an example. Assume that I want to associate my template with some other resources (other templates let's say). I want to enumerate their IDs as tags. Apparently, I cannot enumerate the IDs as I have a limitation of 255 characters. Additionally, I want to have those tags grouped somehow:) Krasi You might want to have a look at this piece of python: https://github.com/NOPping/gCloud/blob/master/gcloud/controllers/operations.py#L80-L105 we used tags to store ssh keys and had to break them up un 255 segments. On Fri, Oct 18, 2013 at 3:47 PM, Stanislav Petrov devilfighter1...@gmail.com wrote: Hi Krasimir, The official documentation ( https://cloudstack.apache.org/docs/en-US/Apache_CloudStack/4.2.0/html/Admin_Guide/tagging-resources.html) says There is no limit on the number of tags you can define. Regards, Stanislav Petrov On Fri, Oct 18, 2013 at 3:15 PM, Krasimir Baylov kbai...@gmail.comwrote: Hi all, Is there any limit on the number of tags I can use for a resource in CloudStack? I use some tags for the templates and I am concerned whether I can use as many tags as I want. Thank you! Krasimir -- Best regards, Stanislav Petrov
Re: Setting user data removes SSH public key data on a running VM
Hi again, Did you stop start the vm after updating the Userdata (after step 4).Even after stop and start if same behavior exists then this is an issue. - Yes, I did restart the VM and the behavior is the same (available user data, removed SSH public key). The public key being removed from the VR is the correct behavior. The VR has a cache that stores the ssh key when you make the API call and once it has been requested is removed from the cache. - Nope, this is NOT true. VR stores user data and SSH public key on two distinct files. I looked at the source code here saveUserDataToRouter() in file server/src/com/cloud/network/router/VirtualNetworkApplianceManagerImpl.java. When saving user data to VR, it specifies SSH public key as null and that removes the public key from the VR. I think the correct behavior should be to check if there is existing SSH public key for this VM. If this is true to specify its value as method argument. Definitely, this seems to be an issue. On Fri, Oct 11, 2013 at 3:13 PM, Travis Graham tgra...@tgraham.us wrote: The public key being removed from the VR is the correct behavior. The VR has a cache that stores the ssh key when you make the API call and once it has been requested is removed from the cache. Travis On Oct 11, 2013, at 4:31 AM, Stanislav Petrov devilfighter1...@gmail.com wrote: Hi all, My Environment is: CS 4.1.1, KVM hypervisor Here is the use case steps: 1. Create SSH keypair keypairname through the REST API. 2. Deploy VM with the created keypair from step 1. 3. On the VM, the url http://dhcpserverip/latest/public-keys returns the public key. 4. Update the running VM with userdata. 5. On the VM, the url http://dhcpserverip/latest/public-keys returns nothing and the user data url http://dhcpserverip/latest/user-data returns the user-data which was set in step 4. Result: User data is set but public key is gone (cleaned). When I update user data on a stopped VM, both SSH public key and user data are available. Is this behavior expected or is this some kind of an issue ? -- Best regards, Stanislav Petrov -- Best regards, Stanislav Petrov
Re: Creating advanced network
Bjoern, Sorry that commit, only fixes part of the problem. Still there are two more issues (source NAT and SG + source NAT combination is not permitted and public traffic type is not allowed in security group based shared network). I opened a feature enhancement CLOUDSTACK-4891 bug for this issue. You may want to try basic zone model of CloudStack which provides security group based L3 isolation with EIP(1:1 NAT) ELB services with NetScaler. Thanks, Murali From: Bjoern Teipel bjoern.tei...@internetbrands.commailto:bjoern.tei...@internetbrands.com Reply-To: users@cloudstack.apache.orgmailto:users@cloudstack.apache.org users@cloudstack.apache.orgmailto:users@cloudstack.apache.org Date: Thursday, 17 October 2013 10:29 AM To: users@cloudstack.apache.orgmailto:users@cloudstack.apache.org users@cloudstack.apache.orgmailto:users@cloudstack.apache.org Subject: Re: Creating advanced network Hi Murali, I saw your git commits. I want to compile now your changes into our source code. Do i need just the one for 4.2 or also the master commits: Commit 4d07493a5e6e13462b80ba09c3535fa4af0ebdc7 in branch refs/heads/4.2 from Murali Reddyhttps://issues.apache.org/jira/secure/ViewProfile.jspa?name=murali.reddy https://issues.apache.org/jira/browse/CLOUDSTACK-4717# [cid:part3.09050002.00050203@internetbrands.com]ASF subversion and git serviceshttps://issues.apache.org/jira/secure/ViewProfile.jspa?name=jira-bot added a comment - Today 06:18 Commit df3b09944968718111d9b6b29d4c7f5a5cfaf630 in branch refs/heads/master from Murali Reddyhttps://issues.apache.org/jira/secure/ViewProfile.jspa?name=murali.reddy https://issues.apache.org/jira/browse/CLOUDSTACK-4717# [cid:part3.09050002.00050203@internetbrands.com]ASF subversion and git serviceshttps://issues.apache.org/jira/secure/ViewProfile.jspa?name=jira-bot added a comment - Today 14:45 Commit df3b09944968718111d9b6b29d4c7f5a5cfaf630 in branch refs/heads/ui-restyle from Murali Reddyhttps://issues.apache.org/jira/secure/ViewProfile.jspa?name=murali.reddy Thanks, Bjoern On 10/16/2013 2:35 AM, Murali Reddy wrote: On 16/10/13 12:23 PM, Bjoern Teipel bjoern.tei...@internetbrands.commailto:bjoern.tei...@internetbrands.com wrote: Murali, That would be great if you're right. But I'm now in a dead lock: Adding new network offering including LB: 2013-10-15 23:34:50,920 WARN [network.element.VirtualRouterElement] (catalina-exec-19:null) Virtual router can't enable services [Dns Dhcp UserData Lb ] without source NAT service 2013-10-15 23:34:50,924 ERROR [cloud.api.ApiServer] (catalina-exec-19:null) unhandled exception executing api command: createNetworkOffering com.cloud.exception.UnsupportedServiceException: Provider VirtualRouter doesn't support services combination: [Dns, Dhcp, UserData, Lb] That forces me to add source nat, but once I want add a guest network in the zone I get the opposite error. I can't mix SG + sourceNat 013-10-15 23:46:30,896 INFO [cloud.api.ApiServer] (catalina-exec-22:null) Service SourceNat is not allowed in security group enabled zone First issue is know issue (CLOUDSTACK-4717) is getting addressed in 4.2.1. Not sure why source NAT should not be allowed in SG network. Sorry, this is indeed a dead lock situation. It does not look like you can use LB with in shared network with SG in advanced zone. So no internal lb ? Thanks, Bjoern On 10/15/2013 11:28 PM, Murali Reddy wrote: On 16/10/13 7:17 AM, Bjoern Teipel bjoern.tei...@internetbrands.commailto:bjoern.tei...@internetbrands.com wrote: Wow, all user@cloudstack mails got catched in my spam filter, so sorry for the late response. After tinkering the whole day I gave up using a tagged VLAN for the storage traffic, seems not to work. It ignores the VID and doesn't create the VLAN on the hypervisor. I added the vlan to the hypervisor now and bound cloudbr1 to it and using it untagged in cloudstack. Finally all is up. :-) Now I was looking how to use a load balancer like the internal cloudstack one or even the F5 and it seems it's not supported. No cloudstack support for internal LB (the VR one) or F5 ? Really !!! According to the advanced network and security groups specification ( https://cwiki.apache.org/confluence/display/CLOUDSTACK/Isolation+based+o n+ Security+Groups+in+Advance+zone) AddF5LoadBalancerCmd api commands will just fail in SG enabled zone. That's just a joke. 4.1 did not support PF/NAT/LB services in shared network. From 4.2, all network services are supported in shared network with or without SG so you should be able to use F5/VR/Netscaler for LB. I'm really close to end the cloudstack adventure and move on with open stack. Having a shared network with SG and loadbalancer is not really a uncommon solution
Re: Upgrade from 4.1.1 to 4.2 failed. Could not restart system VMs. Roll back to 4.1.1 backup initially failed: could not start management service. Finally got it to work. I hope this helps someone el
Hi, This post might help you out. http://cloud.kelceydamage.com/cloudfire/blog/2013/10/08/conquering-the-cloudstack-4-2-dragon-kvm/ If you need further help, please let me know. -Kelcey Sent from my HTC - Reply message - From: sebgoa run...@gmail.com To: users@cloudstack.apache.org, Kelcey Jamison Damage kel...@backbonetechnology.com, Travis Graham tgra...@tgraham.us Subject: Upgrade from 4.1.1 to 4.2 failed. Could not restart system VMs. Roll back to 4.1.1 backup initially failed: could not start management service. Finally got it to work. I hope this helps someone else avoid my same pain... Date: Fri, Oct 18, 2013 12:12 AM On Oct 16, 2013, at 10:48 AM, Milamber milam...@apache.org wrote: Hello, curl: (7) couldn't connect to host Your integration.api.port isn't define in your global settings. See below. Current instructions to upgrade from 4.1.1 to 4.2 are not complete on official docs (missing 2 important steps before upgrade). https://cloudstack.apache.org/docs/en-US/Apache_CloudStack/4.2.0/html/Release_Notes/upgrade-instructions.html#upgrade-from-4.0-to-4.1 (4.2 in reality) Please note too, in 4.2.0, currently the HA with CLVM (iscsi) don't works for Guest VM (works for VR/SystemVM). This have been fixed for 4.2.1 See : https://issues.apache.org/jira/browse/CLOUDSTACK-4627 https://issues.apache.org/jira/browse/CLOUDSTACK-4777 You can cherry-pick the commit to fix on 4.2.0 tag. Here my process for Centos 6.4, KVM, CS 4.1.1 to CS 4.2.0: BEFORE UPGRADE 1/ With the Web UI, add a new template with the name systemvm-kvm-4.2 (exactly) from this link: http://download.cloud.com/templates/4.2/systemvmtemplate-2013-06-12-master-kvm.qcow2.bz2 Don't start upgrade until the status change to Download complete 2/ With the Web UI, go to Global settings, check if the integration.api.port is sets to 8096. If not, set up. Restart Mngt service. (you can revert this change after upgrade for security reasons) For point 1/ ant 2/, See in the docs, the start of process 3.0.x to 4.2 https://cloudstack.apache.org/docs/en-US/Apache_CloudStack/4.2.0/html/Release_Notes/upgrade-instructions.html#upgrade-from-3.0.x-to-4.0 START OF UPGRADE 3/ Stop CS services (all mgnt, all nodes) /etc/init.d/cloudstack-management stop /etc/init.d/cloudstack-usage stop /etc/init.d/cloudstack-agent stop mysqldump -u root -p cloud cloudstack-cloud-backup.sql mysqldump -u root -p cloud_usage cloudstack-cloud_usage-backup.sql mysqldump -u root -p cloudbridge cloudstack-cloudbridge-backup.sql 4/ Modify /etc/yum.repos.d/cloudstack.repo to change to 4.2 repo 5/ Upgrade (all mgnt, all nodes) yum clean all yum update (full update CS + CentOS) OR for only CS update yum upgrade cloudstack-management yum upgrade cloudstack-agent See step 10.a-e in docs 6/ Start agent (all nodes) /etc/init.d/cloudstack-agent start 7/ Start Mngt (first mgnt) /etc/init.d/cloudstack-management start Check log tail -f /var/log/cloudstack/management | grep -v DEBUG 8/ Restart (recreate) the SSVM, Console proxy and all VR cloudstack-sysvmadm -d ipOfDatabase -u cloud -pyourDBpassword -a (long process, you can follow on console and see in Web UI the status of system vms/VR) 9/ Start Usage service /etc/init.d/cloudstack-usage start (and start all mgnt service if needs) Milamber Milamber, thanks for writing this down, I am copying Kelcey who hit this issue and Travis who has been working on docs. Le 15/10/2013 22:10, Adam a ecrit : I upgraded my 4 host private dev cloud running on CentOS 6.4 x86_64 from version 4.1.1 to 4.2 per upgrade instructions. I'm just using 4 HP Z 600 workstations (box 1 has two nics, the cloudstack-management server, the mysql db, primary and secondary storage and a cloudstack-agent, and boxes 2 - 4 each just have primary storage and a cloudstack-agent). It's nothing fancy, but it's been working perfectly now for months. All seemed to go very smoothly except for the very last step: {code} nohup cloudstack-sysvmadm -d cs-east-dev1 -u root -p support -a sysvm.log 21 {code} It could not restart the system vms for some reason. (I do not have the original sysvm.log as I've now been playing with this failed upgrade for two days). However, here is the sysvm.log from the very last attempt: {code} nohup: ignoring input Stopping and starting 1 secondary storage vm(s)... curl: (7) couldn't connect to host ERROR: Failed to stop secondary storage vm with id 14 Done stopping and starting secondary storage vm(s) Stopping and starting 0 console proxy vm(s)... No running console proxy vms found Stopping and starting 1 running routing vm(s)... curl: (7) couldn't connect to host 2 Done restarting router(s). {code} As I mentioned above, I've been playing around with this for 2 days now and actually got the 4.2 management server to finally start, but none of the System VMs worked. I even
Re: Creating advanced network
My problem is that I want to integrate f5 load balancer also. So I'm stuck with advanced mode. Also I don't like that VMs have per default public IPs in basic mode.. Bjoern On Oct 18, 2013, at 6:22 AM, Murali Reddy murali.re...@citrix.commailto:murali.re...@citrix.com wrote: Bjoern, Sorry that commit, only fixes part of the problem. Still there are two more issues (source NAT and SG + source NAT combination is not permitted and public traffic type is not allowed in security group based shared network). I opened a feature enhancement CLOUDSTACK-4891 bug for this issue. You may want to try basic zone model of CloudStack which provides security group based L3 isolation with EIP(1:1 NAT) ELB services with NetScaler. Thanks, Murali From: Bjoern Teipel bjoern.tei...@internetbrands.commailto:bjoern.tei...@internetbrands.com Reply-To: users@cloudstack.apache.orgmailto:users@cloudstack.apache.org users@cloudstack.apache.orgmailto:users@cloudstack.apache.org Date: Thursday, 17 October 2013 10:29 AM To: users@cloudstack.apache.orgmailto:users@cloudstack.apache.org users@cloudstack.apache.orgmailto:users@cloudstack.apache.org Subject: Re: Creating advanced network Hi Murali, I saw your git commits. I want to compile now your changes into our source code. Do i need just the one for 4.2 or also the master commits: Commit 4d07493a5e6e13462b80ba09c3535fa4af0ebdc7 in branch refs/heads/4.2 from Murali Reddyhttps://issues.apache.org/jira/secure/ViewProfile.jspa?name=murali.reddy https://issues.apache.org/jira/browse/CLOUDSTACK-4717# [cid:part3.09050002.00050203@internetbrands.com]ASF subversion and git serviceshttps://issues.apache.org/jira/secure/ViewProfile.jspa?name=jira-bot added a comment - Today 06:18 Commit df3b09944968718111d9b6b29d4c7f5a5cfaf630 in branch refs/heads/master from Murali Reddyhttps://issues.apache.org/jira/secure/ViewProfile.jspa?name=murali.reddy https://issues.apache.org/jira/browse/CLOUDSTACK-4717# [cid:part3.09050002.00050203@internetbrands.com]ASF subversion and git serviceshttps://issues.apache.org/jira/secure/ViewProfile.jspa?name=jira-bot added a comment - Today 14:45 Commit df3b09944968718111d9b6b29d4c7f5a5cfaf630 in branch refs/heads/ui-restyle from Murali Reddyhttps://issues.apache.org/jira/secure/ViewProfile.jspa?name=murali.reddy Thanks, Bjoern On 10/16/2013 2:35 AM, Murali Reddy wrote: On 16/10/13 12:23 PM, Bjoern Teipel bjoern.tei...@internetbrands.commailto:bjoern.tei...@internetbrands.com wrote: Murali, That would be great if you're right. But I'm now in a dead lock: Adding new network offering including LB: 2013-10-15 23:34:50,920 WARN [network.element.VirtualRouterElement] (catalina-exec-19:null) Virtual router can't enable services [Dns Dhcp UserData Lb ] without source NAT service 2013-10-15 23:34:50,924 ERROR [cloud.api.ApiServer] (catalina-exec-19:null) unhandled exception executing api command: createNetworkOffering com.cloud.exception.UnsupportedServiceException: Provider VirtualRouter doesn't support services combination: [Dns, Dhcp, UserData, Lb] That forces me to add source nat, but once I want add a guest network in the zone I get the opposite error. I can't mix SG + sourceNat 013-10-15 23:46:30,896 INFO [cloud.api.ApiServer] (catalina-exec-22:null) Service SourceNat is not allowed in security group enabled zone First issue is know issue (CLOUDSTACK-4717) is getting addressed in 4.2.1. Not sure why source NAT should not be allowed in SG network. Sorry, this is indeed a dead lock situation. It does not look like you can use LB with in shared network with SG in advanced zone. So no internal lb ? Thanks, Bjoern On 10/15/2013 11:28 PM, Murali Reddy wrote: On 16/10/13 7:17 AM, Bjoern Teipel bjoern.tei...@internetbrands.commailto:bjoern.tei...@internetbrands.com wrote: Wow, all user@cloudstack mails got catched in my spam filter, so sorry for the late response. After tinkering the whole day I gave up using a tagged VLAN for the storage traffic, seems not to work. It ignores the VID and doesn't create the VLAN on the hypervisor. I added the vlan to the hypervisor now and bound cloudbr1 to it and using it untagged in cloudstack. Finally all is up. :-) Now I was looking how to use a load balancer like the internal cloudstack one or even the F5 and it seems it's not supported. No cloudstack support for internal LB (the VR one) or F5 ? Really !!! According to the advanced network and security groups specification ( https://cwiki.apache.org/confluence/display/CLOUDSTACK/Isolation+based+o n+ Security+Groups+in+Advance+zone) AddF5LoadBalancerCmd api commands will just fail in SG enabled zone. That's just a joke. 4.1 did not support PF/NAT/LB services in shared network. From 4.2, all network services are supported in shared network with or without SG so you should be able to use F5/VR/Netscaler for LB. I'm really close to end the cloudstack adventure and move on with open stack. Having a shared network with SG and
Re: Limits on tags per resource in CS
Hi Sebastien, Thank you for your reply. I was thinking of using nearly the same approach. Regards, Krasimir On Fri, Oct 18, 2013 at 4:03 PM, Sebastien Goasguen run...@gmail.comwrote: On Oct 18, 2013, at 8:55 AM, Krasimir Baylov kbai...@gmail.com wrote: Wonderful! Each tag value cannot be longer than 255 characters. Is there any mechanism for simulating lists with tags then? Let me give an example. Assume that I want to associate my template with some other resources (other templates let's say). I want to enumerate their IDs as tags. Apparently, I cannot enumerate the IDs as I have a limitation of 255 characters. Additionally, I want to have those tags grouped somehow:) Krasi You might want to have a look at this piece of python: https://github.com/NOPping/gCloud/blob/master/gcloud/controllers/operations.py#L80-L105 we used tags to store ssh keys and had to break them up un 255 segments. On Fri, Oct 18, 2013 at 3:47 PM, Stanislav Petrov devilfighter1...@gmail.com wrote: Hi Krasimir, The official documentation ( https://cloudstack.apache.org/docs/en-US/Apache_CloudStack/4.2.0/html/Admin_Guide/tagging-resources.html ) says There is no limit on the number of tags you can define. Regards, Stanislav Petrov On Fri, Oct 18, 2013 at 3:15 PM, Krasimir Baylov kbai...@gmail.com wrote: Hi all, Is there any limit on the number of tags I can use for a resource in CloudStack? I use some tags for the templates and I am concerned whether I can use as many tags as I want. Thank you! Krasimir -- Best regards, Stanislav Petrov
Building 4.2 -nonoss
Hi, I've followed the instructione for building 4.2 -nonoss from source here: https://cwiki.apache.org/confluence/display/CLOUDSTACK/How+to+build+CloudStack# Couple things...the install-non-oss.sh is a little different for 4.2 than it was for 4.1.1 and 4.1 in that it does not process the cloud-netscaler.jar any more? Is this intentional? Second thing is this from the management-server.log - I seem to be missing a load balancer bean? 2013-10-18 07:33:13,382 ERROR [web.context.ContextLoader] (main:null) Context initialization failed org.springframework.beans.factory.BeanCreationException: Error creating bean with name 'apiDBUtils': Injection of autowire d dependencies failed; nested exception is org.springframework.beans.factory.BeanCreationException: Could not autowire field: private com.cloud.region.ha.GlobalLoadBalancingRulesService com.cloud.api.ApiDBUtils.gslbService; nested exception is org.springframework.beans.factory.NoSuchBeanDefinitionException: No matching bean of type [com.cloud.region.ha.GlobalLoadB alancingRulesService] found for dependency: expected at least 1 bean which qualifies as autowire candidate for this depend ency. Dependency annotations: {@javax.inject.Inject()} at org.springframework.beans.factory.annotation.AutowiredAnnotationBeanPostProcessor.postProcessPropertyValues(Aut owiredAnnotationBeanPostProcessor.java:287) at org.springframework.beans.factory.support.AbstractAutowireCapableBeanFactory.populateBean(AbstractAutowireCapab leBeanFactory.java:1106) at org.springframework.beans.factory.support.AbstractAutowireCapableBeanFactory.doCreateBean(AbstractAutowireCapab leBeanFactory.java:517) at org.springframework.beans.factory.support.AbstractAutowireCapableBeanFactory.createBean(AbstractAutowireCapable BeanFactory.java:456) Suggestions? Thanks for your help, Fred
Cloudstack and JuJu
Hello guys, Was wondering if there has been any work on integrating JuJu with cloudstack? If so, could you please point me in the right direction? Many thank Andrei
Re: Cloudstack and JuJu
There aren't currently any charms written for CloudStack that I can find, but they shouldn't be too hard to write. You'd need to break them down into a few different charms, depending on how you or others might want to deploy them. You'll want to browse some of the other charms to borrow from existing work such as things needed for MySQL and NFS, I would think those for sure have been written and could be reused. Other things like KVM, libvirt, qemu and the like are probably also lurking around in the charms store to be borrowed from. If there's a large enough community backing for Juju I could probably work on this at some point, but it probably wouldn't been as soon as you'd need it if you're looking to roll your own install now. You may want to start a [DISCUSS] thread to solicit feedback from the community on interest and implementation options. Getting an idea of what people would like before digging in would help anyone who's willing to work on making it possible. Personally, I think for those using MASS it might be something they would like to see. Especially for CEPH support since CentOS support is lagging. Travis On Oct 18, 2013, at 11:50 AM, Andrei Mikhailovsky and...@arhont.com wrote: Hello guys, Was wondering if there has been any work on integrating JuJu with cloudstack? If so, could you please point me in the right direction? Many thank Andrei
Re: Secondary Storage IP Subnet: Is it Optional or Mandatory?
Hi, Thanks Sanjeev and Geoff for your replies. I have a few follow-up questions. A. The document I referenced states if a physical storage interface was not specified when creating the zone this interface will not exist in reference to eth3 (the SSVM storage network interface). Is that no longer the case? I don't see the logic of having a second virtual NIC on the same subnet. B. I know that having multiple NICs on the same system connected to the same subnet can sometimes cause network issues. If this is to be the resulting configuration on the SSVM, is there anything I should watch out for? What has been done to head off any issues that could result from this configuration? C. Aside from performance/redundancy concerns, can you offer any general comments regarding the wisdom of opting not to have a secondary storage IP subnet? Beyond being technically possible and permitted, is this a common and accepted approach? For my better understanding, and because I am weighing the options, I'd still appreciate answers to my questions about the Secondary Storage network, re-posted below: 4. I haven't found any documentation that helps with planning the size of the Secondary Storage network -- besides the NFS server and the SSVM, what other devices must have an IP address on this network? 5. How are IP addresses for devices on the Secondary Storage network assigned? In the case of the SSVM it must be dynamic, but if I were to manually put a CloudStack host onto this network (for convenience), what assures that the same IP address won't also get handed out to an SSVM? 6. Is it permitted to place the Secondary Storage traffic onto its own dedicated vSwitch on ESXi? According to the Installation Guide under section 8.3.5.1.1 Separating Traffic: *CloudStack allows you to use vCenter to configure three separate networks per ESXi host. *...* The allowed networks for configuration are public, guest, and private (for management and usually storage traffic).* This would suggest that storage traffic must share the same vSwitch as management traffic. Thanks again, Rob From: Geoff Higginbottom geoff.higginbot...@shapeblue.com Date: Thu, Oct 17, 2013 at 2:56 AM Subject: Re: Secondary Storage IP Subnet: Is it Optional or Mandatory? To: users@cloudstack.apache.org users@cloudstack.apache.org Hi Rob, The 'storage' network is indeed optional so your plan to not use it is perfectly valid. Simply put the NFS server on the management network or ensure there is a route to it via the management network. The SSVM will still get a virtual interface dedicated to storage, but it will be assigned an additional IP address from the management ip range. There is no need to configure any ip range for storage, and if using the add zone wizard it will not prompt you for one. Regards Geoff Higginbottom CTO / Cloud Architect D: +44 20 3603 0542tel:+442036030542 | S: +44 20 3603 0540tel: +442036030540 | M: +447968161581tel:+447968161581 geoff.higginbot...@shapeblue.commailto:geoff.higginbot...@shapeblue.com | www.shapeblue.com ShapeBlue Ltd, 53 Chandos Place, Covent Garden, London, WC2N 4HS On 17 Oct 2013, at 01:08, Rob C. iose...@gmail.commailto: iose...@gmail.com wrote: Hi, According to some (possibly old) documentation I read, I have the impression that it is *optional* to define a secondary storage IP subnet. Note, my questions are all in the context of: - CloudStack 4.2 - ESXi hosts / vSphere 5.1 - Primary Storage is VMFS via fiber channel - Advanced Networking Here is the document to which I am referring, along with the relevant quotes: CloudStack System VMs: https://cwiki.apache.org/confluence/download/attachments/30149076/Cloudstack+System+VMs.pdf *Network Terminology - Storage: As it relates to CloudStack, this is an optional network dedicated to secondary storage. If not specified, the management network will be assumed for this role.* *SSVM - eth3 (storage): Note if a physical storage interface was not specified when creating the zone this interface will not exist. Storage traffic will assume the management interface* My Questions: 1. Is the above-referenced document accurate for CloudStack 4.2? 2. Do I interpret correctly that I don't need to define an IP subnet for Secondary Storage? 3. I have noticed that the CloudStack 4.2 Installation Guide, section 2.8.3 Advanced Zone Network Traffic Types, for the type Storage says: *You must configure the IP range to use for the storage network.* http://cloudstack.apache.org/docs/en-US/Apache_CloudStack/4.2.0/html-single/Installation_Guide/index.html#advanced-zone-network-traffic-types How am I to interpret the quoted statement? Is it saying that I must configure a range regardless, or only if I *opt* to have a Secondary Storage network? If there is no Secondary Storage IP subnet, would I enter a range selected from the Management IP subnet? I am aware of the redundancy and performance
Re: Cloudstack and JuJu
On Oct 18, 2013, at 12:08 PM, Travis Graham tgra...@tgraham.us wrote: There aren't currently any charms written for CloudStack that I can find, but they shouldn't be too hard to write. You'd need to break them down into a few different charms, depending on how you or others might want to deploy them. You'll want to browse some of the other charms to borrow from existing work such as things needed for MySQL and NFS, I would think those for sure have been written and could be reused. Other things like KVM, libvirt, qemu and the like are probably also lurking around in the charms store to be borrowed from. If there's a large enough community backing for Juju I could probably work on this at some point, but it probably wouldn't been as soon as you'd need it if you're looking to roll your own install now. You may want to start a [DISCUSS] thread to solicit feedback from the community on interest and implementation options. Getting an idea of what people would like before digging in would help anyone who's willing to work on making it possible. Personally, I think for those using MASS it might be something they would like to see. Especially for CEPH support since CentOS support is lagging. I think it would be a great addition to the overall ecosystem/config mgt. Go for it ! -sebastien Travis On Oct 18, 2013, at 11:50 AM, Andrei Mikhailovsky and...@arhont.com wrote: Hello guys, Was wondering if there has been any work on integrating JuJu with cloudstack? If so, could you please point me in the right direction? Many thank Andrei
Re: Missing jar vmware-lib-jaxrpc.jar?
Hi Daan, I was reluctant to do that because of this: git commit: CLOUDSTACK-346: Replace vmware-lib-jaxrpc with axis-jaxrpc The fact that I was seeing something trying to load this class meant that something was very wrong. Which makes me suspect the dependencies section on this page: https://cwiki.apache.org/confluence/display/CLOUDSTACK/How+to+build+CloudStack# Is this documentation still good? Thanks, Fred On Thu, Oct 17, 2013 at 1:38 PM, Daan Hoogland daan.hoogl...@gmail.comwrote: Fred, Attachements are stripped from maila to the list. It sounds like this is your problem yes. Why don't you try adding the jar and restarting the ms? regards, Daan On Wed, Oct 16, 2013 at 3:04 PM, Fred Messinger fredmessin...@gmail.com wrote: Hi All, I'm trying to upgrade from 4.1.1 to 4.2 using vmware. And I think I am SO close! My management server log file is attached. It describes some deserialization problems and it also shows that a VMware resource class file cannot be loaded. From googling the exception, I found this jira ticket: https://issues.apache.org/jira/browse/CLOUDSTACK-211 Is the reason I can't start my ssvm because of this? Thanks, Fred
Re: Secondary Storage IP Subnet: Is it Optional or Mandatory?
Rob, Answers in line prefixed GH Regards Geoff Higginbottom CTO / Cloud Architect D: +44 20 3603 0542tel:+442036030542 | S: +44 20 3603 0540tel:+442036030540 | M: +447968161581tel:+447968161581 geoff.higginbot...@shapeblue.commailto:geoff.higginbot...@shapeblue.com | www.shapeblue.com ShapeBlue Ltd, 53 Chandos Place, Covent Garden, London, WC2N 4HS On 18 Oct 2013, at 18:18, Rob C. iose...@gmail.commailto:iose...@gmail.com wrote: Hi, Thanks Sanjeev and Geoff for your replies. I have a few follow-up questions. A. The document I referenced states if a physical storage interface was not specified when creating the zone this interface will not exist in reference to eth3 (the SSVM storage network interface). Is that no longer the case? I don't see the logic of having a second virtual NIC on the same subnet. GH - there are always 4 vNICs, link local, management, storage and public. B. I know that having multiple NICs on the same system connected to the same subnet can sometimes cause network issues. If this is to be the resulting configuration on the SSVM, is there anything I should watch out for? What has been done to head off any issues that could result from this configuration? GH - if the storage network is not configured CloudStack creates static routes mapping the IPs of the sec storage devices to the 'storage' vNIC. Default gateway is the public interface. C. Aside from performance/redundancy concerns, can you offer any general comments regarding the wisdom of opting not to have a secondary storage IP subnet? Beyond being technically possible and permitted, is this a common and accepted approach? GH - I've been involved in lots of deployments and I'd say about half have used the storage network and half have not. If you have hosts with lots of 1GB NICs you can dedicate some to Sec Storage to help with the transferring of Snapshot data. If you only have 2x10GB NICs then the common approach is to simply put all traffic over a LACP bond. For my better understanding, and because I am weighing the options, I'd still appreciate answers to my questions about the Secondary Storage network, re-posted below: 4. I haven't found any documentation that helps with planning the size of the Secondary Storage network -- besides the NFS server and the SSVM, what other devices must have an IP address on this network? GH - The IPs you allocate to the storage network for the POD are used by the SSVMs running in that POD. Every Host also needs to have a connection to all Secondary Storage devices within the Zone, as do the Management Servers. 5. How are IP addresses for devices on the Secondary Storage network assigned? In the case of the SSVM it must be dynamic, but if I were to manually put a CloudStack host onto this network (for convenience), what assures that the same IP address won't also get handed out to an SSVM? GH - The IPs are for SSVM only. The IPs you allocate to Hosts must be outside of the range you allocate to the POD. FYI Hosts HAVE to have a connection to Sec storage, either directly or via their default gateway. 6. Is it permitted to place the Secondary Storage traffic onto its own dedicated vSwitch on ESXi? According to the Installation Guide under section 8.3.5.1.1 Separating Traffic: *CloudStack allows you to use vCenter to configure three separate networks per ESXi host. *...* The allowed networks for configuration are public, guest, and private (for management and usually storage traffic).* This would suggest that storage traffic must share the same vSwitch as management traffic. GH - when allocating the different traffic types to different NICs you map the traffic to the vSwitch so you have to configure a dedicated vSwitch for each network. As a side note, if you have a zone with multiple hypervisors, say VMWare and XenServer, the networks need to be common across the Zone, so all your hosts should have the same number of NICs. For VMware you map the traffic to the vSwitch, and for XenServer you map it to the name of the network or bond, and for KVM the bridge etc. Thanks again, Rob From: Geoff Higginbottom geoff.higginbot...@shapeblue.commailto:geoff.higginbot...@shapeblue.com Date: Thu, Oct 17, 2013 at 2:56 AM Subject: Re: Secondary Storage IP Subnet: Is it Optional or Mandatory? To: users@cloudstack.apache.orgmailto:users@cloudstack.apache.org users@cloudstack.apache.orgmailto:users@cloudstack.apache.org Hi Rob, The 'storage' network is indeed optional so your plan to not use it is perfectly valid. Simply put the NFS server on the management network or ensure there is a route to it via the management network. The SSVM will still get a virtual interface dedicated to storage, but it will be assigned an additional IP address from the management ip range. There is no need to configure any ip range for storage, and if using the add zone wizard it will not prompt you for one. Regards Geoff Higginbottom CTO /
Migrate volume from local to shared storage
Hi, I have 2 hosts with all local primary storage, and we have obtained new shared primary storage now. What would be a good/possible way to migrate existing volumes from local to shared primary storage? I have already done migration of some templates from one NFS secondary storage to another NFS secondary storage (on another host , by changing records inside template_host_ref, template_vm tables...so scp images, and change host to another host, etc.. But with primary storage, not sure if this is possible, since disk offering for these local volumes is local type, not shared type... is it possible at all ? Thanks, -- Andrija Panić
4.2.0: KVM snapshorts are not working
Does anyone tried creating disk snapshots on KVM and it fails, although the qcow files got created. Filed CLOUDSTACK-4878 just to be sure 2013-10-15 21:45:59,415 DEBUG [agent.transport.Request] (AgentManager-Handler-8:null) Seq 1-621806172: Processing: { Ans: , MgmtId: 110493122496, via: 1, Ver: v1, Flags: 110, [{org.apache.cloudstack.storage.command.CopyCmdAnswer:{newData:{org.apache.cloudstack.storage.to.SnapshotObjectTO:{path:snapshots/4/5/57ef69cd-1bc8-4e47-94c1-84834befaaad,id:0}},result:true,wait:0}}] } 2013-10-15 21:45:59,415 DEBUG [agent.manager.AgentAttache] (AgentManager-Handler-8:null) Seq 1-621806172: No more commands found 2013-10-15 21:45:59,415 DEBUG [agent.transport.Request] (Job-Executor-2:job-40 = [ 81b303ad-acb2-483f-9571-d09064d1d086 ]) Seq 1-621806172: Received: { Ans: , MgmtId: 110493122496, via: 1, Ver: v1, Flags: 110, { CopyCmdAnswer } } 2013-10-15 21:45:59,417 DEBUG [storage.snapshot.SnapshotServiceImpl] (Job-Executor-2:job-40 = [ 81b303ad-acb2-483f-9571-d09064d1d086 ]) Failed to update snapshot state com.cloud.utils.exception.CloudRuntimeException: DB Exception on: com.mysql.jdbc.JDBC4PreparedStatement@79fd21d8: SELECT snapshot_store_ref.id, snapshot_store_ref.store_id, snapshot_store_ref.store_role, snapshot_store_ref.snapshot_id, snapshot_store_ref.created, snapshot_store_ref.last_updated, snapshot_store_ref.size, snapshot_store_ref.physical_size, snapshot_store_ref.parent_snapshot_id, snapshot_store_ref.job_id, snapshot_store_ref.install_path, snapshot_store_ref.update_count, snapshot_store_ref.updated, snapshot_store_ref.state, snapshot_store_ref.ref_cnt, snapshot_store_ref.volume_id FROM snapshot_store_ref WHERE snapshot_store_ref.snapshot_id = 1 AND snapshot_store_ref.store_id = 1 AND snapshot_store_ref.store_role = 'Image' ORDER BY RAND() LIMIT 1 at com.cloud.utils.db.GenericDaoBase.searchIncludingRemoved(GenericDaoBase.java:414) at com.cloud.utils.component.ComponentInstantiationPostProcessor$InterceptorDispatcher.intercept(ComponentInstantiationPostProcessor.java:125) at com.cloud.utils.db.GenericDaoBase.searchIncludingRemoved(GenericDaoBase.java:349) at com.cloud.utils.component.ComponentInstantiationPostProcessor$InterceptorDispatcher.intercept(ComponentInstantiationPostProcessor.java:125) at com.cloud.utils.db.GenericDaoBase.findOneIncludingRemovedBy(GenericDaoBase.java:859) at com.cloud.utils.component.ComponentInstantiationPostProcessor$InterceptorDispatcher.intercept(ComponentInstantiationPostProcessor.java:125) at com.cloud.utils.db.GenericDaoBase.findOneBy(GenericDaoBase.java:870) at com.cloud.utils.component.ComponentInstantiationPostProcessor$InterceptorDispatcher.intercept(ComponentInstantiationPostProcessor.java:125) at org.apache.cloudstack.storage.image.db.SnapshotDataStoreDaoImpl.findByStoreSnapshot(SnapshotDataStoreDaoImpl.java:178) at com.cloud.utils.component.ComponentInstantiationPostProcessor$InterceptorDispatcher.intercept(ComponentInstantiationPostProcessor.java:125) at org.apache.cloudstack.storage.snapshot.SnapshotObject.processEvent(SnapshotObject.java:253) at org.apache.cloudstack.storage.snapshot.SnapshotServiceImpl.copySnapshotAsyncCallback(SnapshotServiceImpl.java:315) at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) at java.lang.reflect.Method.invoke(Method.java:616) at org.apache.cloudstack.framework.async.AsyncCallbackDispatcher.dispatch(AsyncCallbackDispatcher.java:142) at org.apache.cloudstack.framework.async.InplaceAsyncCallbackDriver.performCompletionCallback(InplaceAsyncCallbackDriver.java:26) at org.apache.cloudstack.framework.async.AsyncCallbackDispatcher.complete(AsyncCallbackDispatcher.java:120) at org.apache.cloudstack.storage.motion.AncientDataMotionStrategy.copyAsync(AncientDataMotionStrategy.java:423) at org.apache.cloudstack.storage.motion.DataMotionServiceImpl.copyAsync(DataMotionServiceImpl.java:55) at org.apache.cloudstack.storage.snapshot.SnapshotServiceImpl.backupSnapshot(SnapshotServiceImpl.java:265) at org.apache.cloudstack.storage.snapshot.XenserverSnapshotStrategy.backupSnapshot(XenserverSnapshotStrategy.java:138) at org.apache.cloudstack.storage.snapshot.XenserverSnapshotStrategy.takeSnapshot(XenserverSnapshotStrategy.java:264) at com.cloud.storage.snapshot.SnapshotManagerImpl.takeSnapshot(SnapshotManagerImpl.java:1013) at com.cloud.utils.component.ComponentInstantiationPostProcessor$InterceptorDispatcher.intercept(ComponentInstantiationPostProcessor.java:125) at
agent not running in ssvm
Hi guys, I'm running out of time on cloudstack implementation. java agent is not running in ssvm, but running console system vm. In ssvm, I'm getting repeated error: /var/log/cloud.log shows, ERROR [cloud.agent.AgentShell] (main:null) Unable to start agent: Resource class not found: com.cloud.storage.resource.PremiumSecondaryStorageResource due to: java.lang.ClassNotFoundException: com.cloud.storage.resource.PremiumSecondaryStorageResource. Env: cloudstack 4.2.0 mgmt serve: centos6.4 kvm hosts: ubuntu13.04 sorry for duplicating the issue here. the original thread is: http://mail-archives.apache.org/mod_mbox/cloudstack-users/201310.mbox/%3CCAGUH%3DuZMdDN17JuXEx-itp8PjU7kqZtPs8p%2B%3DBbpLmskJ-BFdw%40mail.gmail.com%3E