[jira] [Commented] (CLOUDSTACK-6065) No HA for shutdown VM
[ https://issues.apache.org/jira/browse/CLOUDSTACK-6065?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13900820#comment-13900820 ] ASF subversion and git services commented on CLOUDSTACK-6065: - Commit 1283919f02e2c4d2652ca18693b6cf2906b2263d in branch refs/heads/4.3-forward from [~kelveny] [ https://git-wip-us.apache.org/repos/asf?p=cloudstack.git;h=1283919 ] CLOUDSTACK-6065: Fix NPE problem caused by the lack of context setup in threads from agent manager thread pool No HA for shutdown VM - Key: CLOUDSTACK-6065 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-6065 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Management Server, UI Affects Versions: 4.3.0 Environment: CentOS 6.5 Mgmt HVs, KVM Reporter: Nux Assignee: Kelven Yang Priority: Critical Labels: ha, kvm I create a service offering with HA, launch a VM with it. I stop the VM from its console or SSH by issuing `poweroff` and the VM is not started back on. This is what I see in the management log: http://fpaste.org/75667/67633139/raw/ The UI keeps showing this VM as Running. However, if I force a reboot of its hypervisor ACS does do the right thing and starts the VM on another HV. -- This message was sent by Atlassian JIRA (v6.1.5#6160)
[jira] [Commented] (CLOUDSTACK-6065) No HA for shutdown VM
[ https://issues.apache.org/jira/browse/CLOUDSTACK-6065?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13900941#comment-13900941 ] ASF subversion and git services commented on CLOUDSTACK-6065: - Commit adf4dd59270917b1607940d7064748bd438cbbcc in branch refs/heads/4.3 from [~kelveny] [ https://git-wip-us.apache.org/repos/asf?p=cloudstack.git;h=adf4dd5 ] CLOUDSTACK-6065: Fix NPE problem caused by the lack of context setup in threads from agent manager thread pool (cherry picked from commit 1283919f02e2c4d2652ca18693b6cf2906b2263d) Signed-off-by: Animesh Chaturvedi anim...@apache.org No HA for shutdown VM - Key: CLOUDSTACK-6065 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-6065 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Management Server, UI Affects Versions: 4.3.0 Environment: CentOS 6.5 Mgmt HVs, KVM Reporter: Nux Assignee: Kelven Yang Priority: Critical Labels: ha, kvm I create a service offering with HA, launch a VM with it. I stop the VM from its console or SSH by issuing `poweroff` and the VM is not started back on. This is what I see in the management log: http://fpaste.org/75667/67633139/raw/ The UI keeps showing this VM as Running. However, if I force a reboot of its hypervisor ACS does do the right thing and starts the VM on another HV. -- This message was sent by Atlassian JIRA (v6.1.5#6160)
[jira] [Commented] (CLOUDSTACK-6065) No HA for shutdown VM
[ https://issues.apache.org/jira/browse/CLOUDSTACK-6065?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13897849#comment-13897849 ] Simon Weller commented on CLOUDSTACK-6065: -- Animesh mentioned in the CLOUDSTACK-5731 notes that the new vmsync code isn't in the 4.3 release. The NPE looks virtually identical though. Can someone clarify this? This is a blocker in my opinion. No HA for shutdown VM - Key: CLOUDSTACK-6065 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-6065 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Management Server, UI Affects Versions: 4.3.0 Environment: CentOS 6.5 Mgmt HVs, KVM Reporter: Nux Assignee: Kelven Yang Priority: Critical Labels: ha, kvm I create a service offering with HA, launch a VM with it. I stop the VM from its console or SSH by issuing `poweroff` and the VM is not started back on. This is what I see in the management log: http://fpaste.org/75667/67633139/raw/ The UI keeps showing this VM as Running. However, if I force a reboot of its hypervisor ACS does do the right thing and starts the VM on another HV. -- This message was sent by Atlassian JIRA (v6.1.5#6160)
[jira] [Commented] (CLOUDSTACK-6065) No HA for shutdown VM
[
https://issues.apache.org/jira/browse/CLOUDSTACK-6065?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13896636#comment-13896636
]
prashant kumar mishra commented on CLOUDSTACK-6065:
---
i tried same scenario , UI shows vm is in stop state , but CS did not tried to
start it on other host
logs:
===
2014-02-10 14:01:59,372 INFO [c.c.v.VirtualMachineManagerImpl]
(AgentManager-Handler-6:null) VM i-2-3-VM is sync-ed to at Stopped state
according to power-off report from hypervisor
2014-02-10 14:01:59,383 DEBUG [c.c.a.m.AgentManagerImpl]
(AgentManager-Handler-6:null) SeqA 4-10: Sending Seq 4-10: { Ans: , MgmtId:
7145449848920, via: 4, Ver: v1, Flags: 100010,
[{com.cloud.agent.api.PingAnswer:{_command:{newGroupStates:{},newStates:{i-2-3-VM:Stopped},_hostVmStateReport:{i-2-3-VM:{state:PowerOff,host:Rack1Pod1Host7}},_gatewayAccessible:true,_vnetAccessible:true,hostType:Routing,hostId:4,wait:0},result:true,wait:0}}]
}
No HA for shutdown VM
-
Key: CLOUDSTACK-6065
URL: https://issues.apache.org/jira/browse/CLOUDSTACK-6065
Project: CloudStack
Issue Type: Bug
Security Level: Public(Anyone can view this level - this is the
default.)
Components: Management Server, UI
Affects Versions: 4.3.0
Environment: CentOS 6.5 Mgmt HVs, KVM
Reporter: Nux
Labels: ha, kvm
I create a service offering with HA, launch a VM with it.
I stop the VM from its console or SSH by issuing `poweroff` and the VM is not
started back on.
This is what I see in the management log:
http://fpaste.org/75667/67633139/raw/
The UI keeps showing this VM as Running.
However, if I force a reboot of its hypervisor ACS does do the right thing
and starts the VM on another HV.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
[jira] [Commented] (CLOUDSTACK-6065) No HA for shutdown VM
[ https://issues.apache.org/jira/browse/CLOUDSTACK-6065?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13896653#comment-13896653 ] Wei Zhou commented on CLOUDSTACK-6065: -- In my point of view, HA means vm will restart if it is stopped unexpectedly, for example, the host which vm is running on is enforced to power off It you stop the vm, the vm will not start. No HA for shutdown VM - Key: CLOUDSTACK-6065 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-6065 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Management Server, UI Affects Versions: 4.3.0 Environment: CentOS 6.5 Mgmt HVs, KVM Reporter: Nux Labels: ha, kvm I create a service offering with HA, launch a VM with it. I stop the VM from its console or SSH by issuing `poweroff` and the VM is not started back on. This is what I see in the management log: http://fpaste.org/75667/67633139/raw/ The UI keeps showing this VM as Running. However, if I force a reboot of its hypervisor ACS does do the right thing and starts the VM on another HV. -- This message was sent by Atlassian JIRA (v6.1.5#6160)
[jira] [Commented] (CLOUDSTACK-6065) No HA for shutdown VM
[ https://issues.apache.org/jira/browse/CLOUDSTACK-6065?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13896693#comment-13896693 ] Nux commented on CLOUDSTACK-6065: - Wei, The docs say When an HA-enabled VM crashes, CloudStack detects the crash and restarts the VM automatically within the same Availability Zone.. If I poweroff the VM from within or `virsh destroy` it on the HV, according to the above Cloudstack should restart it. No HA for shutdown VM - Key: CLOUDSTACK-6065 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-6065 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Management Server, UI Affects Versions: 4.3.0 Environment: CentOS 6.5 Mgmt HVs, KVM Reporter: Nux Labels: ha, kvm I create a service offering with HA, launch a VM with it. I stop the VM from its console or SSH by issuing `poweroff` and the VM is not started back on. This is what I see in the management log: http://fpaste.org/75667/67633139/raw/ The UI keeps showing this VM as Running. However, if I force a reboot of its hypervisor ACS does do the right thing and starts the VM on another HV. -- This message was sent by Atlassian JIRA (v6.1.5#6160)
[jira] [Commented] (CLOUDSTACK-6065) No HA for shutdown VM
[ https://issues.apache.org/jira/browse/CLOUDSTACK-6065?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13896842#comment-13896842 ] edison su commented on CLOUDSTACK-6065: --- It's a bug in vm sync: 2014-02-09 17:38:43,962 DEBUG [c.c.v.VirtualMachineManagerImpl] (AgentManager-Handler-14:null) VM i-2-13-VM: cs state = Running and realState = Stopped 2014-02-09 17:38:43,962 DEBUG [c.c.v.VirtualMachineManagerImpl] (AgentManager-Handler-14:null) VM i-2-13-VM: cs state = Running and realState = Stopped 2014-02-09 17:38:43,962 DEBUG [c.c.h.HighAvailabilityManagerImpl] (AgentManager-Handler-14:null) VM does not require investigation so I'm marking it as Stopped: VM[User|ha1] 2014-02-09 17:38:43,962 WARN [o.a.c.f.j.AsyncJobExecutionContext] (AgentManager-Handler-14:null) Job is executed without a context, setup psudo job for the executing thread 2014-02-09 17:38:43,962 WARN [c.c.a.m.AgentManagerImpl] (AgentManager-Handler-14:null) Caught: java.lang.NullPointerException at org.apache.cloudstack.framework.jobs.AsyncJobExecutionContext.getCurrentExecutionContext(AsyncJobExecutionContext.java:172) at com.cloud.vm.VirtualMachineManagerImpl.advanceStop(VirtualMachineManagerImpl.java:1299) at com.cloud.ha.HighAvailabilityManagerImpl.scheduleRestart(HighAvailabilityManagerImpl.java:346) at com.cloud.vm.VirtualMachineManagerImpl.compareState(VirtualMachineManagerImpl.java:2816) at com.cloud.vm.VirtualMachineManagerImpl.deltaHostSync(VirtualMachineManagerImpl.java:2422) at com.cloud.vm.VirtualMachineManagerImpl.processCommands(VirtualMachineManagerImpl.java:2952) at com.cloud.agent.manager.AgentManagerImpl.handleCommands(AgentManagerImpl.java:292) at com.cloud.agent.manager.AgentManagerImpl$AgentHandler.processRequest(AgentManagerImpl.java:1184) at com.cloud.agent.manager.AgentManagerImpl$AgentHandler.doTask(AgentManagerImpl.java:1270) at com.cloud.agent.manager.ClusteredAgentManagerImpl$ClusteredAgentHandler.doTask(ClusteredAgentManagerImpl.java:677) at com.cloud.utils.nio.Task.run(Task.java:83) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1146) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) No HA for shutdown VM - Key: CLOUDSTACK-6065 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-6065 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Management Server, UI Affects Versions: 4.3.0 Environment: CentOS 6.5 Mgmt HVs, KVM Reporter: Nux Labels: ha, kvm I create a service offering with HA, launch a VM with it. I stop the VM from its console or SSH by issuing `poweroff` and the VM is not started back on. This is what I see in the management log: http://fpaste.org/75667/67633139/raw/ The UI keeps showing this VM as Running. However, if I force a reboot of its hypervisor ACS does do the right thing and starts the VM on another HV. -- This message was sent by Atlassian JIRA (v6.1.5#6160)
[jira] [Commented] (CLOUDSTACK-6065) No HA for shutdown VM
[ https://issues.apache.org/jira/browse/CLOUDSTACK-6065?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13896870#comment-13896870 ] Nux commented on CLOUDSTACK-6065: - Thanks, I've changed status to Critical. No HA for shutdown VM - Key: CLOUDSTACK-6065 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-6065 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Management Server, UI Affects Versions: 4.3.0 Environment: CentOS 6.5 Mgmt HVs, KVM Reporter: Nux Priority: Critical Labels: ha, kvm I create a service offering with HA, launch a VM with it. I stop the VM from its console or SSH by issuing `poweroff` and the VM is not started back on. This is what I see in the management log: http://fpaste.org/75667/67633139/raw/ The UI keeps showing this VM as Running. However, if I force a reboot of its hypervisor ACS does do the right thing and starts the VM on another HV. -- This message was sent by Atlassian JIRA (v6.1.5#6160)
[jira] [Commented] (CLOUDSTACK-6065) No HA for shutdown VM
[ https://issues.apache.org/jira/browse/CLOUDSTACK-6065?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13896879#comment-13896879 ] sadhu suresh commented on CLOUDSTACK-6065: -- Due to following bug its fail to perform HA CLOUDSTACK-5713vmsync:KVM:NPE when agentmanager tries to update the destroyed vm state during vmsync Also A VM is scheduled for HA when the following conditions occur. - A host is found to be Down. - A host is disconnected for more than 30 minutes (time is configurable). - An agent for a host reports that a VM has stopped running. - VMs that have been stuck at the Starting state for more than an hour. - A management server determined that its peer has gone down and is unwinding any VM that the peer was trying to start. - A management server on restart is unwinding any VM that it itself was trying to start. No HA for shutdown VM - Key: CLOUDSTACK-6065 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-6065 Project: CloudStack Issue Type: Bug Security Level: Public(Anyone can view this level - this is the default.) Components: Management Server, UI Affects Versions: 4.3.0 Environment: CentOS 6.5 Mgmt HVs, KVM Reporter: Nux Assignee: Kelven Yang Priority: Critical Labels: ha, kvm I create a service offering with HA, launch a VM with it. I stop the VM from its console or SSH by issuing `poweroff` and the VM is not started back on. This is what I see in the management log: http://fpaste.org/75667/67633139/raw/ The UI keeps showing this VM as Running. However, if I force a reboot of its hypervisor ACS does do the right thing and starts the VM on another HV. -- This message was sent by Atlassian JIRA (v6.1.5#6160)
