Uh oh!
There was an error while loading. Please reload this page.
Fix stop VM issue on basic zones - #3501
Conversation
nvazquez
commented
Jul 18, 2019
@blueorangutan package |
blueorangutan
commented
Jul 18, 2019
@nvazquez a Jenkins job has been kicked to build packages. I'll keep you posted as I make progress. |
Uh oh!
There was an error while loading. Please reload this page.
blueorangutan
commented
Jul 18, 2019
Packaging result: ✔centos6 ✔centos7 ✔debian. JID-144 |
nvazquez
commented
Jul 18, 2019
@blueorangutan package |
blueorangutan
commented
Jul 18, 2019
@nvazquez a Jenkins job has been kicked to build packages. I'll keep you posted as I make progress. |
blueorangutan
commented
Jul 18, 2019
Packaging result: ✔centos6 ✔centos7 ✔debian. JID-145 |
nvazquez
commented
Jul 18, 2019
@blueorangutan test |
blueorangutan
commented
Jul 18, 2019
@nvazquez a Trillian-Jenkins test job (centos7 mgmt + kvm-centos7) has been kicked to run smoke tests |
blueorangutan
commented
Jul 18, 2019
Trillian test result (tid-174)
|
server24
commented
Jul 18, 2019
@rhtyd @anuragaw@nvazquez I have recompiled with this test and the DHCP-entries are correct in the Basic Networking (create, start, stop, destroy, expunge VM tested). |
borisstoyanov
left a comment
There was a problem hiding this comment.
IDK if this is expected but if you try to restart the network with clean up in a basic zone you get the following exception:
2019-07-19 09:04:23,245 ERROR [c.c.a.ApiAsyncJobDispatcher] (API-Job-Executor-20:ctx-4eb94b11 job-40) (logid:bfca90d8) Unexpected exception while executing org.apache.cloudstack.api.command.user.network.RestartNetworkCmd
com.cloud.exception.ResourceUnavailableException: Resource [DataCenter:1] is unreachable: Can't find all necessary running routers!
at com.cloud.network.element.VirtualRouterElement.implement(VirtualRouterElement.java:244)
at org.apache.cloudstack.engine.orchestration.NetworkOrchestrator.implementNetworkElements(NetworkOrchestrator.java:1287)
at org.apache.cloudstack.engine.orchestration.NetworkOrchestrator.rollingRestartRouters(NetworkOrchestrator.java:3063)
at org.apache.cloudstack.engine.orchestration.NetworkOrchestrator.restartNetwork(NetworkOrchestrator.java:2940)
at com.cloud.network.NetworkServiceImpl.restartNetwork(NetworkServiceImpl.java:1902)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
this is an issue that I had already reported the 3rd of January, 2019 to the developers list, relevant excerpt: ... "Can't find all necessary running routers!" I have looked up the source code for it and came to the following relevant snippet in server/src/main/java/com/cloud/network/element/VirtualRouterElement.java starting at line 234: From the log files I see that a rollingRestart is indeed triggered, but why does it expect two routers then? Because there might be two routers running simultaneously during the restart? If that was the case, then why in the following lines, relevant snippet (starting at line 237): is throwing an Exception when current_routers.size() is smaller than 2 and not smaller than 1? I understand that the logic is designed towards more complex networks, but will this logic not always fails for an already implemented BASIC network on KVM, with no redudant routers enabled and one VR already active? Back then nobody replied or gave feedback, glad that finally somebody is looking at this. Generally I have the feeling that no testing is ever done for Basic Networking... is everybody just using Advanced Networking with CS? |
yadvr
commented
Jul 22, 2019
yadvr
commented
Jul 22, 2019
@blueorangutan package |
blueorangutan
commented
Jul 22, 2019
@rhtyd a Jenkins job has been kicked to build packages. I'll keep you posted as I make progress. |
blueorangutan
commented
Jul 22, 2019
Packaging result: ✔centos6 ✔centos7 ✔debian. JID-164 |
yadvr
commented
Jul 22, 2019
@blueorangutan test |
blueorangutan
commented
Jul 22, 2019
@rhtyd a Trillian-Jenkins test job (centos7 mgmt + kvm-centos7) has been kicked to run smoke tests |
DaanHoogland
commented
Jul 22, 2019
Aren't we going to drop support for basic zones? How does this fit in that scenario? cc @GabrielBrascher@wido @rhtyd |
yadvr
commented
Jul 22, 2019
@DaanHoogland it has been discussed but a concrete plan has not been proposed. I think that's a very long term thing, till then the feature should be supported like other features. |
yadvr
commented
Jul 22, 2019
@server24 can you test the PR? |
nvazquez
commented
Jul 22, 2019
@rhtyd @server24 I'll address the open comments today |
blueorangutan
commented
Jul 22, 2019
Trillian test result (tid-188)
|
@server24 I have pushed a fix based on your previous comment, thanks! Regarding network restart with cleanup, it is possible by previously changing the setting 'network.rolling.restart' to false. |
nvazquez
commented
Jul 23, 2019
@blueorangutan package |
blueorangutan
commented
Jul 23, 2019
@nvazquez a Jenkins job has been kicked to build packages. I'll keep you posted as I make progress. |
blueorangutan
commented
Jul 23, 2019
Packaging result: ✔centos6 ✔centos7 ✔debian. JID-170 |
nvazquez
commented
Jul 23, 2019
@blueorangutan test matrix |
blueorangutan
commented
Jul 23, 2019
@nvazquez a Trillian-Jenkins matrix job (centos6 mgmt + xs71, centos7 mgmt + vmware65, centos7 mgmt + kvmcentos7) has been kicked to run smoke tests |
borisstoyanov
left a comment
There was a problem hiding this comment.
I've manually checked the fix on a basic zone and it looks good. Let's wait for the final results
blueorangutan
commented
Jul 23, 2019
Trillian test result (tid-196)
|
nvazquez
commented
Jul 23, 2019
@rhtyd there are still vmware and xenserver test jobs running |
blueorangutan
commented
Jul 23, 2019
Trillian test result (tid-197)
|
Description
CloudStack reports an error while stopping/expunging VMs on basic zones
Fixes: #3497
Types of changes
Screenshots (if appropriate):
How Has This Been Tested?
VM lifecycle