[jira] [Commented] (CLOUDSTACK-9763) vpc: can not ssh to instance after vpc restart
[ https://issues.apache.org/jira/browse/CLOUDSTACK-9763?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16134247#comment-16134247 ] ASF subversion and git services commented on CLOUDSTACK-9763: - Commit ea52a1312f2345dc2cdcbff31e326b7202453b42 in cloudstack's branch refs/heads/debian9-systemvmtemplate from [~serbaut] [ https://gitbox.apache.org/repos/asf?p=cloudstack.git;h=ea52a13 ] CLOUDSTACK-9763: set ssh public key when creating vm data (#1919) > vpc: can not ssh to instance after vpc restart > -- > > Key: CLOUDSTACK-9763 > URL: https://issues.apache.org/jira/browse/CLOUDSTACK-9763 > Project: CloudStack > Issue Type: Bug > Security Level: Public(Anyone can view this level - this is the > default.) > Components: Virtual Router, VPC >Affects Versions: 4.8.0 >Reporter: Joakim Sernbrant > > Restart with Cleanup of a VPC does not update the public-key metadata, it is > explicitly set to null in > https://github.com/apache/cloudstack/blob/master/server/src/com/cloud/network/router/CommandSetupHelper.java#L614 > Rebooting instances relying on metadata (e.g. coreos) will no longer have the > correct public key configured. > Added explanation: > The VPC VR maintains metadata > (http://docs.cloudstack.apache.org/projects/cloudstack-administration/en/4.8/virtual_machines/user-data.html) > as static files in /var/www/html/metadata. When a VR is destroyed and > recreated (by e.g. "restart with cleanup") this metadata is rebuilt by > createVmDataCommandForVMs(). public-keys is missing from that function so it > becomes empty after the rebuild and a request for > latest/meta-data/public-keys no longer returns the correct key. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Commented] (CLOUDSTACK-9763) vpc: can not ssh to instance after vpc restart
[ https://issues.apache.org/jira/browse/CLOUDSTACK-9763?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16134229#comment-16134229 ] ASF subversion and git services commented on CLOUDSTACK-9763: - Commit ea52a1312f2345dc2cdcbff31e326b7202453b42 in cloudstack's branch refs/heads/master from [~serbaut] [ https://gitbox.apache.org/repos/asf?p=cloudstack.git;h=ea52a13 ] CLOUDSTACK-9763: set ssh public key when creating vm data (#1919) > vpc: can not ssh to instance after vpc restart > -- > > Key: CLOUDSTACK-9763 > URL: https://issues.apache.org/jira/browse/CLOUDSTACK-9763 > Project: CloudStack > Issue Type: Bug > Security Level: Public(Anyone can view this level - this is the > default.) > Components: Virtual Router, VPC >Affects Versions: 4.8.0 >Reporter: Joakim Sernbrant > > Restart with Cleanup of a VPC does not update the public-key metadata, it is > explicitly set to null in > https://github.com/apache/cloudstack/blob/master/server/src/com/cloud/network/router/CommandSetupHelper.java#L614 > Rebooting instances relying on metadata (e.g. coreos) will no longer have the > correct public key configured. > Added explanation: > The VPC VR maintains metadata > (http://docs.cloudstack.apache.org/projects/cloudstack-administration/en/4.8/virtual_machines/user-data.html) > as static files in /var/www/html/metadata. When a VR is destroyed and > recreated (by e.g. "restart with cleanup") this metadata is rebuilt by > createVmDataCommandForVMs(). public-keys is missing from that function so it > becomes empty after the rebuild and a request for > latest/meta-data/public-keys no longer returns the correct key. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Commented] (CLOUDSTACK-9763) vpc: can not ssh to instance after vpc restart
[ https://issues.apache.org/jira/browse/CLOUDSTACK-9763?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15887059#comment-15887059 ] ASF GitHub Bot commented on CLOUDSTACK-9763: Github user pdion891 commented on the issue: https://github.com/apache/cloudstack/pull/1919 This is usefull when using VPC. +1 > vpc: can not ssh to instance after vpc restart > -- > > Key: CLOUDSTACK-9763 > URL: https://issues.apache.org/jira/browse/CLOUDSTACK-9763 > Project: CloudStack > Issue Type: Bug > Security Level: Public(Anyone can view this level - this is the > default.) > Components: Virtual Router, VPC >Affects Versions: 4.8.0 >Reporter: Joakim Sernbrant > > Restart with Cleanup of a VPC does not update the public-key metadata, it is > explicitly set to null in > https://github.com/apache/cloudstack/blob/master/server/src/com/cloud/network/router/CommandSetupHelper.java#L614 > Rebooting instances relying on metadata (e.g. coreos) will no longer have the > correct public key configured. > Added explanation: > The VPC VR maintains metadata > (http://docs.cloudstack.apache.org/projects/cloudstack-administration/en/4.8/virtual_machines/user-data.html) > as static files in /var/www/html/metadata. When a VR is destroyed and > recreated (by e.g. "restart with cleanup") this metadata is rebuilt by > createVmDataCommandForVMs(). public-keys is missing from that function so it > becomes empty after the rebuild and a request for > latest/meta-data/public-keys no longer returns the correct key. -- This message was sent by Atlassian JIRA (v6.3.15#6346)
[jira] [Commented] (CLOUDSTACK-9763) vpc: can not ssh to instance after vpc restart
[ https://issues.apache.org/jira/browse/CLOUDSTACK-9763?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15865943#comment-15865943 ] ASF GitHub Bot commented on CLOUDSTACK-9763: Github user rafaelweingartner commented on the issue: https://github.com/apache/cloudstack/pull/1919 Thanks for the explanation @serbaut. That is exactly why I asked; so, it is not a reboot/restart per se. It is a re-deploy; the old virtual machine is destroyed and a new one is deployed (or maybe a reset of the VHD of the VM). By VM here I mean VR (which at the end of the day is a VM). @serbaut, could you add these explanations on the PR description and Jira ticket (https://issues.apache.org/jira/browse/CLOUDSTACK-9763)? At least for me, this was not clear. Thanks for the fix ;) Code LGTM > vpc: can not ssh to instance after vpc restart > -- > > Key: CLOUDSTACK-9763 > URL: https://issues.apache.org/jira/browse/CLOUDSTACK-9763 > Project: CloudStack > Issue Type: Bug > Security Level: Public(Anyone can view this level - this is the > default.) > Components: Virtual Router, VPC >Affects Versions: 4.8.0 >Reporter: Joakim Sernbrant > > Restart with Cleanup of a VPC does not update the public-key metadata, it is > explicitly set to null in > https://github.com/apache/cloudstack/blob/master/server/src/com/cloud/network/router/CommandSetupHelper.java#L614 > Rebooting instances relying on metadata (e.g. coreos) will no longer have the > correct public key configured. -- This message was sent by Atlassian JIRA (v6.3.15#6346)
[jira] [Commented] (CLOUDSTACK-9763) vpc: can not ssh to instance after vpc restart
[ https://issues.apache.org/jira/browse/CLOUDSTACK-9763?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15865547#comment-15865547 ] ASF GitHub Bot commented on CLOUDSTACK-9763: Github user ustcweizhou commented on the issue: https://github.com/apache/cloudstack/pull/1919 I agree with @serbaut , different from password which is applied to VR only once, the public keys should be set in VR each time when we recreate a VR. this LGTM+1 > vpc: can not ssh to instance after vpc restart > -- > > Key: CLOUDSTACK-9763 > URL: https://issues.apache.org/jira/browse/CLOUDSTACK-9763 > Project: CloudStack > Issue Type: Bug > Security Level: Public(Anyone can view this level - this is the > default.) > Components: Virtual Router, VPC >Affects Versions: 4.8.0 >Reporter: Joakim Sernbrant > > Restart with Cleanup of a VPC does not update the public-key metadata, it is > explicitly set to null in > https://github.com/apache/cloudstack/blob/master/server/src/com/cloud/network/router/CommandSetupHelper.java#L614 > Rebooting instances relying on metadata (e.g. coreos) will no longer have the > correct public key configured. -- This message was sent by Atlassian JIRA (v6.3.15#6346)
[jira] [Commented] (CLOUDSTACK-9763) vpc: can not ssh to instance after vpc restart
[ https://issues.apache.org/jira/browse/CLOUDSTACK-9763?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15865524#comment-15865524 ] ASF GitHub Bot commented on CLOUDSTACK-9763: Github user serbaut commented on the issue: https://github.com/apache/cloudstack/pull/1919 The VPC VR maintains metadata (http://docs.cloudstack.apache.org/projects/cloudstack-administration/en/4.8/virtual_machines/user-data.html) as static files in /var/www/html/metadata. When a VR is destroyed and recreated (by e.g. "restart with cleanup") this metadata is rebuilt by createVmDataCommandForVMs(). public-keys is missing from that function so it becomes empty after the rebuild and a request for latest/meta-data/public-keys no longer returns the correct key. This PR adds public-key to the rebuild. > vpc: can not ssh to instance after vpc restart > -- > > Key: CLOUDSTACK-9763 > URL: https://issues.apache.org/jira/browse/CLOUDSTACK-9763 > Project: CloudStack > Issue Type: Bug > Security Level: Public(Anyone can view this level - this is the > default.) > Components: Virtual Router, VPC >Affects Versions: 4.8.0 >Reporter: Joakim Sernbrant > > Restart with Cleanup of a VPC does not update the public-key metadata, it is > explicitly set to null in > https://github.com/apache/cloudstack/blob/master/server/src/com/cloud/network/router/CommandSetupHelper.java#L614 > Rebooting instances relying on metadata (e.g. coreos) will no longer have the > correct public key configured. -- This message was sent by Atlassian JIRA (v6.3.15#6346)
[jira] [Commented] (CLOUDSTACK-9763) vpc: can not ssh to instance after vpc restart
[ https://issues.apache.org/jira/browse/CLOUDSTACK-9763?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15864397#comment-15864397 ] ASF GitHub Bot commented on CLOUDSTACK-9763: Github user rafaelweingartner commented on the issue: https://github.com/apache/cloudstack/pull/1919 @serbaut I would only like to make a few questions to see if I understood the issue here, before evaluating the code any further. When you first deploy the VM, the keys are delivered to the newly deployed VM, right? The problem only happens on reboots? Do we need to always keep sending the same access keys? Is not one time enough? Of course, if the keys are changed, we should send them; but otherwise, it feels that we should not need to keep sending them. > vpc: can not ssh to instance after vpc restart > -- > > Key: CLOUDSTACK-9763 > URL: https://issues.apache.org/jira/browse/CLOUDSTACK-9763 > Project: CloudStack > Issue Type: Bug > Security Level: Public(Anyone can view this level - this is the > default.) > Components: Virtual Router, VPC >Affects Versions: 4.8.0 >Reporter: Joakim Sernbrant > > Restart with Cleanup of a VPC does not update the public-key metadata, it is > explicitly set to null in > https://github.com/apache/cloudstack/blob/master/server/src/com/cloud/network/router/CommandSetupHelper.java#L614 > Rebooting instances relying on metadata (e.g. coreos) will no longer have the > correct public key configured. -- This message was sent by Atlassian JIRA (v6.3.15#6346)
