-
Notifications
You must be signed in to change notification settings - Fork 285
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Backports for UPD-1009 / CP-50580 #5935
Conversation
Backport of f25cee6 Do not check permissions of a task if it is acting on itself. In the case the session is lost, the task needs to be able to update itself. If the session were to be lost, we would not able to retrieve back the correct `auth_user_id`. If the task ids match implies that we know we were already authorized. Signed-off-by: Gabriel Buica <[email protected]> Co-authored-by: Christian Lindig <[email protected]>
Backport adjustments. Signed-off-by: Gabriel Buica <[email protected]> Signed-off-by: Christian Lindig <[email protected]>
Add ``shutdown` to the list of operation that `hard_shutdown` can cancel. Currently, ``shutdown` is missing from the list of operation that `hard_shutdown` needs to cancel. Signed-off-by: Gabriel Buica <[email protected]> Co-authored-by: Christian Lindig <[email protected]>
Adds ``shutdown` to `vm_operation_table`. `VM.shutdown` operations were displayed as `(unknown operation)`. This should display the operation correctly in logs and during `xe task-list`. Signed-off-by: Gabriel Buica <[email protected]> Co-authored-by: Christian Lindig <[email protected]>
Backport 71c3960 Non-running VMs' metrics are stored in the coordinator. When the coordinator is asked about the metrics try to unarchive them instead of failing while trying to fetch the coordinator's IP address. This needs to force the HTTP method of the query to be POST Also returns a Service Unavailable when the host is marked as Broken. Signed-off-by: Pau Ruiz Safont <[email protected]> Signed-off-by: Christian Lindig <[email protected]>
Backport 71c3960 Backport adjustments Signed-off-by: Christian Lindig <[email protected]>
xapi-project/xen-api-libs-transitional#116 updates a library; the CI fails because it depends on it. |
There is also 7fe1955 which says it fixed the underlying cause for 71c3960 (backported here). Do we need this 2nd commit backported too? Also 6bb7702 further simplifies the code, and we probably only tested/reviewed all 3 commits together. They were all part of this PR: https://github.com/xapi-project/xen-api/pull/5855/commits (which has 2 more commits without a ticket, do we need them?) |
Thanks for pointing out additional commits. I'll look into this. We really need to make sure commits have the correct CA as it's easy to miss related commits. |
Backport 7fe1955 Forces users to use an address, instead of being implicit, this avoid the underlying cause for the issue fixed in the previous commit: it allowed a coordinator to call Pool_role.get_master_address, which always fails. Signed-off-by: Pau Ruiz Safont <[email protected]> Co-authored-by: Christian Lindig <[email protected]> Signed-off-by: Christian Lindig <[email protected]>
…ble at Backport 6bb7702 This makes the selection of the action obvious, previously the two booleans made it hazy to understand the decision, and was part of the error why the coordinator tried to get the coordinator address from the pool_role file (and failed badly) Signed-off-by: Pau Ruiz Safont <[email protected]> Signed-off-by: Christian Lindig <[email protected]>
@edwintorok I've backported two more commits |
DCO unhappy about a commit authored by @GabrielBuica with Co-authored-by Christian (and committer Christian), but I've set it to pass, we don't need to be so strict about this, I'd consider it a bug in the DCO app that it doesn't recognize Co-authored-by. To keep DCO happy it probably wants also a sign-off by Christian. |
vm_operation_table