* 4.20:
VR: fix site-2-site VPN if split connections is enabled (#10067)
UI: fix cannot open 'Edit tags' modal for static routes (#10065)
Update ownership selection component to be language independent (#10052)
Support to enable/disable VM High Availability manager and related alerts (#10118)
- Adds new config 'vm.ha.enabled' with Zone scope, to enable/disable VM High Availability manager. This is enable by default (for backward compatibilty).
When enabled, the VM HA WorkItems (for VM Stop, Restart, Migration, Destroy) can be created and the scheduled items are executed.
When disabled, new VM HA WorkItems are not allowed and the scheduled items are retried until max retries configured at 'vm.ha.migration.max.retries' (executed in case HA is re-enabled during retry attempts), and then purged after 'time.between.failures' by the cleanup thread that runs regularly at 'time.between.cleanup'.
- Adds new config 'vm.ha.alerts.enabled' with Zone scope, to enable/disable alerts for the VM HA operations. This is enabled by default.
* 4.20:
VR: apply iptables rules when add/remove static routes (#10064)
Certificate and VM hostname validation improvements (#10051)
set ulimit for server according to redhat spec (#10040)
kvm-storage: provide isVMMigrate information to storage plugins (#10093)
Allow config drive deletion of migrated VM, on host maintenance (#10045)
linstor: improve heartbeat check with also asking linstor (#10105)
server: simplify role change validation (#9173)
UI: create VPC network offering with conserve mode (#10082)
server: fix typo removeaccessvpn in VirtualRouterElement (#10086)
UI: remove duplicated Instance Name in Public IP details page (#10087)
UI: Fixes in the Usage UI (#10000)
SAML2: add cookie with HttpOnly too #10013 (#10047)
ui: Allow font-awesome icon usage and optimise icon size inconsistency (#9744)
Particular Linstor needs can use this information to only allow
dual volume access for live migration and not enable it in general,
which can and will lead to data corruption if for some reason
2 VMs get started on 2 different hosts.
If a node doesn't have a DRBD connection to another node,
additionally ask Linstor-Controller if the node is alive.
Otherwise we would have simply said no and the node might still be alive.
This is always the case in a non hyperconverged setup.
* fixes on usage ui
* dont call API when caller does not have access to it
* handle when lastheartbeat is undefined
* hide last collection datetime when lastheartbeat is undefined
* apply Pearl's suggestions
Co-authored-by: Pearl Dsilva <pearl1594@gmail.com>
* apply Pearl's suggestions
Co-authored-by: Pearl Dsilva <pearl1594@gmail.com>
* apply Pearl's suggestions
Co-authored-by: Pearl Dsilva <pearl1594@gmail.com>
* apply Pearl's suggestions
Co-authored-by: Pearl Dsilva <pearl1594@gmail.com>
* apply Pearl's suggestions
Co-authored-by: Pearl Dsilva <pearl1594@gmail.com>
---------
Co-authored-by: Pearl Dsilva <pearl1594@gmail.com>
John Bampton
Daan Hoogland
Wei Zhou
Bryan Lima
Suresh Kumar Anaparti
Pearl Dsilva
Bernardo De Marco Gonçalves
Rene Peinthor
Abhishek Kumar
Ruben Bosch
Fabricio Duarte
Rohit Yadav
Felipe
Lucas Martins