Xen Orchestra 5.58
Xen Orchestra 5.58 is available and comes with the third wave of improvements regarding our backup features.
It's april's release time and the third wave of improvements for Xen Orchestra backup features are coming with it. In addition, we are also making progress regarding our Amazon S3 plugin as well as our load balancing features which will become more and more complete in the next few months.
NFS version update
Xen Orchestra is no longer using NFS 3 version for NFS remote. We are now using the highest available version of NFS available. It should brings better performances and security for the concerned remote.
However, if you want to stick with the previous behavior, you can put
vers=3 in the field
Custom mount options in the remote view of your appliance.
Xen Developers & Design summit
This year again we are proud to sponsor the Xen Developers & Design summit alonside Arm!
As last year, because of the COVID-19 pandemics, the event will be held online from 25th to 28th of May. You can visit the official website to register.
And this year, we have 4 talks!
Olivier, our CEO, will talk about the emergence of DPU in the datacenter and provide some insight about XCP-ng project state and perspectives.
Bobby, our Xen specialist, will have also 2 talks:
Backup performance & security improvements
Xen Orchestra 5.58 is the third wave of improvements regarding our backup features.
Recoverable VHD merge
During delta backup jobs, an important phase of the process is the merging of the VHD files on the remote. Until now, this merging was one of the weak points of the solution, because an interruption of the job at this point often meant that the entire backup chain was corrupted. This meant it now had to redo a full backup the next time the job ran in order to restart on a healthy basis.
With our new backup code, we have now introduced the ability to recover from an interruption during this step. This is a big step forward in improving the reliability and robustness of our delta backup functionality.
Lock VMs dir during backup
We have added an extra security layer in our backup features to avoid corruption by locking the VMs backup directory on the remote when a backup job is running.
The main objective of this lock is to avoid a single backup job running twice on the same VM, at the same moment. This event could occur in some rare situations. For example, if you deploy a new appliance, using the configuration of your old appliance, without deactivating the previous one.
Clean VM during backup job
Before starting a backup job, there will now be an automatic cleaning of the VM's backup directory on the remote.
This operation will:
- delete VHDs, XVAs files that are not linked to a backup
- delete broken VHDs
- delete delta VHDs with missing parent
- delete backups with missing files (VHDs or XVAs)
This cleaning function will allow to check the integrity of the backups, to avoid bad surprises at the restoration and also make free space on disk by checking that all broken or useless files or deletes as expected.
VDI destruction retrial
When a backup job is over, we are supposed to destroy the VDI used for the job. We noticed that sometimes, the VDI was not ready to be destroyed at the end of the job (as it should be). This issue is an XAPI bug, and we are currently working with the Xen developer team in order to fix this.
However, as a workaround, we have implemented an automatic retry of the VDI destruction when it fails - 10 retries every 5 seconds.
Create template from snapshot
You could already create VMs from a snapshot - now you have the ability to create a template from a snapshot.
Display proxy upgrade on XO updates
Until now, when updates were available for the Xen Orchestra proxy, you had to go into the proxy view to get the information which was not very convenient.
Available proxy updates are now visible directly in the update view of your main Xen Orchestra Appliance.
[Beta] Amazon S3 improvement
We fixed a bug impacting our AWS S3 plugin and also added the ability to select the region, if you want to backup is specific places. Please test and report!
Ignore load balancing VMs via tag
You can now ignore VMs in the load balancer plugin by using tags. We also did various load balancer improvements to fit in more use cases. On the long run, this plugin will be more and more a "resource scheduler" than just a load balancer.
The beta of XOSTOR has started. If you are interested in our hyperconvergence solution developed jointly with LINSTOR, the DRBD technology specialists - you can register on this page.