Hi, I am trying to configure high availability on a critical VM. My ovirt version is 4.3.10.4-1.el7. My question is about the best settings to specify, since it's not completely clear to me how they influence the availability. What I want is simply to have a specific machine ALWAYS running, or at least, to be available as much as possible. I have no special requirements on the VM, if it goes down, simply restarting it is fine. Here is my current setting. Is it good or I'd better chage anything? Thanks, Andrea -- Andrea Chierici - INFN-CNAF Viale Berti Pichat 6/2, 40127 BOLOGNA Office Tel: +39 051 2095463 SkypeID ataruz --
On Tuesday, 2 February 2021 11:37:01 CET Andrea Chierici wrote:
Hi, I am trying to configure high availability on a critical VM. My ovirt version is 4.3.10.4-1.el7. My question is about the best settings to specify, since it's not completely clear to me how they influence the availability.
What I want is simply to have a specific machine ALWAYS running, or at least, to be available as much as possible. I have no special requirements on the VM, if it goes down, simply restarting it is fine.
Here is my current setting. Is it good or I'd better chage anything?
you may want to configure VM lease. If you don't do so, in case of split brain, there can be two VMs running and accessing the same storage, leading to the data corruption. With VM leases this cannot happen - only one VM can access the storage. To speed up the switch, you may want to configure also watchdog. But there can be use cases (e.g. stateless server which just forwards received packet somewhere) where leases are not needed or not using them won't cause any harm. See [1] for more details. [1] https://www.ovirt.org/develop/ha-vms.html
Thanks, Andrea
On Tue, Feb 2, 2021 at 1:57 PM Vojtech Juranek <vjuranek@redhat.com> wrote:
On Tuesday, 2 February 2021 11:37:01 CET Andrea Chierici wrote:
Hi, I am trying to configure high availability on a critical VM. My ovirt version is 4.3.10.4-1.el7. My question is about the best settings to specify, since it's not completely clear to me how they influence the availability.
What I want is simply to have a specific machine ALWAYS running, or at least, to be available as much as possible. I have no special requirements on the VM, if it goes down, simply restarting it is fine.
Here is my current setting. Is it good or I'd better chage anything?
you may want to configure VM lease. If you don't do so, in case of split brain, there can be two VMs running and accessing the same storage, leading to the data corruption. With VM leases this cannot happen - only one VM can access the storage. To speed up the switch, you may want to configure also watchdog. But there can be use cases (e.g. stateless server which just forwards received packet somewhere) where leases are not needed or not using them won't cause any harm. See [1] for more details.
Indeed, configuring a VM lease and a watchdog device can be useful. Not only that a VM lease can generally prevent split brains but in case the host that the VM runs on becomes non-responsive, the system would automatically try to restart the VM elsewhere (otherwise, the system waits for a confirmation from the admin that the host is down which can take time, and yes, if the VM still runs on the non-responsive host can lead to a split-brain). The watchdog device identifies problems from within the guest that are otherwise hard to identify from outside of the virtual machine. Richard Jones wrote a nice post about the watchdog device few years ago [2] [2] https://rwmj.wordpress.com/2010/03/03/what-is-a-watchdog/
Thanks, Andrea
_______________________________________________ Users mailing list -- users@ovirt.org To unsubscribe send an email to users-leave@ovirt.org Privacy Statement: https://www.ovirt.org/privacy-policy.html oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/ List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/IYHLDVK7UNFBHA...
participants (3)
-
Andrea Chierici -
Arik Hadas -
Vojtech Juranek