Actually I was using a bash script works like semaphore in
before/after_vm_start. As a result that a parallel problem occured: four
vms read the same number at the same time, so I abandoned this way.
Now I just add a random delay period before vm start, not very well.
2013/5/24 Itamar Heim <iheim(a)redhat.com>
On 05/23/2013 04:23 PM, lofyer wrote:
> Can host hooks be used to implement the following policy to avoid "boot
> storm"?
>
> 1.count and sort VMs awaiting for boot.
> 2.If the count is less than or equal to, for example 5, then boot all
> VMs and exit the script.
> 3.Otherwise, boot and dequeue vm0 to vm4, and minus the count by 5.
> 4.Sleep some time, and go to step 2.
> ______________________________**_________________
> Users mailing list
> Users(a)ovirt.org
>
http://lists.ovirt.org/**mailman/listinfo/users<http://lists.ovirt.org...
>
hooks run one on vm-start one by one. would be a bit tricky to do the
above.
roy - didn't you add some mechanism to throttle boot storms in the engine?