It is particulary a problem for cloudboot hypervisors. When we restart it, all logical volumes that reside on a fibre channel datastore are being activated. This slows down the booting process and produces lots of input/output errors.
For static hypervisors, we can handle this by editing the lvm.conf in the custom config of the hypervisor. However, this is not applicable to cloudboot hypervisors because the volumes are being activated early in the booting process and are already active when the custom config is applied.
Please provide some solution, so the volumes don't get activated on every restart. You could for example edit the lvm.conf file in the cloudboot image (auto_activation_volume_list = ["vg list"]) or apply a patch to the os release.
Kind regards !
Please sign in to leave a comment.