Current procedure for upgrading 3.5 HE with RHEL7 hosts, specifically chapter 5.2[1], is not good enough, The procedure advises first to upgrade the Manager and then the hosts. While this is sometimes impossible to manage. And specifically - HE and VDSM have the same channel for both 3.5 and 3.6 packages. So, once the host got upgraded using `yum update` command, it will get to latest 3.6 without much control from the end user, if the user does not know he is not supposed to do so. And since running `yum update` on a rhel box is a pretty common operation, we should modify the procedure or create a prevention mechanism to update the hosts to 3.6 before the manager, if this is a problem that the hosts would be updated before the manager. [1] https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Virtualization/3.6/html/Self-Hosted_Engine_Guide/Upgrading_the_Self-Hosted_Engine.html
Also, once the hosts get upgraded before the manager, they start generating the following error in agent logs on the host: ~~~ MainThread::ERROR::2016-05-21 14:06:29,594::config::234::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine.config::(refresh_local_conf_file) Unable to get vm.conf from OVF_STORE, falling back to initial vm.conf ~~~ So, probably the procedure is not good enough.
The issues here appear to be similar to the ones discussed in bug 1328382. To ensure that we cover all of the issues related to the self-hosted engine upgrade procedure, I am closing this one as a duplicate. Please reopen this bug if you think it's necessary to track something else here. *** This bug has been marked as a duplicate of bug 1328382 ***