Hide Forgot
Description of problem: Among all hosts if one host dont have connection with storage and if we try adding storage domain then, storage domain gets added successfully. Further another host which don't have connection with storage, move into non-operational state. RHEV should check if all hosts in Data center have access to the storage domain, before adding the storage domain. If any of the host don't have access then, RHEV should throw some error message while adding storage it-self. For ex: "host <hostname> dont have access to the storage". Or storage domain can be added, but activation of new storage domain should fail if one of the host fails to access it. Hypervisor should not go to Non-Operational while trying to activate a new storage domain. Version-Release number of selected component (if applicable): RHEV3.2/3.2 How reproducible: 100% Steps to Reproduce: 1. Add two hosts. 2. unmap any of the host from storage. 3. Try creating new storage domain. Actual results: Among all hosts if one host dont have connection with storage and if we try adding storage domain then, storage domain gets added successfully. Further another host which don't have connection with storage, move into non-operational state. Expected results: I think, In 3.2, storage domain can be added, but activation of new storage domain should fail if one of the host fails to access it. Hypervisor should not go to Non-Operational while trying to activate a new storage domain. Or RHEV should check if all hosts in Data center have access to the storage domain, before adding the storage domain. If any of the host don't have access then, RHEV should throw some error message while adding storage it-self. For ex: "host <hostname> dont have access to the storage". Additional info:
Allon, how do you want to handle it? possibly we can make it to be a 3.5 RFE, basically same issue as "prepare for maintenace" (here it will be "prepare for activation or something like that) if we do want to proceed with it.
(In reply to Liron Aravot from comment #4) > Allon, how do you want to handle it? possibly we can make it to be a 3.5 > RFE, basically same issue as "prepare for maintenace" (here it will be > "prepare for activation or something like that) if we do want to proceed > with it. We need to see how this fits in our 3.5 capacity, but definitely an RFE that should be tackled.
could this bring down many hypervisors or even whole clusters if they can't access the storage domain? I only had the problem with one hypervisor like the the original reporter.
This is fixed in 4.1, can be closed
BZ<2>Jira Resync