]
eedri updated OVIRT-1491:
-------------------------
Resolution: Duplicate
Status: Done (was: To Do)
Change oVirt DC production storage
-----------------------------------
Key: OVIRT-1491
URL:
https://ovirt-jira.atlassian.net/browse/OVIRT-1491
Project: oVirt - virtualization made easy
Issue Type: Task
Reporter: eedri
Assignee: infra
Priority: Highest
It seems that during the history of the DC, there wasn't one time we really used the
DRDB features and it makes our storage configuration more complex and hard to
maintain/debug than it brings value.
Lets think on a new SIMPLE storage setup that will increase the availability of oVirt
services and will prevent incidents like before where we have a single point of failure if
the storage is down. without multiple services that adds more failure points and
complexity to the setup.
One suggestion was to install 2 new storage servers running ISCSI and connect them
independently to oVirt and spread the load over both of them, so if one is down the other
will keep running.
We can consider having a few redundant VMs on both servers if possible like the resources
server and maybe others as well.
Please add more suggestions on this ticket, it should be our number one priority right
now.