On Sun, Aug 8, 2021 at 7:53 PM Gilboa Davara <gilboad@gmail.com> wrote:
Hello all,
During the night, one of my (smaller) setups, a single node self hosted engine (localhost NFS) crashed due to what-looks-like a massive disk failure (Software RAID6, with 10 drives + spare). After a reboot, I let the RAID resync with a fresh drive) and went on to start oVirt. However, no such luck. Two issues: 1. ovirt-ha-broker fails due to broken hosted engine state (log attached). 2. ovirt-ha-agent fails due to network test (tcp) even though both remote-host and DNS servers are active. (log attached).
Two questions: 1. Can I somehow force the agent to disable the network liveliness test? 2. Can I somehow force the broker to rebuild / fix the hosted engine state?
- Gilboa
FWIW switching agent network test to none (via hosted-engine --set-shared-config network_test none --type=he_local) doesn't seem to work. (Unless I'm missing the point and the agent is failing due to broker issues and not due to a failed network liveliness check). - Gilboa