"No data" image shown on dashboard instead of usage graphs
This is because metric names provided by monitor-client and extracted from graphite do not match. There are two causes for this:
- Red Hat uses a more advanced parametering of the kvm/qemu executable, messing up monitor-client which parses the kvm command line to determine vm name (used in metric name). See Red Hat Bug 1140121 & Bug 1276485
- If host name of node is set in /etc/hostname or with hostname command as an FQDN (e.g. cloud-12345.vm.ik.bme.hu instead of cloud-12345), metrics are sent with the FQDN name as metric name, but extracted by the portal with the short host name as metric name. From hostname man page:
"/etc/hostname Historically this file was supposed to only contain the hostname and not the full canonical FQDN. Nowadays most software is able to cope with a full FQDN here."Solution/workaround: do not use FQDN in /etc/hostname. Should be documented in aquickstart.md
Please
register
or
sign in
to comment