« Previous 1 2 3 4
What's new in Ceph
Well Kept
Improved Dashboard
Finally, I'll take a quick look at the dashboard, which originated with OpenAttic, although it probably no longer has much of its original code on board. However, the wide-ranging functionality of the official Ceph dashboard has grown in leaps and bounds. In addition to monitoring, alerting, and trending (MAT) with Prometheus and Grafana, it now even offers the ability to change cluster configuration details (Figures 3 and 4).
Since adding new OSDs with ceph-mgmt
, logic is always the same process, which can now be started from the dashboard, as well. Basic changes to the CRUSH map are also possible. The developers are continuously fine-tuning the dashboard's user interface, as evidenced by occasional changes to the design and to the program's menus. The dashboard is still not entirely suitable as a slot-in shell replacement, but that is ultimately not its intent. On top of that, it is part of the standard scope of a normal Ceph setup rolled out by cephadm
and ceph-mgmt
. So, if you need a Ceph monitoring dashboard to display on a large TV in the operations room, the Ceph dashboard will serve you well.
Conclusions
Ceph's quality has developed considerably of late. Many classic desired features have been added in the meantime, and the organizational changes behind the scenes do not give rise to worries of the developers stopping their good work any time soon. That said, much is left to be done. No matter how the Ceph developers tweak the software's many performance bottlenecks, RADOS is still not free of the massive latency that is mainly caused by the Ethernet transport path, on the one hand, and the internal CRUSH algorithm on the other. If the developers succeed in making noticeable improvements there, Ceph would be able to open up a completely new target group. Currently, no specific activity in this respect is evident. Future developments remain exciting.
Infos
« Previous 1 2 3 4
Buy this article as PDF
(incl. VAT)