News, tips, partners, and perspectives for the Oracle Solaris operating system
- February 5, 2018
What is this BUI thing anyway?
The Solaris Analytics WebUI (or "bui" for short) is what we use to tie together all our data gathering from the Stats Store. Comprised of two web apps (titled "Solaris Dashboard" and "Solaris Analytics"), enable the webui service via
# svcadm enable webui/server
Once the service is online, point your browser at https://127.0.0.1:6787 and log in. [Note that the self-signed certificate is that generated by your system, and adding an exception for it in your browser is fine]. Rather than roll our own toolkit, we make use of Oracle Jet, which means we can keep a consistent look and feel across Oracle web applications.
After logging in, you'll see yourself at the Oracle Solaris Web Dashboard, which shows an overview of several aspects of your system, along with Faults (FMA) and Solaris Audit activity if your user has sufficient privileges to read them.
Mousing over any of the visualizations on this page will give you a brief description of what the visualization provides, and clicking on it will take you to a more detailed page.
If you click on the hostname in the top bar (next to Applications), you'll see what we call the Host Drawer. This pulls information from svc:/system/sysstat.
Click the 'x' on the top right to close the drawer.
Selecting Applications / Solaris Analytics will take you to the main part of the bui:
I've select the NFS Client sheet, resulting in the dark shaded box on the right popping up with a description of what the sheet will show you.
Building blocks: faults, utilization and audit events
In the previous installment I mentioned that we wanted to provide a way for you to tie together the many sources of information we provide, so that you could answer questions about your system. This is a small example of how you can do so.
The host these screenshots were taken from is a single processor, four-core Intel-based workstation. In a terminal window I ran
# psradm -f 3Followed a few minutes later by
# psradm -n 3
You can see those events marked on each of the visualizations with a blue triangle here:
Now if I mouseover the triangle marking the second offline/online pair, in the Thread Migrations viz, I can see that the system generated a Solaris Audit event:
This allows us to observe that the changes in system behaviour (primarily load average and thread migrations across cores) were correlated with the offlining of a cpu core.
Finally, let's have a look at the Audit sheet. To view the stats on this page, you need to login to the bui as a suitably-privileged user - either root, or a user with the solaris.sstore.read.sensitive privileges.
# usermod -A +solaris.sstore.read.sensitive $USER
For this screenshot I not only redid the psradm operations from earlier, I also tried making an ssh connection with an unknown user, and logged in on another of this system's virtual consoles. There are many other things you could observe with the audit subsystem; this is just a glimpse:
Tune in next time for a discussion of using the C and Python bindings to the Stats Store so you can add your own statistics.
For some lower-level information, please have a read of my post https://blogs.oracle.com/jmcp/solaris-analytics%3a-an-overview