Project Description

WebFamulus is a web application using the Quartz scheduler framework for process automation. It provides a web interface to schedule, monitor, and interact with Java processes that are configured as 'jobs'. Access to user and administrative interfaces is based on permissions and group ownership. Likewise, the configuration of jobs, alerts, performance monitoring and statistics is based on group ownership and permissions granted to an individual user. The application is currently under development and a beta-version is available to selected clients.


Typical applications for the Quartz framework are network and hardware monitors, database maintenance, reporting, etc. Tedious, time-consuming tasks can run repeatedly whenever possible following sophisticated schedules, including selected dates and weekdays, black-out times, holidays, etc. The field of application, however, is limited only by what the programming language cannot provide. Quartz makes it possible to schedule and execute automatically just about anything that can be programmed in Java.


WebFamulus, in addition, supports jobs to be executed as sequences based on the (optional) evaluation of the preceding job's exit status. Furthermore, the application is backed by an extensive database that makes it possible to let non-administrative users create, configure, interact with, and monitor their own jobs, alerts, and reports on-line or via electronic mail.


2013-12-05

3. SMART temperature surveys

Data Collection

Hard disks with SCT (SMART Command Transport) capabilities maintain a log of the working temperature of the drive. Most drives register entries at one minute intervals. The header of the SCT section of the smartctl output indicates the registration interval.

SCT-enabled drives are capable of recording temperatures of ±127°C. When the device is powered up, an illegal value is entered so that the starting point is is registered. (In smartctl output, this value appers as as a question mark '?'.) The log itself is a circular buffer. Hence, when the temperature log is read, only the values after the last question mark (if there is one) have a correct time reference (namely, N entries at the logging interval time before the date and time the SMART device was queried).

Reporting

I. Daily Reports

Temperature surveys reflect usage patters of computers. The following graph illustrates the typical use of laptop; this one having been put into sleep mode twice on the day sampled.


The same laptop running continuously as a low-volume server.




The following example shows an office workstation during a regular pattern utilization. The machine is used for image processing.



The same machine showing a pattern of high utilization. This pattern, a high temperature maintained over a long period of time, occurs occasionally. The pattern, when it occurs, is always the same: an ascending curve to a high temperature that maintained for several hours. Unfortunately, it has been impossible to ascertain what software was running at the time. The staff reported a normal use. A mechanical problem, like a malfunctioning fan, did not appear to have been the cause and was ruled out.



Monthly Reports

Summary reports over longer time periods help to pinpoint occurrences of irregularities. Of particular interest are, of course, unusual high and low temperatures. A rise in the average curve is indicative a prolonged operation a higher temperatures. (See 09/13 and 09/15, which reflect a brief and prolonged high disk use, respectively.) Upward spikes in low temperature values are due to the fact that the first SMART test of the day ran some time after the computer was turned on.



Annual Reports

Temperature measurements from computers that do not operate continuously—desktop workstations and laptops, for example—reflect the ambient temperatures. The data of the following graph was collected from a laptop that is usually turned off over night. This usually provided the lowest temperature of the day. (The high spikes in the low temperature are an indication that a SMART test was not run within the first two hours after computer was started.)
Since the power supply that recharges the battery is connected, the temperatures inside the casing are of course higher than the ambient room temperature. The low temperature range (in blue) before July and in September reflects seasonal temperatures in Italy: warm, summerly temperatures before July and in September, fall temperatures in October and November, and finally the onset of winter in January. The periods of July, August, mid-December and early January, were measured in Germany: colder than Italy during the summer, but warmer in the winter. Interestingly, unlike the ambient temperatures in Italy, those in Germany remain more or less constant. Conclusion: Building construction in Germany proves to be more weatherproof!

No comments:

Post a Comment