1.1.1 The threshold alarm template of weekend is not valid.
[Testing process]
1. Set the threshold template of CPU as weekend and the threshold is 10.
2. Wait a few minutes without reporting CPU threshold alarm;
[Problem Location]
1. Set the log level to INFO by setLogLevel.sh tool.
2. Looking at the log of pms, we find that the baseline template is everyday, and we can't get the weekend template.
3. Increase the debugging log and print out the day of week = 0 for today (Sunday) (range 0-6);
4. But in the code, it is weekend if DayOfWeek is greater than or equal to 5 and workday if DayOfWeek is less than or equal to 5.
5. The actual situation is that the value of DayOfWeek corresponds to Monday to Friday when it is 1-5; the value of DayOfWeek equals 6:00 is Saturday, and the value of DayOfWeek equals 0:00 is Sunday.
[Amendment Program]
According to the above analysis results, add judgment, when DayOfWeek is greater than or equal to 1 and less than or equal to 5, return to the workday template;
When DayOfWeek equals 0 or 6, return the weekend template
1.1.2 [Warning] Stop the main OMS reporting that many services are unavailable
1. Report HDFS service unavailability alarm after main and standby switching. This is known to have been amended bill of lading.
Problem analysis:
When checking the unavailability alarm of HDFS service, we did not consider the initial installation stage. We only judged whether the number of name services of BAD was equal to the number of all current name services. But in the installation stage, the number of name services was 0 and bad was 0, which led to the false alarm of unavailability alarm of HDFS service.
Print all name services in the Controller log in BAD state. so need to send HDFS alarm.
However, there is no alarm for anomalous name service on the interface. The alarm ID is 14010.
The single number is DTS20150705959.
Entry time: 2015/7/9
The environment Manager version is:
Fusion Insight Fuison Insight Manager V100R002C50 2015-07-01 17:43
