Alarm list is not retrieving alarms more than 2 days
Alarm list is not retrieving alarms more than 2 days for specific HI controllers
Redudant AC800m connectivity servers
Redudant Aspect servers
Are we talking alarms or events here?
The answer is different depending on what you aim for.
- Stored only in RAM of Alarm Manager service
- Default storage is 10.000 shared among all categories
- Storage is First In First Out (FIFO)
- Unless changed, HW alarms may flood and cause process alarms to be pushed out of FIFO
- Do change storage settings to avoid flooding (search AKS for how!)
- An alarm that go inactive and become acknowledged will vanish from system
- Each alarm will be sent as event to Event Storage
- Stored persistently on disk by Event Storage service
- Default storage is 10.000 per category
- FIFO storage
An alarm or event list can only filter for either alarm OR event, not both at same time. Only alarm list offer Alarm State (the square box to acknowledge with).
It is a common problem that alarm storage has not been split up between system and process alarms. A flood of system alarms may then force process alarms out of the FIFO storage. This problem is easy to fix - search AKS for how to split (reserve) separate storage for system vs process categories.
It is also a common problem that event categories wrap earlier than expected. 10.000 is not that much if you have lots of IO and a upset process situation with many alarms and events. Event Storage can be increased to 50.000 events per category (read manual how to do it). An Information Manager can store up to 12.000.000 events, but is less user friendly for operators to search in. SQL access method is preferred, but require extra skills.