Planned maintenance: PAMS
It is possible to implement a monitoring system as to the "state of health" and performance on Incas system servers; this system is capable of sending alarms in the case of malfunctioning or critical situations.
PAMS is the service that constantly monitors the state of operation of the system and the software, to anticipate the causes of the breakdown allowing preventive maintenance.
An internet connection puts the Incas control room in contact with the customer's system on which the indices to be controlled have been configured.
If these critical thresholds are exceeded an automatic mail alarm system is activated.
The requirements to activate the system:
- SMTP Server
- Broad band Internet line
The first is a necessary condition for the functioning of the procedure that will be installed and is used to send the alarm mail.
The second is essential in order to access the remote server in an efficient way in the case of an intervention to check any alarms.
The aspects to be monitored are not uniform amongst themselves, thus it is possible to act with two different instruments:
- the MRTG console is used to monitor the performance of the server
- in order to check the proper execution of the automatic procedures an auto check module capable of giving an alarm is added to the latter
We set out below examples of the parameters to be monitored that from a historical analysis have been shown to create data loss or performance problems.
|Use of physical disks
(%Total time and %Preferential time)
|Space occupied on the logical disks
|Fixed memory occupied
(Bytes available and bytes tied up)
|Space occupied inside the database
(Mb free inside the tablespace)
|Utilisation of the disks
(reading and writing)
(number of instructions carried out)
(volume of transactions generated)
|Number of sessions||Database||MRTG|
Applicative parameters monitored via MRTG console
Via the MRTG console one controls the arrangement of the records inside the data exchange tables with the management system.
In particular the total number of records inside the tables: HOST_TX and HOST_RX are graphically represented, comparing them with the number in status 0.
Sending the alarm
Alarm threshold values can be set on every monitored parameter. If these values are exceeded an alarm event is generated and an e-mail is automatically sent to one or more e-mail addresses.
Certain threshold activities are activated on the server, which execute backup and database maintenance procedures on a daily basis.
Once terminated, these procedures are capable of checking for any errors and consequently the problem can be signalled via e-mail.
It is also possible to have a signal of correct operation.
Activation of the monitoring procedure
The installation time and the time required to create the monitoring environment on the customer's server is 4 hours.
Using the same technological approach the historical sequence of failure causes of the various equipment was analysed and the planned control strategies were defined.