How are "Archive Manager's disk is full" Alerts calculated?
search cancel

How are "Archive Manager's disk is full" Alerts calculated?


Article ID: 188894


Updated On:


CA Spectrum DX NetOps


I received the alert "Archive Manager's disk is 90%" full alert.  I'm trying to understand what is Spectrum using to determine Archive manager's disk is 90% full. 


Release : All Supported Releases

Component : SPCCSS - Spectrum Core / SpectroSERVER


In Spectrum, Archive Manager process calculates its installed drive used space value at every one minute interval and then compares that calculated value with DISK_ALARM_PERCENTAGE threshold value (By default 90) configured in $(SPECROOT)/SS/DDM/.configrc file.

If the disk used space is more than the configured threshold value, then Archive manager sends an event message:
"The Archive Manager's disk is 74% full"

with eventType as 0x10c11 to SpectroSERVER. SS will internally raises an "ARCH MANGER DISK Capacity has violated Alarm" after processing the eventType(0x10c11)

During calculation, If Archive manager encounters a disk used space value less than the configured threshold value, then there is no event type that tells SpectroSERVER to clear the already existing Alarm. In case Archive Manager finds that the used disk space value comes down below threshold value, and again goes beyond the configured threshold value, then again Archive Manager sends a 0x10c11 event message to SpectroSERVER which will check if there is any already existing disk capacity Alarm or not. If there is an Alarm then it will increment the "Occurrences" column. If not, SS simply creates "Disk capacity violation" Alarm with "Occurrence" as 1.

In Spectrum, we are using the Microsoft Windows standard GetDiskFreeSpaceEx() API for calculating Disk free and used space.
This API commonly used in SpectroSERVER, Archive Manager & SRADMIN processes. 
SpectroSERVER process calculates the DiskFreeSpace while performing the OnlineBackup operation on the VNM model 
and also uses the same process to update "PercentDiskUsed" attribute on the VNM model at every 2 minutes interval.