Archive manager is shown Down in OC-Console even task is up and running

Document ID : KB000121295
Last Modified Date : 16/11/2018
Show Technical Document Details
Introduction:
Primary SpectroSERVER Archive manager is shown down for some reason as well as secondary spectrosevrer is also down.
Checking the "Archivemanager" tasks (ArchMgr executable) is active in the task-list and ./SS/DDM/ARCHMGR.OUT looks fine.
 
Background:
Communication facility / toolset used to establish communication for Oneclick-server to ArchiveManager is the CORBA toolset. 
Means - the OneClick server implementation is checking and "registering" to the ArchMgr-CORBA resource object running at the SpectroSERVER host.
The CORBA resource object name is CsCEventDomain

Make use of tool -> ./bin/VBNS/ nsutil list <hostname>   to list the available CORBA objects/resources when all Spectrum services are running.

It appears under very rare conditions, the ArchiveManager startup is fine and runs stable but the CORBA resource object CsCEventDomain is not created. 
This then causes at OneClick service to show the ArchiveManager (Events) being down. 
 
Environment:
This applies to all CA Spectrum releases and used installation platform
Instructions:
At primary SpectroSERVER it appears the CORBA resource layer was "garbaged". Even ArchMgr is running and at TCP level and the port 14003 (14003 is the ArchiveManager CORBA socket) is open and requested by "ArchMgr" executable (see netstat -an[b|p] output (differs for linux/windows)), there is no CORBA resource allocation to CsCEventDomain possible.

Checking for the CORBA resources/objects per command ./bin/VBNS/nsutil list <hostname> .
Output looks like:
[spectro@labspec03 VBNS]$ ./nsutil list labspec03
Bindings in labspec03
Object : LmtManagerDomain
Object : SLMDomain
Object : SpectrumICMP
Object : RTMDomain
Object : CsCLocServMapInt
Object : TelcoManagerDomain
Object : CsCEventDomain              # this is the ArchiveManager
Object : ScmService                  # this is the NetworkConfigManager task
Object : EventDispDomain
Object : CsCModelDomain              # this is the SpectroSERVER
Object : SnmpServ
Object : REXConnectorDomain
Object : ProcedureDomain
Object : OnlineMibImport
Object : CsCAlarmDomain



So - even we did stop and restart single service (here the ArchMgr) - it did not bring back the CORBA resource CsCEventDomain -- so we finally did a full Spectrum services stop and startup for processd first - and SpectroSERVER then again. This brought back the correct CORBA resources so the OC-servers could proper connect.