APAR status
Closed as program error.
Error description
Topology sync does not exclude ghosted agents when querying for connected agents with offline status. When agent record is ghosted, but the agent reconnects, a new record is created with the same endpoint ID. Topology sync found the ghosted offline record for the online agent and incorrectly tries to correct the discrepency by queueing work to the agent task batcher to set in to online status. The batcher will send an "update metadata" message to the actual connected agent and set its unghosted record to connecting status. When the agent responds, the unghosted record is set back to online. This process will repeat at each topology sync because the two records will have the same status they had at the beginning. In the case that the agent has been cloned and two or more instances are connected to the system, all will respond to the message. Because of factors like different hostnames or IP addresses, the property hash in the in the "update metadata" message will only match for at most one agent. The others will send updated properties to the server, which will load new property sheets into versioned config. Because of the repetition described above, the database will eventually fill with these property sheets.
Local fix
Problem summary
**************************************************************** * USERS AFFECTED: * * All end users on all supported browsers. * **************************************************************** * PROBLEM DESCRIPTION: * * Topology sync does not exclude ghosted agents when querying * * for * * connected agents with offline status. * * * * When agent record is ghosted, but the agent reconnects, a * * new * * record is created with the same endpoint ID. Topology sync * * found * * the ghosted offline record for the online agent and * * incorrectly * * tries to correct the discrepency by queueing work to the * * agent * * task batcher to set in to online status. The batcher will * * send * * an "update metadata" message to the actual connected agent * * and * * set its unghosted record to connecting status. When the * * agent * * responds, the unghosted record is set back to online. This * * process will repeat at each topology sync because the two * * records will have the same status they had at the beginning. * * * * In the case that the agent has been cloned and two or more * * instances are connected to the system, all will respond to * * the * * message. Because of factors like different hostnames or IP * * addresses, the property hash in the in the "update metadata" * * message will only match for at most one agent. The others * * will * * send updated properties to the server, which will load new * * property sheets into versioned config. Because of the * * repetition * * described above, the database will eventually fill with * * these * * property sheets. * **************************************************************** * RECOMMENDATION: * * Fixed in version 7.1.0.0 * ****************************************************************
Problem conclusion
Fix is provided in IBM UrbanCode Deploy 7.1.0.0
Temporary fix
Comments
APAR Information
APAR number
PH22782
Reported component name
UC DEPLOY
Reported component ID
5725M5400
Reported release
704
Status
CLOSED PER
PE
NoPE
HIPER
NoHIPER
Special Attention
NoSpecatt / Xsystem
Submitted date
2020-02-28
Closed date
2020-07-22
Last modified date
2020-07-22
APAR is sysrouted FROM one or more of the following:
APAR is sysrouted TO one or more of the following:
Fix information
Fixed component name
UC DEPLOY
Fixed component ID
5725M5400
Applicable component levels
[{"Business Unit":{"code":"BU053","label":"Cloud & Data Platform"},"Product":{"code":"SS4GSP","label":"IBM UrbanCode Deploy"},"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"704","Line of Business":{"code":"LOB45","label":"Automation"}}]
Document Information
Modified date:
23 July 2020