Symptoms

The issue is actual for OSA 6.0 version only.

OA Control Panel becomes unavailable during "Get usage data from PACI Instance Manager" periodic task run.

OA Management Node runs on top of a server with >16 processor units (>64 CPU cores).

The task log shows:

Feb  8 10:02:14 core : DBG [task:18627834:1071 1:4983:ee8f7b90 PACI ]: [ PACI::PACI_impl::runUpdateSubscriptionUsage] Detected 16 CPUs on master node
Feb  8 10:02:14 core : DBG [task:18627834:1071 1:4983:ee8f7b90 PACI ]: [ PACI::PACI_impl::runUpdateSubscriptionUsage] Use 64 threads to update subscription usage

The issue occurs in case there are simultaneous change orders getting processed for OACI subscriptions.

Cause

The issue is acknowledged with internal ID POA-109894.

The amount of threads, automatically calculated for the periodic task, is too large and may lead to a system deadlock in case of additional activities on changing subscription limits (e.g. change orders).

Resolution

To minimize the impact from the task, reschedule it to the night time.

Contact TAM/PTA to get a permanent resolution for the issue.

Internal content

Link on internal Article