Symptoms

A periodic task with the name Office 365 Synchronization with Office 365 Portal fails sometimes with 45 minutes timeout.

Cause

When the task satrts, it runs the following query to the Office 365 endpoint: /aps/2/resources/ca97eb1d-3aa5-44c5-85f8-53fe5dd05f0a/azureSynchronize, and waits for its completion.

At this query execution, a block of 100 (or less) most outdated tenants is taken and processed. The query duration depends on amount of the objects to sync, such as users, domains, etc. Sometimes the quey execution time runs out of the default REST timeout of 45 minutes, therefore the task does not receive a response, despite the sync itself was successful:

2017-08-20 20:51:27,972 [  19] DEBUG aps_endpoint: 6363885908797297207 - Request: GET https://10.0.10.5/O365App/aps/office365app/0de04eeb-6666-4181-94ab-9eebc24203a4/azureSynchronize
...
2017-08-20 21:48:36,223 [  19] DEBUG aps_endpoint: Office365AppController.AzureSynchronize finished

Resolution

Such timeout error can be ignored, since it does not affect the synchronization process. However, the timeout issue can be fixed by increasing REST timeout value:

  1. Use setRestTimeout.py python script on OA Management Node to set the required timeout for the operation to 1 hour:

    # python setRestTimeout.py 60

  2. Restart pa-agent service (no need to restart pau, it is only needed to reload SaaS):

    # service pa-agent stop
# service pa-agent start

See also:

KB130404: readCSPAccounts.py fails in 45 minutes: No message recieved

Internal content