Nov 28 2019 03:36 AM
Nov 28 2019 03:36 AM
I have configured hybrid federated search from SharePoint server to SPO. It works but we are experiencing a strange behavior. When we do a search from the search center (SP on-prem) from time to time we need to press search several times in order to get a search result from SPO.
I have traced it down in the ULS and see the following entry every time when the search fail to get a result from SPO:
NodeRunner.exe (0x1C08) 0x24F4 SharePoint Server Search Query aqdw1 Medium QueryRouterEvaluator: Received results for already timed out query with query type ResultBlock 543aaf9c-0a9c-481a-9cfa-dda4fbcc4e5a
We use SP2019 with latest CU. We have this problem in two different environments on 2 different o365 tenants and data centers for SP on-prem. I have attached a trimmed ULS log. Where we can see the problem occur 2 times and by the 3 time the same search query is done, it works and SPO result is shown correctly.
Anyone experienced this problem?
Jun 05 2020 09:31 AM
@Martin_Ericsson Did you resolve this issue? I am trying to set up outbound hybrid federated search and I am getting this same timeout error in the ULS logs. In my case, I am trying to display the SP online search results in a result block and the majority of time they are not displayed at all and only the on-prem results are displayed. Were you able to track down the root cause?
Jun 05 2020 09:50 AM
Jun 05 2020 10:20 AM
@Lewis-H In my case, the result source and the query rule seem to be configured correctly as it intermittently works properly. Also, if I do the opposite where the primary results are the online results and the result block displays the on-prem results then it works fine.
There are a lot of components involved here and most of them are not easily accessible (on-prem search components, SP online search, etc.) to view exactly what is happening. I was interested in knowing what causes may have been found and what they used for troubleshooting.
Feb 10 2021 03:35 AM
@Rob_DF Sorry for the late answer. No we didn't resolve this. We choose a different path and stopped using the Hybrid Federated Search since it was so unstable and unreliable. Were you able to resolve it?