Text Index Event Process (Never end)

danielcuadros
Tera Contributor

Hello Everyone!

I'm having some problems with the text index. I activated the text index on a table that I have created, but the Text Index Event Process never ends. It has been there for one day and still in state Ready.

What can I do? It's normal to expend so much time in there?

I wanted to activate de text index of that table to allow find that records on the search box.

Thank you for your time.

3 REPLIES 3

sergiu_panaite
ServiceNow Employee
ServiceNow Employee

State 'Ready' is the normal state. When it runs, it should show state 'Running'. When you activate text indexing on a table a text index event is created in the Events Log and that event needs to be processed by 'text index events process' job. Was the case for you that the event was not processed?


raves12
Kilo Sage

HI Daniel,

 

We are running into same type of issue with our subprod instance flooded with text_index events which are in ready state. I sorted by created date and coverted the state of oldest one to error but no success.

Currently there are over 83000 text_index events which are in ready state and don't know from whwre should I begin.

mike_s
Kilo Contributor

Same case here. Takes about a day to process through the items queued for text indexing. Here is a URL for pulling up text index events that may be running/queued:
https://(your URL here)/sysevent_list.do?sysparm_query=stateNOT%20INready%2Cprocessed%2Cerror%2Ctransferred%5Equeue%3Dtext_index

In our case, we have a process that causes over 1000 (sometimes 10k+) records to have work notes added. We see the work notes being included in this text indexing. We opened a HI ticket and support mentioned about maybe not indexing on work notes for the record type. We are in agreement there, just seeing if that will be possible without impacting other records using that field (work notes comes from the task table).

 

May want to consider the same steps to resolve. See if there are fields that you do not truly need to be indexed. Otherwise, we are also considering less bulk updates at once and possibly performing fewer updates more frequently throughout the day to spread out the processing load.