Skip to main content
Ex Libris Knowledge Center

Indexing process fails quickly after starting

  • Article Type: General
  • Product: Primo
  • Product Version: 4
  • Relevant for Installation Type: Dedicated-Direct; Local;

Problem Symptoms:
? Includes Indexing_and_Hotswapping and Indexing_and_Didumean_and_Hotswapping processes
? Indexing process fails consistently
? Failure occurs quickly after Indexing process starts

Cause:
Indexing will fail for 2-Tier or 2-Tier MFE topology when all SEs (Indexers) located on the BE server are not running

Resolution:
1. use ssh as primo user to access BE server.
2. use jps -m to determine which Primo processes are running
2a. 2-Tier BE should display:
<pid1> AgentServer -port:9501 (This is SE)
<pid2> Bootstrap start (This is Back Office)

2b. 2-Tier MFE BE should show
<pid1> AgentServer -port:9501 (This is slot 1 SE)
<pid1> AgentServer -port:9502 (This is slot 2 SE)
<pid2> Bootstrap start (This is Back Office)

3. If any AgentServer processes are missing, log into the server selecting the appropriate slot*, and run se_start to start the SE

4. When the SE has completed starting, re-Execute the appropriate Indexing process from the Back Office

Additional Information

*A 2-Tier topology BE (non-MFE) will log user into slot 1 by default without prompting the user.
The SE process on the FE server serves as the search Slice, whereas the SE process on the BE server serves as the Indexer, and is required to be running by the Indexing process.

For a 2-Tier MFE topology:
The SE (slice) on FE1 server communicates with the SE (indexer) on slot 1 (p4_1) of BE server on port 9501
The SE (slice) on FE2 server communicates with the SE (indexer) on slot 2 (p4_2) of BE server on port 9502
BOTH of the SEs on BE server need to be running

Category: Search Engine

Subject: Indexing


  • Article last edited: 7/7/2015