"SRCH-02001: Error while retrieving search contents. Contact your system administrator." while searching or retrieving or adding documents to the knowledge store

Document ID : KB000019522
Last Modified Date : 17/04/2018
Show Technical Document Details
Introduction:

Description:

Searching/ Retrieving or uploading documents to the knowledge store or project - collaboration tabs errors out:

SRCH-02001: Error while retrieving search contents. Contact your system administrator.

This could happen if there is corruption in indexes and this can due to a variety of reasons and can affect different specific areas.

For example, an issue with lucene locking can occur. The lucene references a document on your system. These references should be automatically removed once a document has been opened, saved, and stored. Sometimes, these locks remain in the system and might corrupt the normal functionality of the indexing.

Instructions:

Solution:

1. Pause the job "Index contents and documents for search" only if it is in the 'WAITING' or 'SCHEDULED' state.

2. Turn off synchronous indexing.
a. Go to the <PPM_HOME>\.setup\scripts folder
b. Open the search.xml file to edit.
c. Find the following section that starts with  <jvmarg
d. Add the following line:
<jvmarg value="-Djava.io.tmpdir=<JAVA_TEMP>"/>

Excerpt example:

    <presetdef name="search.java">
      <java fork="true" classpathref="lib.path">
        <jvmarg line="-Xms64m -Xmx${scriptMaxMem} -XX:-UseGCOverheadLimit"/>
        <jvmarg line="-Dfile.encoding=UTF-8 -Djava.net.preferIPv4Stack=true"/>
        <jvmarg value="-Dinstall.dir=${install.dir}"/>
        <jvmarg value="-Djava.io.tmpdir=c:\javatemp"/>
        <arg file="${install.dir}"/>
      </java>

e. Save the file.

3. Log in to the CSA
a. Click on the Application page.
b. In the "Application instance: app" section,
append the following field to the JAVA VM Parameters field:
-Djava.io.tmpdir=<JAVA_TEMP>"

Be sure the actual location exists.

c. Click Save.

d. Click on the Background page.
e. Go to the "Background Server" section
f. append the following field to the JAVA VM Parameters field:
-Djava.io.tmpdir=<JAVA_TEMP>

Be sure the actual location exists.

4. Search and delete any files that end in write.lock on the application server.
They will look something like this:
lucene-62db5567fbeaae16396e79fb8fe42895-write.lock

5. Restart all APP and BG services.

6. Delete the nspider and nspiderfile folders from the file server.

7. Log into the app server as someone who has full control of the filestore
folders including the ability to create new folders/files like the nspiderfiles folders

a. Run the command: admin search recreate-index-data
b. Run the command: admin search recreate-index-files

8. When the command completes,
run the job "Index contents and documents for search"

IMPORTANT NOTE ABOUT RE-INDEXING:

If you are getting frequent search errors that cause you to re-index too frequently, consider running only the synchronous indexing or only the "Index contents and documents" job. There have been some instances where the running both at the same time have caused more errors to occur. Both methods accomplish essentially the same thing. All documents will be indexed using either one. The synchronous indexing indexes documents as they are added to the system which allows users to search for their document or its contents almost immediately after the document is added to the CA PPM system. The disadvantage to this method is that it can provide extra system overhead in systems where lots of documents are being added and/or changed. The "Index Contents and Documents" job indexes documents only when the job is run. Most customers schedule this job to run anywhere from once every 15 minutes to once every hour. The job also does more optimization of the index than the synchronous indexing does. The disadvantage to this method is that users cannot search successfully for recently added documents until the next the job is run. If you choose to use the synchronous indexing option, I recommend running steps 11 and 12 to recreate your indexes once every month or once every 3 months depending on how heavily used your system is. This will optimize the index and improve the performance of your searches. If you are using the "Index Contents and Documents" job, you do not need to run these commands quite so frequently. Once or twice a year is probably sufficient.

IMPORTANT SCHEDULING NOTE: 

Please perform the activity after hours, as this very resource-intensive and could impact performance.

Additional Information:
https://docops.ca.com/ca-ppm/15-4/en/administration/csa-ca-ppm-system-administration/csa-document-index-search-storage-and-file-scanning