Collect Vertica Scrutinize data for problem analysis

Products

• CA Infrastructure Management
• Network Observability
• CA Performance Management

Issue/Introduction

Vertica support has requested something called Scrutinize output for analysis while investigating an issue reported against the Data Repository Database servers. How is this data gathered to honor the request?

How to generate Vertica scrutinize logging diagnostics package for support analysis.

Environment

Data Repository, all versions

Resolution

To generate the requested Scrutinize diagnostic data in order to help us investigate the issue involved, complete the following steps.

1. Run the utility as the dradmin user.
2. Execute the following command:
   
   /opt/vertica/bin/scrutinize -U '<username>' -P '<password>' -m '<CaseNumber>'
   • Where '<username>' is the DB admin username, default is usually dradmin. Don't forget to include the single quotes!
   • Where '<password>' is the DB password, the same one used to stop/start the DB via the adminTools UI. Don't forget to include the single quotes!
   • Where '<CaseNumber>' is the Support Case Number involved with the request. Don't forget to include the single quotes!
     Note that the username, password and CaseNumber values must be in single quotes for the scrutinize process to gather all of the data necessary to debug many issues.
     The scrutinize process will still run and collect system data if this isn't done, but it will be missing this vital information if the command is not formatted correctly.

For example: If the DB user name is dradmin, the password (the same one used to stop/start the Vertica DB from the adminTools UI) is dbpass (default from the guides), the command run might look like this:

/opt/vertica/bin/scrutinize -U 'dradmin' -P 'dbpass' -m '00123456'

You can also include the -z switch, to include a certain number of old vertica.logs. The following example includes 5 previous Vertica logs.

/opt/vertica/bin/scrutinize -z5 -U 'dradmin' -P 'dbpass' -m 'Case 00123456'

When the process is completed, please upload the resulting VerticaScrutinize.<date>.zip file to the support case for analysis.

Additional Information

• This utility collects information from all nodes in your cluster at once.
• It only needs to run on one node in the cluster, any node will do.
• Creates a file named VerticaScrutinize.<date>.zip where date is replaced with a numeric date representation.
• This date naming allows multiple calls to the scrutinize utility, resulting in different files created for each execution for easy differentiation between runs.
• Use the --tmpdir=TMP_DIR to use a directory other than the default /tmp.
  • Sample to use /opt/files instead of /tmp:
    • --tmpdir=/opt/files
• Need additional days of logging?
  • Scrutinize gathers current day +1 by default, meaning two days of logging. Current day plus the day prior. 
  • If there is an issue under review that occurred 4 days ago we'd add this to the command.
  • It will gather current day +5 days prior covering the issue and the day before for logging clues.
    • -z 5
  • Change the number of days as needed.