I posted the problem before, but didn't get any answer. Since we have to resolve the issue, I post again.
I loaded data to Vertica using multiple (parallel) COPY DIRECTs (via a old version of Vertica connector). It took about 40 mins to finish this. But after it finished, from our Ganglia charts, we found the CPU utilization was still high for about 3 hours. What was the CPU doing? Can we speed up the computing needed after COPY DIRECTs? Thanks.
Ey-Chih Chow
If you are running many COPY DIRECT statements, you may be creating work for the Tuple Mover. Specifically Mergeout. The documentation ought to be able to point you towards monitoring and tuning choices for Tuple Mover.
- Derrick
Thanks for the information. We have some reporting jobs running on the background. When the CPU utilization is high due to mergeout of the tuple mover, the reporting jobs are not progressing. Can we lower the priority of the tuple mover so the jobs for reporting and interactive analytics can be on schedule? Thanks.
Best regards,
Ey-Chih Chow
Ey-Chih Chow,
The "Tuning the Tuple Mover" documentation page may help. It's immediately after the Tuple Mover documentation page I linked in my last reply.
Derrick
Already have an account? Login
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.
77 4th Avenue
Waltham, MA 02451 USA
