Skip to content

Latest commit

 

History

History
83 lines (57 loc) · 3 KB

FAQ.md

File metadata and controls

83 lines (57 loc) · 3 KB
Raw

FAQ

DDC is resource intensive

do the following in the ddc.yaml:

  • set number-threads: 1
  • set dremio-jstack-freq-seconds: 10
  • set dremio-queries-json-num-days: 7
  • if using --dremio-pat-prompt when running ddc or setting dremio-pat-token in the ddc.yaml then set number-job-profiles: 50

DDC tarball is too big

do the following in the ddc.yaml:

  • set dremio-queries-json-num-days: 7
  • set collect-gc-logs: false

DDC is too slow

do the following in the ddc.yaml:

  • set number-threads: 4
  • if using --dremio-pat-prompt when running ddc or setting dremio-pat-token in the ddc.yaml then set number-job-profiles: 50

I have a tiny /tmp folder and DDC is filling it up

  • set --transfer-dir at the cli or if doing a local-collect use --tarball-out-dir or set tarball-out-dir in ddc.yaml this will avoid the use the /tmp folder (as of ddc 0.9.0)

DDC didn't capture what I wanted

  • read the ddc-HOSTNAME.log logs and see what errors there are (ie literally grep for ERROR)
  • are the dremio-log-dir, dremio-conf-dir set correctly? (assuming the node is offline or the version of DDC is under 0.8 this may be necessary to set)
  • the job profiles, KV report, WLM report, and system table report all need dremio-pat-token to be set in ddc.yaml or --dremio-pat-prompt to be passed at the command line
  • are you running the latest version of DDC? We had over 15 releases in 2023 containing bug fixes and new functionality, check here https://github.com/dremio/dremio-diagnostic-collector/releases
  • If you are running ssh..did you remember to use --sudo-user as the dremio user or as a user with admin rights?

What is captured by DDC?

DDC has 4 modes - but you can alter the ddc.yaml and override collection modes to suit your own configuration

Light mode

  • system disk usage
  • server.log and 2 days of archives
  • metadata_refresh.log and 2 days of archives
  • reflection.log and 2 days of archives
  • queries.json and up to 2 days of archives
  • all Dremio configurations
  • all GC logs if present

Standard mode

  • perf metrics (cpu and GC usage by thread)
  • system disk usage
  • Java Flight Recorder recording of 60 seconds
  • top output of 60 seconds - older versions used ttop
  • server.log and 7 days of archives
  • metadata_refresh.log and 7 days of archives
  • reflection.log and 7 days of archives
  • queries.json and up to 30 days of archives
  • all Dremio configurations
  • all GC logs if present

Standard+jstack mode

  • everything stated in standard
  • captures 60 seconds of jstack at 1 second intervals

Healthcheck mode (with a Dremio Personal Access Token)

  • everything stated in standard
  • a sampling of job profiles (note 25000 jobs can take 15 minutes to collect)
  • Dremio key value store report
  • Dremio work load manager details
  • system tables and their details

Optionally with the appropriate change to ddc.yaml

  • access.log and 7 days of archives
  • audit.log and 7 days of archives
  • Java heap dump