Long pending time

What causes a query to wait in pending? What resources can be increased to lower this time?

For context this is a kubernetes deployment.

@millius, Pending means the query is waiting to be planned. Sixteen seconds seems somewhat high.

  • What version of Dremio are you using (4.6?)
  • Were there many other jobs submitted simultaneously?
  • Do you frequently see long Pending times (more then several seconds) or just this one instance

@ben

  • ver 4.6.1
  • Yes there are many jobs kicked off around the same time.
  • This happens frequently.

I suspect that bullet #2 may be the key. There are about 12 jobs submitted in a 16 second window. Each gets a progressively longer pending time.

What resource can be expanded to improve the pending time?

The number of cores available to the master coordinator.