12-07-2018 02:00 AM - last edited on 12-07-2018 07:07 AM by cjervis
I am trying to run a big query and it is halted in running state in hue.I have identical cluster where same query is running fine.
both the servers are identical but when i went through the query plan the only difference i saw was max per-host reservation parameter.
it is below where query is running
Max Per-Host Resource Reservation: Memory=1.30 GB
Per-Host Resource Estimates: Memory=75.94 GB
and it is below where it has stuck
Max Per-Host Resource Reservation: Memory=143.56 MB
Per-Host Resource Estimates: Memory=3.84 GB
It is killing our time please help
12-07-2018 05:56 AM
12-08-2018 10:23 AM - edited 12-08-2018 10:27 AM
There's no relation between resource reservations and query states.
There's probably two things going on:
As a mitigation we usually suggest setting an "idle query timeout" in Cloudera manager to automatically cancel queries that have been hanging around for a while with no client activity.
Edit: second observation was wrong. See my next post.
12-08-2018 10:25 AM