Dremio with docker swarm and reflections status

1

We are using Dremio with docker swarm as the orchestration tool, and also using anNFS share. Every time we cycle the stack, all reflections gets into an INVALID status. I have two questions:

  1. Are they suppose to self heal on the next refresh cycle?
  2. How can we avoid this from happening?

Thanks

2

@martinocando

How many executors do you have? Is this NFS share mounted on all executors? Can you please provide us the dremio.conf file?

Thanks
Bali

3

We have 1 master and 4 executors. And yes, the NFS share is mounted on all executors.

This is dremio.conf from the master node:

#########################

paths: {
local: ${DREMIO_HOME}"/data"

dist: “pdfs://”${paths.local}"/pdfs"
}
services: {
coordinator.enabled: true,
coordinator.master.embedded-zookeeper.enabled: false,
executor.enabled: false
}
zookeeper: “zk_zoo1:2181,zk_zoo2:2181,zk_zoo3:2181”
registration.publish-host:“172.25.24.213”

And the dremio.conf of one of the executors:

###############################

paths: {
local: ${DREMIO_HOME}"/data"

dist: “pdfs://”${paths.local}"/pdfs"
}

services: {
coordinator.enabled: false
}

zookeeper: “zk_zoo1:2181,zk_zoo2:2181,zk_zoo3:2181”
registration.publish-host:"${LOCAL_CONTAINER_IP}"

4

@martinocando

The dremio.conf setting below is configured as a PDFS, which means the reflection files are distributed across the executors and not like a central store,

dist: “pdfs://”${paths.local}“/pdfs”

please fix it using the below syntax

dist: “file://”${paths.local}“/pdfs”

http://docs.dremio.com/deployment/dist-store-config.html#nas

5

Thanks for the information. I’ll test it and report back.