Auto restart for master(coordinator)

Hi, somehow the dremio master will shut down automatically (but all 4 workers are working fine)

seek for help:

  1. the dremio master node was in the same instance with other bi tool, any way could I avoid the dremio master shut down nexttime? was it related to os memory? was there a way to check the heap & direct memory usage (via UI)?

  2. whenever the master restarted, i refreshed the existing reflection with below error:
    IllegalStateException: Selected partition 28563bb4-c476-4d0c-91af-848d5bfb12af_1671505352475_0 does not have associated splits. Please report to administrator.

even i retry to refresh many times but still fail, my workaround is to restart also the 4 workers and then refresh on reflection works again!! any solution to avoid this case happen?

many thanks in advance! :smiley:

@crazyisjen This error is from an older version, what version of Dremio are you on?

Cn you please explain what you mean by “the dremio master node was in the same instance with other bi tool”?

my dremio ver. is 19.0.0 (community ver.)
dremio master node and superset bi tool locate in same cloud server

@crazyisjen

Can you check in /var/log/messages if you see a oom-killer?