What is going on when this error happens? Does this happen all the time, or did it just start? Have you looked in the coordinator or zookeeper logs? Any network blips happening around that time?
What are your load rules set to? And zookeeper is healthy?
Hi - did you resolve this?
If not, suggest checking “/testapp/druid/var/segment-cache” to ensure that each Historical process is able to write to that location successfully? There may be more information in the Historical log itself?