This message was deleted.
# troubleshooting
s
This message was deleted.
a
Could you please check the historical logs corresponding to the stuck load queue? You could try restarting the coordinator to see if it helps clear the bad segment from the load queue. (Also, please switch from
curator
to
http
loadqueues before restarting the coordinator in case you are using the former)
v
Copy code
Unannouncing segment[<segment>] at path[..]
org.apache.druid.server.coordination.SegmentLoadDropHandler - Completely removing <segment> in [30,000] millis
org.apache.druid.server.SegmentManager - Attempting to close segment <segment>
only these 3 log lines
a
Could you please share the replication factor of your segments and also run the following sys table query from the druid console?
select num_replicas, count(*) from sys.segments group by 1
v
Replication factor is 2
a
The above query should give the count of segments with various replication levels. It's possible that the cluster is overreplicated and segments are being dropped
v
Copy code
1, 503
2, 103229
a
Oh, overreplication is not the problem here then
only these 3 log lines
Are these the only 3 log lines in total? I interpreted it as these 3 log lines are repeated continuously on the historical earlier
v
I mean they are repeated per segment
Do you have any insight on
org.apache.druid.server.coordinator.ReplicationThrottler:
errors ? Why do they occur ?
Replicant create queue stuck after 3 runs
I increased it from 10 to 100 to see if it can reduce the # of errors.