This message was deleted.
# troubleshooting
s
This message was deleted.
a
Do you have streaming ingestion with late data or other batch ingestion with higher priority for this datasource + interval?
d
The problem is with all data sources not with particular one
a
Could you please share more info such as the task logs, overlord logs and maybe even the coordinator's logs related to compaction?
d
TASK EROOR - " errorMsg": "org.apache.druid.segment.loading.SegmentLoadingException: Failed to load segment data_source_name_202..." LOG -"Request failed with status code 404" COORDINATOR LOGS 2023-03-08T133759,389 WARN [Coordinator-Exec--0] org.apache.druid.server.coordinator.duty.NewestSegmentFirstIterator - total segment size[1737266692] for datasource[data_source_name] and interval[2023-02-13T000000.000Z/2023-02-14T000000.000Z] is larger than inputSegmentSize[1073741824]. Continue to the next interval.
a
Could you please check the segment availability as well?
It seems like the segment data might be missing from deep storage
2023-03-08T133759,389 WARN [Coordinator-Exec--0] org.apache.druid.server.coordinator.duty.NewestSegmentFirstIterator - total segment size[1737266692] for datasource[data_source_name] and interval[2023-02-13T000000.000Z/2023-02-14T000000.000Z] is larger than inputSegmentSize[1073741824]. Continue to the next interval.
This seems problematic as well. Could you please set the inputSegmentSize to a higher value and see if it helps? (Which version of Druid are you using?)
d
Segments are available for only this date ->2023-02-23
inputSegmentSize I didn't increased and tried. As I am not familiar with the impact of doing so. Can you please tell me what it does if possible
a
Could you please share which version you are using?
I think inputSegmentSize has been deprecated
From what I can see this was relevant when compaction ran on a single thread to limit the resources. However, compaction can run with several subtasks now and this config is no longer needed
d
The version is 0.22.1
a
You could increase it to a large value and then increase maxNumConcurrentSubTasks as well
However, I think the compaction failure is due to the segment data not being available
d
Ok I will check