Race on segment loading

[sixtus]

Hi,

I see a lot of "Replicant create queue stuck after 15+ runs!" messages in coordinator.

From the log files it looks like this:

• Indexer builds a segment
• Coordinator schedules to load it
• Auto-merge detects that it can be merged and the merge finishes before the historical node actually downloaded the segment
• Now historical doesn't download the segment anymore (it's not the latest), but it doesn't get flushed from the load queue.

The only way to break it is to restart the historical node. It complains about missing segment on disk and unable to unload, then actually does the right thing.

Unfortunately, it looks like this seems to stall loading of other segments as well (even though I have multiple threads assigned to loading).

Does that make any sense? Some of the above is guesswork.

[agahemre]

I am getting the same error message. This message is related with the same data source segment which is no longer available. Are there any guess what root cause can be ?

[stale]

This issue has been marked as stale due to 280 days of inactivity. It will be closed in 4 weeks if no further activity occurs. If this issue is still relevant, please simply write any comment. Even if closed, you can still revive the issue at any time or discuss it on the dev@druid.apache.org list. Thank you for your contributions.

[stale]

This issue has been closed due to lack of activity. If you think that is incorrect, or the issue requires additional review, you can revive the issue at any time.