[GitHub] [accumulo] keith-turner commented on issue #2667: lastCompactID is inconsistent with metadata Exception

[GitBox <gi...@apache.org>]

keith-turner commented on issue #2667:
URL: https://github.com/apache/accumulo/issues/2667#issuecomment-1269905344

   > I wonder if this error occurred when a table was compacting and splitting at the same time and the split failed due to the bug I fixed in https://github.com/apache/accumulo/issues/2977.
   
   I will look into that.  I have been trying to recreate your scenario with a slight modification to the RW test graph. Instead of not ever verifiying and therefore not ending.  I changed my graph locally to start over after verify.  So doing the following locally.
   
   ```
   $ git diff
   diff --git a/src/main/resources/randomwalk/modules/Bulk.xml b/src/main/resources/randomwalk/modules/Bulk.xml
   index 26004b1..1880944 100644
   --- a/src/main/resources/randomwalk/modules/Bulk.xml
   +++ b/src/main/resources/randomwalk/modules/Bulk.xml
   @@ -55,7 +55,7 @@
    </node>
    
    <node id="bulk.Verify">
   -  <edge id="END" weight="1"/>
   +  <edge id="bulk.Setup" weight="1"/>
    </node>
    
    </module>
   ```
   
   I have been bumping into a few little issues with the test when trying to verify so i have not have not had a super long run yet.  But in the runs so far I don't think I saw any problems.  I will take a look at the split thing you mentioned.  I also found I needed to add nohup to your little script so that the processes would not terminate.
   
   ```
   for (( i=1; i<5; i++)); do nohup bin/rwalk Bulk.xml > rwalk"$i".log 2>&1 & done
   ```
   
   I think I am close to getting a test that could run for days on end going.  I think this a good to test to run even if this issue may be fixed by the change you mentioned.
   
   @milleruntime when you did see the error on the tserver, did it cause the RW test to hang or error?
   
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: notifications-unsubscribe@accumulo.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org