ivakegg opened a new issue, #3552:
URL: https://github.com/apache/accumulo/issues/3552

   In this case, doing an accumulo-cluster stop, the shutdown process was stuck 
but for a different reason this time.  The manager in this case was repeating 
the message that 1308 non metadata tablets were still hosted.  I scanned the 
metadata table for the "loc" cf and found that they were all hosted by the same 
tserver.  In that tserver's log, it was getting spammed with "told to unload 
tablet that was not being served ..."  The solution was of course to kill that 
tserver manually and then the manager finally recovered from that situation and 
completed the shutdown.
   
   Tracking one of the tablets in the tserver's log, I did not find any 
indication that it had unloaded the tablet.  I see a compaction message, and 
then 40 minutes later the "told to unload tablet that was not being served" 
message.  I did not find any indication of the tablet being loaded on any other 
tserver.  So all I can figure is that something did not get flushed to the 
accumulo metadata table or the tserver that was hosting those tablets simply 
forgot about them.  This is a little scary.
   
   Accumulo 2.1.1


-- 
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: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to