EdColeman commented on issue #2085: URL: https://github.com/apache/accumulo/issues/2085#issuecomment-839753763
When you do a restart, do you start the tservers and THEN the master(s) - if you start the master first, it will see assign everything to the first tserver(s) that it sees, resulting in an unbalanced cluster that takes a while for everything to be processed and then rebalanced because only one (or a few) tservers are processing everything. There is a property to have the master wait for a required number of tservers that was added in 1.10. It is best if the tservers are all up and then the master starts - taht way all tservers are available for assignments. Other factors could be if the tables are performing batch writes (vs bulk import) - that will generate WAL files that need to be processed during recovery. Do you every compact the tables - that will reduce the number of files as well as process any info in the WAL so that there is less work that needs to be performed during recovery. -- 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. For queries about this service, please contact Infrastructure at: [email protected]
