Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Accumulo, mail # user - Found two locations for the same extent


Copy link to this message
-
Re: Found two locations for the same extent
dlmarion 2014-01-25, 16:13
I dont recall ever seeing/hearing of this situation. Maybe someone else on the list has. Have you had any recent failures?
Sent via the Samsung GALAXY S®4, an AT&T 4G LTE smartphone

-------- Original message --------
From: Anthony F <[EMAIL PROTECTED]>
Date:01/25/2014  11:03 AM  (GMT-05:00)
To: [EMAIL PROTECTED]
Subject: Re: Found two locations for the same extent

Yes, I tried killing one and waiting for the system to stabilize - nothing changed.  Tried killing both, same deal.  I even tried bouncing the whole system.  I found a previous message about how to grant write permissions on !METADATA to root and was able to delete one of the offending entries.  At that point, the system stabilized.  Any tips on avoiding getting into this situation?
On Sat, Jan 25, 2014 at 10:51 AM, dlmarion <[EMAIL PROTECTED]> wrote:
Have you tried killing one or both of the tservers?
Sent via the Samsung GALAXY S®4, an AT&T 4G LTE smartphone
-------- Original message --------
From: Anthony F
Date:01/25/2014 10:12 AM (GMT-05:00)
To: [EMAIL PROTECTED]
Subject: Found two locations for the same extent

I have a table that appears to be in an inconsistent state.  The master keeps logging the following exception:

java.lang.RuntimeException: org.apache.accumulo.server.master.state.TabletLocationState$BadLocationStateException: found two locations for the same extent

and never balances/migrates or loads any tablets.  I've tried deleting one of the offending rows in the !METADATA table but, not surprisingly, root doesn't have proper permissions.

Anyone know how Accumulo 1.5.0 can get in this state and what can be done to get out of it?