Yes, the root cause seems to be the gap of 4 minutes between
2010-07-16 05:49:26,805 and 2010-07-16 05:53:23,476 . but I checked GC gc-hbase.log and don't see
4 minute gap in gc. I just wonder what could cause this large gap.

I also wonder if there is a configuration that I can do to avoid this long pause, or get around the problem
cause by this long pause.

Jimmy

--------------------------------------------------
From: "Stack" <[email protected]>
Sent: Friday, July 16, 2010 11:44 AM
To: <[email protected]>
Subject: Re: YouAreDeadException with hbase

You'll see this if the server reports to the master after the master
has ruled it 'dead'.

Here is the code that produces the exception:

   if (!isDead(serverName)) return;
String message = "Server " + what + " rejected; currently processing " +
     serverName + " as dead server";
   LOG.debug(message);
   throw new YouAreDeadException(message);

Servers are on the 'dead' list if zk reports their session has
expired.  The master moves then to cleanup after the dead server and
process its logs.  If during this cleanup time the server reports in,
master will return the youaredead exception.

Usually the RS has lost its zk session but has yet to realize it.

St.Ack

On Thu, Jul 15, 2010 at 11:52 PM, Jinsong Hu <[email protected]> wrote:
Hi, There:
I got some YouAreDeadException with hbase. what can cause it ? I do notice
between 5:49 to 5:53 ,
for 4 minutes, there is no log. This doesn't look like GC issue as I checked
the GC log, the longest GC
is only 9.6 seconds.

Jimmy.


2010-07-16 05:49:26,805 DEBUG
org.apache.hadoop.hbase.io.hfile.LruBlockCache: Ca
che Stats: Sizes: Total=3.355194MB (3518176), Free=405.4198MB (425113472),
Max=4
08.775MB (428631648), Counts: Blocks=1, Access=2178914, Hit=1034,
Miss=2177880,
Evictions=0, Evicted=0, Ratios: Hit Ratio=0.04745483165606856%, Miss
Ratio=99.95
254278182983%, Evicted/Run=NaN
2010-07-16 05:53:23,476 DEBUG
org.apache.hadoop.hbase.io.hfile.LruBlockCache: Ca
che Stats: Sizes: Total=3.355194MB (3518176), Free=405.4198MB (425113472),
Max=4
08.775MB (428631648), Counts: Blocks=1, Access=2178915, Hit=1035,
Miss=2177880,
Evictions=0, Evicted=0, Ratios: Hit Ratio=0.04750070511363447%, Miss
Ratio=99.95
250105857849%, Evicted/Run=NaN

....
2010-07-16 05:53:26,171 INFO org.apache.zookeeper.ClientCnxn: Client session
tim
ed out, have not heard from server in 240540ms for sessionid
0x329c88039b0006c,
closing socket connection and attempting reconnect
2010-07-16 05:53:27,333 INFO org.apache.zookeeper.ClientCnxn: Opening socket
con
nection to server t-zookeeper2.cloud.ppops.net/10.110.24.57:2181
2010-07-16 05:53:27,334 INFO org.apache.zookeeper.ClientCnxn: Socket
connection
established to t-zookeeper2.cloud.ppops.net/10.110.24.57:2181, initiating
sessio
n
2010-07-16 05:53:27,335 INFO org.apache.zookeeper.ClientCnxn: Unable to
reconnec
t to ZooKeeper service, session 0x329c88039b0006c has expired, closing
socket co
nnection
2010-07-16 05:53:27,896 INFO org.apache.zookeeper.ClientCnxn: Client session
tim
ed out, have not heard from server in 240520ms for sessionid
0x129c87a7f98007a,
closing socket connection and attempting reconnect


2010-07-16 05:53:39,090 FATAL
org.apache.hadoop.hbase.regionserver.HRegionServer
: Aborting region server serverName=m0002028.ppops.net,60020,1279237223465,
load
=(requests=952, regions=21, usedHeap=575, maxHeap=2043): Unhandled exception
org.apache.hadoop.hbase.YouAreDeadException:
org.apache.hadoop.hbase.YouAreDeadE
xception: Server REPORT rejected; currently processing
m0002028.ppops.net,60020,
1279237223465 as dead server
      at
org.apache.hadoop.hbase.master.ServerManager.checkIsDead(ServerManage
r.java:217)
      at
org.apache.hadoop.hbase.master.ServerManager.regionServerReport(Serve
rManager.java:271)
      at
org.apache.hadoop.hbase.master.HMaster.regionServerReport(HMaster.jav
a:684)
      at sun.reflect.GeneratedMethodAccessor2.invoke(Unknown Source)
      at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAcces
sorImpl.java:25)
      at java.lang.reflect.Method.invoke(Method.java:597)
at org.apache.hadoop.hbase.ipc.HBaseRPC$Server.call(HBaseRPC.java:576)
      at
org.apache.hadoop.hbase.ipc.HBaseServer$Handler.run(HBaseServer.java:
919)

      at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native
Method)


Reply via email to