Hi All,
I recently encountered the following exception when attempting to
write to our hadoop grid
This is from the name-node log:
2008-05-05 16:11:01,966 INFO org.apache.hadoop.ipc.Server: IPC Server
handler 4 on 54310, call create(/2008-05-05/ONE-2008-05-05-14.gz,
DFSClient_-164311132, false, 3, 67108864) from 10.2.15.6:42519: error:
org.apache.hadoop.dfs.AlreadyBeingCreatedException: failed to create
file /2008-05-05/ONE-2008-05-05-14.gz for DFSClient_-164311132 on
client 10.2.15.6 because current leaseholder is trying to recreate
file.
org.apache.hadoop.dfs.AlreadyBeingCreatedException: failed to create
file /2008-05-05/ONE-2008-05-05-14.gz for DFSClient_-164311132 on
client 10.2.15.6 because current leaseholder is trying to recreate
file.
at
org.apache.hadoop.dfs.FSNamesystem.startFileInternal(FSNamesystem.java:850)
at org.apache.hadoop.dfs.FSNamesystem.startFile(FSNamesystem.java:806)
at org.apache.hadoop.dfs.NameNode.create(NameNode.java:276)
at sun.reflect.GeneratedMethodAccessor15.invoke(Unknown Source)
at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
at java.lang.reflect.Method.invoke(Method.java:597)
at org.apache.hadoop.ipc.RPC$Server.call(RPC.java:379)
at org.apache.hadoop.ipc.Server$Handler.run(Server.java:596)
I was wondering if anyone could offer some insight as to possible
causes for this. Only one process was attempting to use copyFromLocal
on the file at that time (although five other machines were attempting
to copy files differently named to the same directory using
copyFromLocal). Our grid was under very heavy load at that time - so
might this have something to do with the following exception? It
should be noted that the following attempt to use copyFromLocal an
hour after this incident was successful.
Further, is there a way to specify that failed copies not leave
artifacts? In this particular case, a size 0 file
ONE-2008-05-05-14.gz was left on the grid.
--
Bo Shi
(207) 469-8264 (M)
25 Kingston St, 5th Fl
Boston, MA 02111 USA
