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

Switch to Threaded View
Flume, mail # user - Exception : Log File is null for id


Copy link to this message
-
Re: Exception : Log File is null for id
Hari Shreedharan 2012-10-05, 18:28
Brock,  

This looks like FLUME-1417. This logs on the jira show when the problem is hit during startup. I actually managed to get the Log Id is null error during runtime when I was testing that issue, if you change to small file size and checkpoint very often.

Thanks,
Hari

--
Hari Shreedharan
On Friday, October 5, 2012 at 11:19 AM, Brock Noland wrote:

> Hi,
>
> Just curious if you got around this or figured out what was going on?
> Makes me a little nervous about a file channel bug.
>
> Brock
>
> On Tue, Oct 2, 2012 at 6:28 AM, Brock Noland <[EMAIL PROTECTED] (mailto:[EMAIL PROTECTED])> wrote:
> > Also, if you could send us your full log that would be great. The
> > email list doesn't take attachements so either:
> >
> > 1) post it on pastbin
> > or
> > 2) zip it and mail it to me directly
> >
> > Brock
> >
> > On Tue, Oct 2, 2012 at 6:06 AM, Brock Noland <[EMAIL PROTECTED] (mailto:[EMAIL PROTECTED])> wrote:
> > > Hi,
> > >
> > > What version of flume? If trunk (1.3.0-SNAPSHOT) what is the last
> > > patch you have?
> > >
> > > Can you how us a ls -la of your data and checkpoint directories?
> > >
> > > Brock
> > >
> > > On Tue, Oct 2, 2012 at 3:43 AM, Raymond Ng <[EMAIL PROTECTED] (mailto:[EMAIL PROTECTED])> wrote:
> > > > Just to add more info to this, I've checked the File channel where a
> > > > "ChannelException: Cannot acquire capacity" is reported against, and can see
> > > > file log-1 has the size of 0 and log-2 has over 300 MB of data, comparing
> > > > with another File channel which has files log-2 and log-3 both with data in
> > > > it but no file log-1 is found.
> > > >
> > > > sounds like log-1 is the one causing the "NullPointerException: LogFile is
> > > > null for id 1" below, and when I restarted flume, I get the following
> > > > warning. I can confirm there was no manual tampering in the file channel
> > > > directory
> > > >
> > > > 2012-10-02 09:38:10,231 INFO [conf-file-poller-0]
> > > > DefaultLogicalNodeManager.java - Starting Channel probeFileChannel1
> > > > 2012-10-02 09:38:10,239 INFO [conf-file-poller-0]
> > > > DefaultLogicalNodeManager.java - Starting Channel probeFileChannel3
> > > > 2012-10-02 09:38:10,313 WARN [lifecycleSupervisor-1-2] ReplayHandler.java -
> > > > Hit EOF on /home/user/flume-ng/filechannel3/data/log-1
> > > > 2012-10-02 09:38:10,314 INFO [lifecycleSupervisor-1-1]
> > > > DirectMemoryUtils.java - Unable to get maxDirectMemory from VM:
> > > > NoSuchMethodException: sun.misc.VM.maxDirectMemory(null)
> > > > 2012-10-02 09:38:10,317 INFO [lifecycleSupervisor-1-1]
> > > > DirectMemoryUtils.java - Direct Memory Allocation: Allocation = 1048576,
> > > > Allocated = 0, MaxDirectMemorySize = 954466304, Remaining = 954466304
> > > > 2012-10-02 09:38:10,381 WARN [lifecycleSupervisor-1-1] LogFile.java -
> > > > Checkpoint for file(/home/user/flume-ng/filechannel1/data/log-2) is:
> > > > 1349166469095, which is beyond the requested checkpoint time: 0.
> > > > 2012-10-02 09:38:10,381 WARN [lifecycleSupervisor-1-2] LogFile.java -
> > > > Checkpoint for file(/home/user/flume-ng/filechannel3/data/log-2) is:
> > > > 1349166991594, which is beyond the requested checkpoint time: 0.
> > > > 2012-10-02 09:41:52,144 ERROR [lifecycleSupervisor-1-2] ReplayHandler.java -
> > > > Pending takes 32103 exist after the end of replay. Duplicate messages will
> > > > exist in destination.
> > > > 2012-10-02 09:41:52,709 INFO [lifecycleSupervisor-1-2]
> > > > MonitoredCounterGroup.java - Component type: CHANNEL, name:
> > > > probeFileChannel3 started
> > > > 2012-10-02 09:42:31,413 WARN [lifecycleSupervisor-1-1] LogFile.java -
> > > > Checkpoint for file(/home/cluster_admin/flume-ng/filechannel1/data/log-3)
> > > > is: 1349166981020, which is beyond the requested checkpoint time: 0.
> > > > 2012-10-02 09:45:14,836 ERROR [lifecycleSupervisor-1-1] ReplayHandler.java -
> > > > Pending takes 8409 exist after the end of replay. Duplicate messages will
> > > > exist in destination.
> > > > 2012-10-02 09:45:15,453 INFO [lifecycleSupervisor-1-1]