Re: Re: aggregated logs on HDFS garbled

[Robert Kanter]

As a shameless self-promotion :), you can try this program that I wrote
which will convert the log files back to plain text files:
https://github.com/rkanter/ALFExploder


- Robert

On Thu, May 10, 2018 at 12:44 AM, <yao...@mail.ustc.edu.cn> wrote:

> Thanks for your advice！
>
> I tried the tflie reader you provided in [2]，but there's still some garble
> appear at the beginnig and the end, looks like this：
> stdout0
>
> and I found another tool that works fine:
> https://github.com/shanyu/hadooplogparser
>
> 在2018-05-10 11:13:38,Hariharan Iyer<hi...@qubole.com>写道：
> > Application logs are not saved in text format. They are saved in a
> > binary format called TFile. You can use the "yarn logs"[1] command to
> > view them. Alternatively you can use a tool like this tfile reader[2].
> >
> > [1] - https://hadoop.apache.org/docs/r2.7.4/hadoop-yarn/
> hadoop-yarn-site/YarnCommands.html#logs
> > [2] - https://github.com/alexandruanghel/hadoop-
> snippets/tree/master/read-tfile
> >
> > Thanks,
> > Hari
> >
> >
> > On Thu, May 10, 2018 at 6:06 AM,  <yao...@mail.ustc.edu.cn> wrote:
> > > Hello，
> > >
> > >
> > >
> > > I set “yarn.log-aggregation-enable” to true, and got logs on HDFS, but
> when
> > > I use
> > >
> > > “bin/hdfs dfs -cat” or download files from the HDFS website, some
> garble
> > > appeared, and it looks like this:
> > >
> > >
> > >
> > > \D1 \D3h\91\B5׶9\DFA@\92\BA\E1P
> > > \00 VERSION \00\00\00  \00 APPLICATION_ACL$\00 VIEW_APP\00 root \00
> > >
> > > MODIFY_APP\00 root
> > >  \00 APPLICATION_OWNER \00 root(\00&container_1525851594379_0001_01_000003
> \00 stderr\00 0\00 stdout\00 0\00 \00\00 \00
> > >
> > > \00 VERSION*(\00&container_1525851594379_0001_01_000003  none  ̡̡
> data:BCFile.index none\CC\ED   data:TFile.index none̷66 data:TFile.meta
> none̱  \00\00\00\00\00\00\00\F8\00 \00\00\D1 \D3h\91\B5׶9\DFA@\92\BA\E1P
> > >
> > >
> > > What can I do to fix this?
> > >
> > >
> > >
> > > Thanks in advance!
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: user-unsubscr...@hadoop.apache.org
> > For additional commands, e-mail: user-h...@hadoop.apache.org
> >
>
>