You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@hbase.apache.org by Stanley Shi <ss...@pivotal.io> on 2014/10/11 03:45:42 UTC
log: ulimit output location
Hi users,
I found that currently the hbase daemon will send the ulimit output to .log
file;
To my understanding:
The .out file is used to store stdout and stderr (in case user
misconfigured log4j, all hbase logs will goto this file).
The .log file is meant to store only the log4j outputs from the hbase codes.
This .out and .log files together makes the "hbase log".
So I was wondering why we put the ulimit output to .log file, why not to
the .out file?
BTW: for hadoop hdfs/yarn daemons, the ulimit information is printed to the
.out file;
--
Regards,
*Stanley Shi,*
Re: log: ulimit output location
Posted by Stack <st...@duboce.net>.
On Sat, Oct 11, 2014 at 10:26 AM, lars hofhansl <la...@apache.org> wrote:
> The ullimit should be in log not out (IMHO). That's the file you look at
> for normal operations.
> Can we log it to both places?
>
We could do that.
St.Ack
> -- Lars
>
> From: Stack <st...@duboce.net>
> To: Hbase-User <us...@hbase.apache.org>
> Sent: Friday, October 10, 2014 10:36 PM
> Subject: Re: log: ulimit output location
>
> On Fri, Oct 10, 2014 at 6:45 PM, Stanley Shi <ss...@pivotal.io> wrote:
>
> > Hi users,
> >
> > I found that currently the hbase daemon will send the ulimit output to
> .log
> > file;
> > To my understanding:
> > The .out file is used to store stdout and stderr (in case user
> > misconfigured log4j, all hbase logs will goto this file).
> > The .log file is meant to store only the log4j outputs from the hbase
> > codes.
> > This .out and .log files together makes the "hbase log".
> >
> > So I was wondering why we put the ulimit output to .log file, why not to
> > the .out file?
> >
>
>
> We've been doing it this way since near inception writing all we can to the
> one log. IIRC, it was thought it just easier having one log carry all
> needed to debug an install.
>
> Anyone care if we change to align with how hdfs/yarn store environment
> info, to .out rather than .log?
>
> St.Ack
>
>
>
>
>
> > BTW: for hadoop hdfs/yarn daemons, the ulimit information is printed to
> the
> > .out file;
> >
> > --
> > Regards,
> > *Stanley Shi,*
> >
>
>
>
>
Re: log: ulimit output location
Posted by lars hofhansl <la...@apache.org>.
The ullimit should be in log not out (IMHO). That's the file you look at for normal operations.
Can we log it to both places?
-- Lars
From: Stack <st...@duboce.net>
To: Hbase-User <us...@hbase.apache.org>
Sent: Friday, October 10, 2014 10:36 PM
Subject: Re: log: ulimit output location
On Fri, Oct 10, 2014 at 6:45 PM, Stanley Shi <ss...@pivotal.io> wrote:
> Hi users,
>
> I found that currently the hbase daemon will send the ulimit output to .log
> file;
> To my understanding:
> The .out file is used to store stdout and stderr (in case user
> misconfigured log4j, all hbase logs will goto this file).
> The .log file is meant to store only the log4j outputs from the hbase
> codes.
> This .out and .log files together makes the "hbase log".
>
> So I was wondering why we put the ulimit output to .log file, why not to
> the .out file?
>
We've been doing it this way since near inception writing all we can to the
one log. IIRC, it was thought it just easier having one log carry all
needed to debug an install.
Anyone care if we change to align with how hdfs/yarn store environment
info, to .out rather than .log?
St.Ack
> BTW: for hadoop hdfs/yarn daemons, the ulimit information is printed to the
> .out file;
>
> --
> Regards,
> *Stanley Shi,*
>
Re: log: ulimit output location
Posted by Stack <st...@duboce.net>.
On Fri, Oct 10, 2014 at 6:45 PM, Stanley Shi <ss...@pivotal.io> wrote:
> Hi users,
>
> I found that currently the hbase daemon will send the ulimit output to .log
> file;
> To my understanding:
> The .out file is used to store stdout and stderr (in case user
> misconfigured log4j, all hbase logs will goto this file).
> The .log file is meant to store only the log4j outputs from the hbase
> codes.
> This .out and .log files together makes the "hbase log".
>
> So I was wondering why we put the ulimit output to .log file, why not to
> the .out file?
>
We've been doing it this way since near inception writing all we can to the
one log. IIRC, it was thought it just easier having one log carry all
needed to debug an install.
Anyone care if we change to align with how hdfs/yarn store environment
info, to .out rather than .log?
St.Ack
> BTW: for hadoop hdfs/yarn daemons, the ulimit information is printed to the
> .out file;
>
> --
> Regards,
> *Stanley Shi,*
>