Re: Specifying the log file name of pgbench -l option - Mailing list pgsql-hackers

From Fabien COELHO
Subject Re: Specifying the log file name of pgbench -l option
Date
Msg-id alpine.DEB.2.20.1611011731100.9978@lancre
Whole thread Raw
In response to Specifying the log file name of pgbench -l option  (Masahiko Sawada <sawada.mshk@gmail.com>)
Responses Re: Specifying the log file name of pgbench -l option
List pgsql-hackers
> The log file generated by pgbench -l option is fixed file name
> 'pgbench_log.<pid>.<thread id>'. And it's a little complicated for the
> script that runs pgbench repeatedly to identify the log file name.
> Attached patch make it possible to specify the log file name. I think
> it's useful for the use who want to run pgbench repeatedly in script
> and collects and analyze the result.
>
> The one thing I concern is that this patch changes -l option so that
> it requires argument.
> But changing its behavior would be good rather than adding new option.
>
> Please give me feedback.

Patch applies but does not compile, because "logfilename" is not declared.
I guess "logfile" was meant instead.

I understand and agree that in some case having only a predefined file 
prefix in the current directory as the only option can be a hindrance for 
scripts which use pgbench and rely on the log.

I'm not at ease either with changing the behavior of such an option, as 
some people may be happy with it and some script may be using it. I would 
suggest not to do so.

Moreover, what is provided is not a file name, but a prefix used to build 
file names.

So I would suggest to: - fix the compilation issue - leave -l/--log as it is, i.e. use "pgbench_log" as a prefix - add
--log-prefix=...(long option only) for changing this prefix
 

-- 
Fabien.



pgsql-hackers by date:

Previous
From: Abbas Butt
Date:
Subject: Using a latch between a background worker process and a thread
Next
From: Adam Brusselback
Date:
Subject: Re: GiST support for UUIDs