linux-nfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Chuck Lever <chuck.lever@oracle.com>
To: Kevin Constantine
	<Kevin.Constantine-FfNkGbSheRGpB8w63BLUukEOCMrvLtNR@public.gmane.org>
Cc: linux-nfs@vger.kernel.org, steved@redhat.com,
	Kevin Constantine
	<Kevin.Constantine-P5ys19MLBK/QT0dZR+AlfA@public.gmane.org>
Subject: Re: [PATCH 1/1] nfs-iostat.py: Fixes several Divide by Zero errors
Date: Thu, 27 May 2010 20:22:17 -0400	[thread overview]
Message-ID: <4BFF0CB9.7010700@oracle.com> (raw)
In-Reply-To: <1275004718-1802-1-git-send-email-kevin.constantine-FfNkGbSheRGpB8w63BLUukEOCMrvLtNR@public.gmane.org>

On 05/27/2010 07:58 PM, Kevin Constantine wrote:
> There was no check to see if sample_time was zero before dividing by it.

I haven't looked at this code in a very long time.  Why was sample_time 
zero?  That seems wrong.

> This was causing ZeroDivisionError's:
>
> Traceback (most recent call last):
>    File "/home/fahome/kconstan/repos/nfs-utils.upstream/tools/nfs-iostat/nfs-iostat.py", line 628, in ?
>      iostat_command(prog)
>    File "/home/fahome/kconstan/repos/nfs-utils.upstream/tools/nfs-iostat/nfs-iostat.py", line 610, in iostat_command
>      print_iostat_summary(old_mountstats, mountstats, devices, sample_time, options)
>    File "/home/fahome/kconstan/repos/nfs-utils.upstream/tools/nfs-iostat/nfs-iostat.py", line 462, in print_iostat_summary
>      stats[device].display_iostats(time, options.which)
>    File "/home/fahome/kconstan/repos/nfs-utils.upstream/tools/nfs-iostat/nfs-iostat.py", line 370, in display_iostats
>      backlog = (float(self.__rpc_data['backlogutil']) / sends) / sample_time
> ZeroDivisionError: float division
>
> Traceback (most recent call last):
>    File "/home/fahome/kconstan/repos/nfs-utils.upstream/tools/nfs-iostat/nfs-iostat.py", line 628, in ?
>      iostat_command(prog)
>    File "/home/fahome/kconstan/repos/nfs-utils.upstream/tools/nfs-iostat/nfs-iostat.py", line 610, in iostat_command
>      print_iostat_summary(old_mountstats, mountstats, devices, sample_time, options)
>    File "/home/fahome/kconstan/repos/nfs-utils.upstream/tools/nfs-iostat/nfs-iostat.py", line 462, in print_iostat_summary
>      stats[device].display_iostats(time, options.which)
>    File "/home/fahome/kconstan/repos/nfs-utils.upstream/tools/nfs-iostat/nfs-iostat.py", line 380, in display_iostats
>      print '%7.2f' % (sends / sample_time),
> ZeroDivisionError: float division
>
> Traceback (most recent call last):
>    File "/home/fahome/kconstan/repos/nfs-utils.upstream/tools/nfs-iostat/nfs-iostat.py", line 630, in ?
>      iostat_command(prog)
>    File "/home/fahome/kconstan/repos/nfs-utils.upstream/tools/nfs-iostat/nfs-iostat.py", line 612, in iostat_command
>      print_iostat_summary(old_mountstats, mountstats, devices, sample_time, options)
>    File "/home/fahome/kconstan/repos/nfs-utils.upstream/tools/nfs-iostat/nfs-iostat.py", line 464, in print_iostat_summary
>      stats[device].display_iostats(time, options.which)
>    File "/home/fahome/kconstan/repos/nfs-utils.upstream/tools/nfs-iostat/nfs-iostat.py", line 386, in display_iostats
>      self.__print_rpc_op_stats('READ', sample_time)
>    File "/home/fahome/kconstan/repos/nfs-utils.upstream/tools/nfs-iostat/nfs-iostat.py", line 350, in __print_rpc_op_stats
>      print '\t\t%7.3f' % (ops / sample_time),
> ZeroDivisionError: float division
>
> Signed-off-by: Kevin Constantine<kevin.constantine-FfNkGbSheRGpB8w63BLUukEOCMrvLtNR@public.gmane.org>
> ---
>   tools/nfs-iostat/nfs-iostat.py |   20 +++++++++++++++-----
>   1 files changed, 15 insertions(+), 5 deletions(-)
>
> diff --git a/tools/nfs-iostat/nfs-iostat.py b/tools/nfs-iostat/nfs-iostat.py
> index 2d0b143..4d6b295 100644
> --- a/tools/nfs-iostat/nfs-iostat.py
> +++ b/tools/nfs-iostat/nfs-iostat.py
> @@ -342,13 +342,19 @@ class DeviceData:
>               retrans_percent = 0.0
>               rtt_per_op = 0.0
>               exe_per_op = 0.0
> +        if sample_time != 0:
> +            ops_per_sample_time = ops / sample_time
> +            kb_per_sample_time = kilobytes / sample_time
> +        else:
> +            ops_per_sample_time = 0.0
> +            kb_per_sample_time = 0.0
>
>           op += ':'
>           print '%s' % op.lower().ljust(15),
>           print '  ops/s\t\t   kB/s\t\t  kB/op\t\tretrans\t\tavg RTT (ms)\tavg exe (ms)'
>
> -        print '\t\t%7.3f' % (ops / sample_time),
> -        print '\t%7.3f' % (kilobytes / sample_time),
> +        print '\t\t%7.3f' % (ops_per_sample_time),
> +        print '\t%7.3f' % (kb_per_sample_time),
>           print '\t%7.3f' % kb_per_op,
>           print ' %7d (%3.1f%%)' % (retrans, retrans_percent),
>           print '\t%7.3f' % rtt_per_op,
> @@ -358,7 +364,9 @@ class DeviceData:
>           sends = float(self.__rpc_data['rpcsends'])
>           if sample_time == 0:
>               sample_time = float(self.__nfs_data['age'])
> -        return (sends / sample_time)
> +            return (sends / sample_time)
> +        else:
> +            return(0.0)
>
>       def display_iostats(self, sample_time, which):
>           """Display NFS and RPC stats in an iostat-like way
> @@ -366,10 +374,12 @@ class DeviceData:
>           sends = float(self.__rpc_data['rpcsends'])
>           if sample_time == 0:
>               sample_time = float(self.__nfs_data['age'])
> -        if sends != 0:
> +        if sends != 0 and sample_time != 0:
>               backlog = (float(self.__rpc_data['backlogutil']) / sends) / sample_time
> +            sends_per_sample_time = sends / sample_time
>           else:
>               backlog = 0.0
> +            sends_per_sample_time = 0.0
>
>           print
>           print '%s mounted on %s:' % \
> @@ -377,7 +387,7 @@ class DeviceData:
>           print
>
>           print '   op/s\t\trpc bklog'
> -        print '%7.2f' % (sends / sample_time),
> +        print '%7.2f' % (sends_per_sample_time),
>           print '\t%7.2f' % backlog
>
>           if which == 0:


-- 
chuck[dot]lever[at]oracle[dot]com

  parent reply	other threads:[~2010-05-28  0:24 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-05-27 23:58 [PATCH 1/1] nfs-iostat.py: Fixes several Divide by Zero errors Kevin Constantine
     [not found] ` <1275004718-1802-1-git-send-email-kevin.constantine-FfNkGbSheRGpB8w63BLUukEOCMrvLtNR@public.gmane.org>
2010-05-28  0:22   ` Chuck Lever [this message]
2010-05-28  0:50     ` Kevin Constantine
2010-05-28 17:00       ` Chuck Lever
2010-05-28 22:51         ` Kevin Constantine
2010-06-01 15:37           ` Chuck Lever
2010-06-02 17:13             ` Kevin Constantine
2010-06-07 22:40               ` [PATCH 1/1] nfs-iostat.py: divide by zero with fresh mount Kevin Constantine
     [not found]                 ` <1275950427-10200-1-git-send-email-kevin.constantine-FfNkGbSheRGpB8w63BLUukEOCMrvLtNR@public.gmane.org>
2010-06-07 23:12                   ` Chuck Lever
2010-06-22 21:52                   ` Steve Dickson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4BFF0CB9.7010700@oracle.com \
    --to=chuck.lever@oracle.com \
    --cc=Kevin.Constantine-FfNkGbSheRGpB8w63BLUukEOCMrvLtNR@public.gmane.org \
    --cc=Kevin.Constantine-P5ys19MLBK/QT0dZR+AlfA@public.gmane.org \
    --cc=linux-nfs@vger.kernel.org \
    --cc=steved@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).