addon iostat for RHEL 5

[stephane]

Hello All

Here is the addon with correct % busy.

Hope this help
Stephane

[Dimitri]

Hi Stephane!

many thanks for contribution!! :-))
on the same time I'm surprised that your "iostat" command is having
less output columns comparing to those used in the STAT-service -
normally the "newest" binary prints more or at least not less
information than the older one.. - Are you sure to use the latest
"iostat" version available for RH5 ?.. (RH should have something like
a sys stats package or similar)..

Rgds,
-Dimitri

> --
> You received this message because you are subscribed to the Google Groups
> "dim_STAT" group.
> To post to this group, send email to dim...@googlegroups.com.
> To unsubscribe from this group, send email to
> dimstat+u...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/dimstat?hl=en.
>
>

[stephane]

Hi Dimitry

Sysstat rpm seems to be up to date on my server.
sysstat-7.0.2-3.el5

There are the output of the iostat command:
[root@XXX ~]# iostat -x
Linux 2.6.18-194.el5 (XXXX)    06/28/2011

avg-cpu:  %user   %nice %system %iowait  %steal   %idle
          11.79    0.00    3.59    9.04    0.00   75.58

Device:         rrqm/s   wrqm/s   r/s   w/s   rsec/s   wsec/s avgrq-sz avgqu-sz   await  svctm  %util
sda              40.74   431.01 144.68 119.46 10143.01  4486.24    55.38     1.84    6.94   1.23  32.55
sda1              0.00     0.00  0.00  0.00     0.00     0.00    10.62     0.00    5.93   4.89   0.00
sda2              0.64    87.13  2.57  7.79   235.76   759.43    96.03     0.58   56.42   3.04   3.15
sda3              0.25     9.49  0.90  1.64    29.33    89.00    46.62     0.07   25.76   6.63   1.68
sda4              0.00     0.00  0.00  0.00     0.00     0.00     2.00     0.00   13.00  13.00   0.00
sda5             24.66    12.29  8.35  5.46   264.06   142.95    29.47     0.56   40.80   2.29   3.16
sda6              0.00   136.96  0.13 29.64     1.00  1332.80    44.80     0.20    6.64   1.06   3.15
sda7              0.02     6.90  0.07  0.87     7.60    62.15    74.38     0.05   55.67   6.79   0.64
sda8              0.01     3.50  0.11  7.41     5.29    87.25    12.30     0.02    2.64   1.97   1.49
sda9             15.16   174.74 132.55 66.65  9599.96  2012.67    58.30     0.35   15.75   1.46  29.07


See you
Stephane

[stephane]

Hello Dimitry

I had tested this addon on VM. It was working fine.
Unfortunately, with physical server, the output of iostat is not well formated (cf below)
The results for each device is on 2 lines ;-(

Thank you again
Stephane

[root@SRV639 bin]# iostat -x -d
Linux 2.6.18-194.el5 (SRV639)   07/05/2011

Device:         rrqm/s   wrqm/s   r/s   w/s   rsec/s   wsec/s avgrq-sz avgqu-sz   await  svctm  %util

cciss/c0d0        0.37   502.65  0.62 11.67    16.23  4114.60   336.26     0.02    1.42   0.48   0.59
cciss/c0d0p1
               0.03     0.00  0.00  0.00     0.00     0.00     0.00     0.00    4.63   4.22   0.00
cciss/c0d0p2
               0.03     0.00  0.00  0.00     0.00     0.00     0.00     0.00    4.92   4.45   0.00
cciss/c0d0p3
               0.13     2.76  0.52  0.80     0.00     0.00     0.00     0.00    2.64   1.40   0.19
cciss/c0d0p4
               0.00     0.00  0.00  0.00     0.00     0.00     0.00     0.00   37.00  37.00   0.00
cciss/c0d0p5
               0.06     0.37  0.00  0.28     0.00     0.00     0.00     0.00    1.15   0.67   0.02
cciss/c0d0p6
               0.06    12.47  0.06  4.18     0.00     0.00     0.00     0.00    0.95   0.23   0.10
cciss/c0d0p7
               0.06   487.05  0.02  6.41     0.00     0.00     0.00     0.01    1.48   0.50   0.32
cciss/c0d1       11.37   860.10 25.16  9.53  5754.04  6957.05   366.46     0.47   13.62   1.28   4.45
cciss/c0d1p1
               0.02     0.00  0.00  0.00     0.00     0.00     0.00     0.00    4.78   4.78   0.00
cciss/c0d1p2
              11.35   860.10 25.16  9.53     0.00     0.00     0.00     0.47   13.62   1.28   4.45
cciss/c1d0        0.00     0.00  0.00  0.00     0.02     0.00    39.27     0.00    6.03   6.03   0.00

[Dimitri]

Hi Stephane,

seems the last version of "sysstat" package is v10.0.1. Within the
STAT-Service looks that I've ship "iostat" from the v6, and I've also
modified the code and that's why there are more columns in output
comparing to the "native iostat" you're using (v7).

I'll prepare the iostat from v10 and will send you back to check if
it'll run better on your server.

Rgds,
-Dimitri

[stephane]

Thanks a lot, sorry to give you some homework.


stef

[Dimitri]

Hi Stef,

(sorry for delay - time is running ;-))

the new "iostat" binary is in attch. - worked well on RHEL5 server,
and hope will be ok now also on your machine..

except if something going wrong, I've already included it into
dim_STAT v.9.0 distribution, it seems to bring more info (as well has
"-p" option to print per partitions details).

Let me know if it's working well for you too :-))

Rgds,
-Dimitri

On 7/6/11, stephane <stephane....@gmail.com> wrote:
> Thanks a lot, sorry to give you some homework.
>
>
> stef
>

[stephane]

Thanks a lot, Dim

I will go for a test on monday.
I let you know then.

[stephane]

Hi Dim

I have a good and a bad news.

Let start with the bad news:
i still have some disk busy > 100% with the new iostat (up to 110%)

The good news is i succeed to reproduce the "issue" with the native iostat (V7 with RHEL5.5); using a dd command.

Then, i guess nothing can be done until redhat fixes the iostat command.

By the way, i don t remember if i told you about cores cpu usage > 100% on RHEL 5.5 as well.
I workaound this using the native mpstat

Thank you
Stephane

[Dimitri]

Hi Stef,

there is no "native" iostat or mpstat from RHEL :-)
all the stat commands are coming from "sysstat" package, and the
"mpstat" shipped with STAT-service for Linux is from the V8 of
sysstat, so should be better than V7 (normally) - or your "native"
mpstat should have the same problem too..

In attch. there is an "mpstat" from V10 of sysstat - hope it'll not
have problems on your system, otherwise it'll be fine to understand
why :-)) (probbaly some light debugging will be needed?) - another
possibility that under VM the stats are reported wrongly.. - any other
way to check it? ;-)

Rgds,
-Dimitri