At a first glance, your production server's numbers are looking fairly
similar to the "small file workload" results of your development
server.

I thought you were saying that the development server has faster performance?

Alex.


On Tue, Aug 11, 2009 at 1:33 PM, Ed Spencer<ed_spen...@umanitoba.ca> wrote:
> I've come up with a better name for the concept of file and directory
> fragmentation which is, "Filesystem Entropy". Where, over time, an
> active and volitile  filesystem moves from an organized state to a
> disorganized state resulting in backup difficulties.
>
> Here are some stats which illustrate the issue:
>
> First the development mail server:
> ==================================
> (Jump frames, Nagle disabled and tcp_xmit_hiwat,tcp_recv_hiwat set to
> 2097152)
>
> Small file workload (copy from zfs on iscsi network to local ufs
> filesystem)
> # zpool iostat 10
>               capacity     operations    bandwidth
> pool         used  avail   read  write   read  write
> ----------  -----  -----  -----  -----  -----  -----
> space       70.5G  29.0G      3      0   247K  59.7K
> space       70.5G  29.0G    136      0  8.37M      0
> space       70.5G  29.0G    115      0  6.31M      0
> space       70.5G  29.0G    108      0  7.08M      0
> space       70.5G  29.0G    105      0  3.72M      0
> space       70.5G  29.0G    135      0  3.74M      0
> space       70.5G  29.0G    155      0  6.09M      0
> space       70.5G  29.0G    193      0  4.85M      0
> space       70.5G  29.0G    142      0  5.73M      0
> space       70.5G  29.0G    159      0  7.87M      0
>
> Large File workload (cd and dvd iso's)
> # zpool iostat 10
>               capacity     operations    bandwidth
> pool         used  avail   read  write   read  write
> ----------  -----  -----  -----  -----  -----  -----
> space       70.5G  29.0G      3      0   224K  59.8K
> space       70.5G  29.0G    462      0  57.8M      0
> space       70.5G  29.0G    427      0  53.5M      0
> space       70.5G  29.0G    406      0  50.8M      0
> space       70.5G  29.0G    430      0  53.8M      0
> space       70.5G  29.0G    382      0  47.9M      0
>
> The production mail server:
> ===========================
> Mail system is running with 790 imap users logged in (low imap work
> load).
> Two backup streams are running.
> Not using jumbo frames, nagle enabled, tcp_xmit_hiwat,tcp_recv_hiwat set
> to 2097152
>    - we've never seen any effect of changing the iscsi transport
> parameters
>      under this small file workload.
>
> # zpool iostat 10
>               capacity     operations    bandwidth
> pool         used  avail   read  write   read  write
> ----------  -----  -----  -----  -----  -----  -----
> space       1.06T   955G     96     69  5.20M  2.69M
> space       1.06T   955G    175    105  8.96M  2.22M
> space       1.06T   955G    182     16  4.47M   546K
> space       1.06T   955G    170     16  4.82M  1.85M
> space       1.06T   955G    145    159  4.23M  3.19M
> space       1.06T   955G    138     15  4.97M  92.7K
> space       1.06T   955G    134     15  3.82M  1.71M
> space       1.06T   955G    109    123  3.07M  3.08M
> space       1.06T   955G    106     11  3.07M  1.34M
> space       1.06T   955G    120     17  3.69M  1.74M
>
> # prstat -mL
>   PID USERNAME USR SYS TRP TFL DFL LCK SLP LAT VCX ICX SCL SIG
> PROCESS/LWPID
>  12438 root      12 6.9 0.0 0.0 0.0 0.0  81 0.1 508  84  4K   0 save/1
>  27399 cyrus     15 0.5 0.0 0.0 0.0 0.0  85 0.0  18  10 297   0 imapd/1
>  20230 root     3.9 8.0 0.0 0.0 0.0 0.0  88 0.1 393  33  2K   0 save/1
>  25913 root     0.5 3.3 0.0 0.0 0.0 0.0  96 0.0  22   2  1K   0 prstat/1
>  20495 cyrus    1.1 0.2 0.0 0.0 0.5 0.0  98 0.0  14   3 191   0 imapd/1
>  1051 cyrus    1.2 0.0 0.0 0.0 0.0 0.0  99 0.0  19   1  80   0 master/1
>  24350 cyrus    0.5 0.5 0.0 0.0 1.4 0.0  98 0.0  57   1 484   0 lmtpd/1
>  22645 cyrus    0.6 0.3 0.0 0.0 0.0 0.0  99 0.0  53   1 603   0 imapd/1
>  24904 cyrus    0.3 0.4 0.0 0.0 0.0 0.0  99 0.0  66   0 863   0 imapd/1
>  18139 cyrus    0.3 0.2 0.0 0.0 0.0 0.0  99 0.0  24   0 195   0 imapd/1
>  21459 cyrus    0.2 0.3 0.0 0.0 0.0 0.0  99 0.0  54   0 635   0 imapd/1
>  24891 cyrus    0.3 0.3 0.0 0.0 0.9 0.0  99 0.0  28   0 259   0 lmtpd/1
>   388 root     0.2 0.3 0.0 0.0 0.0 0.0 100 0.0   1   1  48   0
> in.routed/1
>  21643 cyrus    0.2 0.3 0.0 0.0 0.2 0.0  99 0.0  49   7 540   0 imapd/1
>  18684 cyrus    0.2 0.3 0.0 0.0 0.0 0.0 100 0.0  48   1 544   0 imapd/1
>  25398 cyrus    0.2 0.2 0.0 0.0 0.0 0.0 100 0.0  47   0 466   0 pop3d/1
>  23724 cyrus    0.2 0.2 0.0 0.0 0.0 0.0 100 0.0  47   0 540   0 imapd/1
>  24909 cyrus    0.1 0.2 0.0 0.0 0.2 0.0  99 0.0  25   1 251   0 lmtpd/1
>  16317 cyrus    0.2 0.2 0.0 0.0 0.0 0.0 100 0.0  37   1 495   0 imapd/1
>  28243 cyrus    0.1 0.3 0.0 0.0 0.0 0.0 100 0.0  32   0 289   0 imapd/1
>  20097 cyrus    0.1 0.2 0.0 0.0 0.3 0.0  99 0.0  26   5 253   0 lmtpd/1
> Total: 893 processes, 1125 lwps, load averages: 1.14, 1.16, 1.16
>
> --
> Ed
>
> _______________________________________________
> zfs-discuss mailing list
> zfs-discuss@opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
>



-- 

Ted Turner  - "Sports is like a war without the killing." -
http://www.brainyquote.com/quotes/authors/t/ted_turner.html
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss

Reply via email to