On Fri, Aug 27, 2010 at 05:05:45PM -0400, David Nolan wrote:
> We're hearing a lot of what we consider to be FUD from the vendors
> about each others products, specifically from the dedup vendors.  I'd
> love to get some less biased opinions.

Yes, they're quite good at that. It was one of the most frustrating
parts of our purchasing efforts.

We looked seriously at Sepaton and DataDomain. Initially we had ruled
DataDomain out since they didn't have a big enough system, but they
released the DD880 while we were looking.

We ended up going with DataDomain, particularly since we back up a lot
of Oracle database, and using NFS with RMAN is a good combination.

I still have a lot of respect for Sepaton's products, and would suggest
looking hard at what they do to see if it fits your needs. 

We've had pretty good luck with DataDomain, and also have been very
happy with their responsiveness and willingness to involve engineering
about new features and optimizing what we already have.

> I'm also interested in real world experience with WAN replication
> cross country (80ms RTT) and globally (250ms RTT).   What were your
> replication windows and how much bandwidth did you require to
> replicate that data.

We replicate from Boston to Dallas, about 40ms. DataDomain's replication
only sends deduped & compressed data, so it's really efficient. I think
we peak at 300-400Mbit/sec. Here's a snapshot of the replication that's
running right now so you can see the pre-comp/post-comp usage:

sysad...@dd880-1# replication show performance all
08/28 10:53:27
       rctx://1                rctx://2                rctx://3         
 Pre-comp    Network     Pre-comp    Network     Pre-comp    Network    
  (KB/s)     (KB/s)       (KB/s)     (KB/s)       (KB/s)     (KB/s)     
 ---------  ---------    ---------  ---------    ---------  ---------   
    167283        170            0          0            0          0   
    235187        353            0          0            0          0   
    137559        182            0          0            0          0   
    239336        333            0          0            0          0   
    115908        263            0          0            0          0   
     87375        202            0          0            0          0   
    254204        244            0          0            0          0   
    158608        338            0          0            0          0   
    175434        247            0          0            0          0   
    219142        221            0          0            0          0   

Our Oracle databases have a very high rate of change, so we don't get
especially good dedupe. Here's one of their stats commands that show
dedupe over time. We do cumulative backups sunday night thru thursday
night, and our fulls run friday night to sunday morning. 

sysad...@dd880-1# filesys show compression daily-detailed
                      
From: 2010-07-24 12:00 To: 2010-08-28 09:00
                      
    Sun    Mon    Tue    Wed    Thu    Fri     Sat  Weekly                   
------- ------ ------ ------ ------ ------ ------- -------
------------------
   -25-   -26-   -27-   -28-   -29-   -30-    -31-         Date              
10027.9 1548.3 2083.6 2280.8 2571.3 2831.0  8667.8 30010.7 Pre-Comp          
  400.5  340.9  444.5  459.5  516.5  562.0   580.9  3304.8 Post-Comp         
   6.9x   1.2x   1.3x   1.3x   1.4x   1.4x    3.6x    2.4x Global-Comp Factor
   3.6x   3.8x   3.7x   3.7x   3.7x   3.7x    4.1x    3.8x Local-Comp Factor 
  25.0x   4.5x   4.7x   5.0x   5.0x   5.0x   14.9x    9.1x Total-Comp Factor 
   96.0   78.0   78.7   79.9   79.9   80.1    93.3    89.0 Reduction %       
                                                                             
    -1-    -2-    -3-    -4-    -5-    -6-     -7-                           
11878.4 2524.1 2712.0 2962.9 3650.6 3338.0  9386.9 36452.9                   
  662.9  468.4  538.7  559.9  636.9  568.4   606.9  4042.2                   
   5.1x   1.5x   1.4x   1.4x   1.5x   1.6x    4.0x    2.5x                   
   3.5x   3.6x   3.6x   3.7x   3.7x   3.7x    3.9x    3.7x                   
  17.9x   5.4x   5.0x   5.3x   5.7x   5.9x   15.5x    9.0x                   
   94.4   81.4   80.1   81.1   82.6   83.0    93.5    88.9                   
                                                                             
    -8-    -9-   -10-   -11-   -12-   -13-    -14-                           
10177.3 3158.6 3354.5 3665.9 3886.6 4088.8 11441.3 39772.8                   
  283.4  635.8  606.8  657.0  660.4  686.1  1078.2  4607.8                   
   9.7x   1.3x   1.5x   1.5x   1.6x   1.6x    2.7x    2.3x                   
   3.7x   3.7x   3.6x   3.7x   3.7x   3.7x    4.0x    3.8x                   
  35.9x   5.0x   5.5x   5.6x   5.9x   6.0x   10.6x    8.6x                   
   97.2   79.9   81.9   82.1   83.0   83.2    90.6    88.4                   
                                                                             
   -15-   -16-   -17-   -18-   -19-   -20-    -21-                           
18124.2 1390.1 1913.9 2238.9 2631.0 3047.1 15407.8 44753.0                   
 1143.1  319.9  434.7  467.7  510.7  569.2  1248.4  4693.6                   
   4.1x   1.2x   1.2x   1.3x   1.4x   1.4x    3.1x    2.5x                   
   3.8x   3.8x   3.7x   3.7x   3.7x   3.7x    3.9x    3.8x                   
  15.9x   4.3x   4.4x   4.8x   5.2x   5.4x   12.3x    9.5x                   
   93.7   77.0   77.3   79.1   80.6   81.3    91.9    89.5                   
                                                                             
   -22-   -23-   -24-   -25-   -26-   -27-    -28-                           
13911.3 1417.4 2070.1 2439.3 2735.1 3458.3 12029.0 38060.6                   
  713.1  308.8  457.4  505.7  555.6  596.1   960.7  4097.5                   
   5.4x   1.2x   1.2x   1.3x   1.3x   1.5x    3.2x    2.5x                   
   3.6x   3.7x   3.7x   3.7x   3.7x   3.9x    3.9x    3.8x                   
  19.5x   4.6x   4.5x   4.8x   4.9x   5.8x   12.5x    9.3x                   
   94.9   78.2   77.9   79.3   79.7   82.8    92.0    89.2                   
------- ------ ------ ------ ------ ------ ------- -------
                      
                Pre-Comp Post-Comp Global-Comp Local-Comp    Total-Comp
                   (GiB)     (GiB)      Factor     Factor        Factor
                                                          (Reduction %)
--------------- -------- --------- ----------- ---------- -------------
Currently Used: 318551.7   43533.3           -          -   7.3x (86.3)
Written:*                                                              
  Last 34 days  189050.0   20745.9        2.4x       3.8x   9.1x (89.0)
  Last 24 hrs    12093.5     961.0        3.2x       3.9x  12.6x (92.1)
--------------- -------- --------- ----------- ---------- -------------
 * Does not include the effects of pre-comp file deletes/truncates
   since the last cleaning on 2010/08/23 12:02:17.
Key:                                                          
       Pre-Comp = Data written before compression             
       Post-Comp = Storage used after compression             
       Global-Comp Factor = Pre-Comp / (Size after de-dupe)   
       Local-Comp Factor = (Size after de-dupe) / Post-Comp   
       Total-Comp Factor = Pre-Comp / Post-Comp               
       Reduction % = ((Pre-Comp - Post-Comp) / Pre-Comp) * 100


***

The CLI on the DataDomain is excellent. I never use the web interface.
When we were looking at the Sepaton, they had no CLI.

Let me know if you've got other questions!

-j
_______________________________________________
Tech mailing list
Tech@lopsa.org
http://lopsa.org/cgi-bin/mailman/listinfo/tech
This list provided by the League of Professional System Administrators
 http://lopsa.org/

Reply via email to