List:Replication« Previous MessageNext Message »
From:Johan De Meersman Date:September 14 2010 6:29am
Subject:Re: Filtered data replication
View as plain text  
Well, traditionally, partitioning :-)

I honestly don't know if the partition/row replication/replicate-do-table
thing will work, but I suppose it's worth a shot.

The easy, straightforward option, of course, would be to partition and
replicate the whole junk, then just query the remote slaves on the partition
you're interested in; but I suppose there's a reason for not doing that.
Bandwidth, storage, ... ?



On Tue, Sep 14, 2010 at 8:19 AM, Ritesh Nadhani <riteshn@stripped> wrote:

> Hi
>
> Thanks for the response. We had actually looked into partitioning
> before but there seems to be no standard way of doing this and
> anything we tried to do was hacky at its best.
>
> Since there is no way to do it, makes me think if this is a good design or
> not.
>
> How does one suggestion splitting data like this?
>
> On Mon, Sep 13, 2010 at 11:12 PM, Johan De Meersman <vegivamp@stripped>
> wrote:
> > As far as I can remember (what'd I eat this morning again ?) there's no
> > possibility to stick a where-clause on a replication slave.
> >
> > You could MAYBE (and depending on your mysql version) partition your
> > issue_table on the issue_from field, and use replicate-do-table to only
> > replicate the appropriate partition to the corresponding slave.
> >
> > IF it works like that, it'll probably only work by using row-based
> > replication, not statement-based. The latter is however the default
> > configuration - read up on the doc for the caveats.
> >
> > That configuration is, however, almost certainly not supported, to the
> point
> > that people will give you strange looks when you mention it, and keep
> their
> > children away from you.
> >
> > Also, I have not actually tried this, so do set up a test environment :-)
> >
> >
> > Another option I thought of, is to partion and replicate, and switch the
> > useless partitions on each slave to the blackhole engine - but I strongly
> > suspect that the partitioning engine will throw a fit if you do.
> >
> > On Tue, Sep 14, 2010 at 5:51 AM, Ritesh Nadhani <riteshn@stripped>
> wrote:
> >>
> >> Hello
> >>
> >> In a new system that we are working on, we have a table:
> >>
> >> issue_table
> >> =============
> >> id, issue_from, details
> >> 1, machine1, "some details"
> >> 2, machine1, "some details again"
> >> 3, machine2, "some details"
> >> 4, machine2, "some details"
> >>
> >> Is it possible to setup a master slave replication where, where I can
> >> filter what data goes to which slave? In simple terms, after
> >> replication I would like something like:
> >>
> >> slave1.issue_table
> >> ===============
> >> id, issue_from, details
> >> 1, machine1, "some details"
> >> 2, machine1, "some details again"
> >>
> >> slave2.issue_table
> >> =================
> >> id, issue_from, details
> >> 3, machine2, "some details"
> >> 4, machine2, "some details"
> >>
> >> Is replication the right tool to do this kind of work? Or there is a
> >> better design to handle such problems. Basically, due to security and
> >> management decision, we would like to keep the data to the slaves
> >> sharded.
> >>
> >> Even a pointer to the right tutorial/document would be helpful.
> >>
> >> --
> >> Ritesh
> >>
> >> --
> >> MySQL Replication Mailing List
> >> For list archives: http://lists.mysql.com/replication
> >> To unsubscribe:
> >>  http://lists.mysql.com/replication?unsub=1
> >>
> >
> >
> >
> > --
> > Bier met grenadyn
> > Is als mosterd by den wyn
> > Sy die't drinkt, is eene kwezel
> > Hy die't drinkt, is ras een ezel
> >
>
>
>
> --
> Ritesh
> http://www.riteshn.com
>
> --
> MySQL Replication Mailing List
> For list archives: http://lists.mysql.com/replication
> To unsubscribe:
> http://lists.mysql.com/replication?unsub=1
>
>


-- 
Bier met grenadyn
Is als mosterd by den wyn
Sy die't drinkt, is eene kwezel
Hy die't drinkt, is ras een ezel

Thread
Filtered data replicationRitesh Nadhani14 Sep
  • Re: Filtered data replicationJohan De Meersman14 Sep
    • Re: Filtered data replicationRitesh Nadhani14 Sep
      • Re: Filtered data replicationJohan De Meersman14 Sep
        • Re: Filtered data replicationRitesh Nadhani14 Sep
          • Re: Filtered data replicationJohan De Meersman14 Sep
            • RE: Filtered data replicationAndrew Morgan14 Sep
              • Re: Filtered data replicationRitesh Nadhani14 Sep
              • RE: Filtered data replicationAndrew Morgan14 Sep
                • Re: Filtered data replicationRitesh Nadhani14 Sep