http://www.maatkit.org/

I use this to checksum masters and insure data integrity.

On Oct 8, 3:47 pm, "Frédéric Sidler" <[EMAIL PROTECTED]>
wrote:
> How do you checksum the two masters and see that they are not quite in sync
> ?
> Do you use master-master for flexibility ?
> Do you have problem because of CPU or because of network traffic or other ?
>
> Do you plan to shard on application level, do you know already ?
>
> On Wed, Oct 8, 2008 at 9:40 PM, mikeytag <[EMAIL PROTECTED]> wrote:
>
> > Hi Frederic,
>
> > I used master-master on my own hardware, and let me just say that it
> > served it's purpose but does not really scale well. We are at the
> > point now that we really only run on 1 db server, although we need 2,
> > simply because replication can not keep up with the amount of writes
> > happening. After getting up and running on Scalr I think I am going to
> > begin to do some form of sharding, but I am not there quite yet.
>
> > Mike
>
> > On Oct 8, 12:04 pm, "Frédéric Sidler" <[EMAIL PROTECTED]>
> > wrote:
> > > Mike, did you use scalr functionnalities only to achieve your goal or do
> > you
> > > have any personal recipe you can share.
>
> > > What do you use for master-master config
> > > Do you already use something for sharding, pyshard, hscale, for example ?
>
> > > --
> > > Frederic Sidler
> > >  <www.mixin.com/users/fredericsidler>
> > > mixin<www.mixin.com/users/fredericsidler>- what are your friends doing
> > > this week
>
> > > On Wed, Oct 8, 2008 at 8:30 PM, mikeytag <[EMAIL PROTECTED]> wrote:
>
> > > > Alex,
>
> > > > Thanks for the quick reply. Turns out I misconfigured my php APC
> > > > settings and thus was seeing the slowdown. It didn't make since to me
> > > > since the DB appeared to be sleeping and not really caring that it was
> > > > being backed up. Once I turned up the apc.shm_size to 128 all is well
> > > > with my application again. I will do some more backups and bundles and
> > > > let you know if I see big performance hits during.
>
> > > > Thanks,
> > > > Mike
>
> > > > On Oct 8, 11:10 am, Alex Kovalyov <[EMAIL PROTECTED]> wrote:
> > > > > Mike,
> > > > > We'd like to assist and investigate your case. I will request some
> > > > details
> > > > > over email shortly.
>
> > > > > As for sharding - it may sound obvious, but it should be planned at
> > > > earliest
> > > > > stage to avoid at least heavy data moves later.
>
> > > > > On 08.10.08 20:38, "mikeytag" <[EMAIL PROTECTED]> wrote:
>
> > > > > > Hi everyone,
>
> > > > > > I am getting ready to move most of our domains over to our Scalr
> > farm
> > > > > > that I have been working with for the past couple weeks and am so
> > > > > > excited to start using it in a production environment. Currently, I
> > am
> > > > > > only running one db in scalr which is an x.large mysqllvm64
> > instance.
> > > > > > I am curious to know what everyone would recommend as far as the
> > > > > > settings for backup and bundle. Whenever either a backup or a
> > bundle
> > > > > > runs on this db, my application becomes painfully slow, so much
> > that I
> > > > > > cannot move our stuff over until I figure this out.
>
> > > > > > A bit of background, our database currently stands around 60GB and
> > at
> > > > > > our co-lo we have a dual master configuration on two dual xeons
> > each
> > > > > > with 16GB of RAM. Part of the reason we are making the move to AWS
> > and
> > > > > > Scalr is the fact that we can distribute heavy read load on the
> > fly.
> > > > > > Which is just one of the many benefits to having your own hardware.
>
> > > > > > I am definitely planning on running at least 2 mysqllvm64 instances
> > > > > > and maybe 3. If that happens and a backup is run, which instance
> > does
> > > > > > it run the backup on? I am hoping that it takes one of the slaves
> > out
> > > > > > of rotation and uses it to backup so there is little impact in
> > > > > > performance on the application. Is that right?
>
> > > > > > In addition, our application has the lovely problem of being heavy
> > on
> > > > > > writes as well as reads. (We have farms of "workhorse" servers that
> > > > > > constantly crunch algorithms related to multivariate testing and
> > the
> > > > > > like and update the db) I am realizing that soon I need to rewrite
> > my
> > > > > > application to start sharding tables, but that is a discussion for
> > > > > > another time.
>
> > > > > > Mike

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"scalr-discuss" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at 
http://groups.google.com/group/scalr-discuss?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to