"Hassan" <Hassan@hotmail.com> wrote in message
news:u3aVch2QHHA.4260@TK2MSFTNGP06.phx.gbl...
> Any issues w/ synchronous ? Would love to do so, but we are concerned
about
> the latency .
> I dont know if you saw my other thread on BCVs, but do you have to kinda
> quiesce the source at times when the destination may get out of sync,etc ?
> What happens when the destination either falls of the n/w or is slow ? Do
> you apps timeout ?
>
> "Linchi Shea" <LinchiShea@discussions.microsoft.com> wrote in message
> news:DC1C9AE1-F8D8-4771-96EA-2AD7A74041A9@microsoft.com...
> > EMC plus synchronous replication for zero data loss and manual site
> > failover
> > in most cases.
> >
> > Linchi
> >
> > "Hassan" wrote:
> >
> >> Linchi,
> >>
> >> Do you use EMC or HDS for your storage level mirroing ? And do you use
> >> synchronous or asynchronous ? Do you use manaul failover for your
remote
> >> storage level mirroing as well ?
> >>
> >>
> >> "Linchi Shea" <LinchiShea@discussions.microsoft.com> wrote in message
> >> news:19711415-07C2-4028-8AAC-761F6BC80696@microsoft.com...
> >> >> Cost is not a factor here.
> >> >
> >> > If cost is not a factor, you may consider both because they protect
> >> > different aspects of your database systems (if there is a business
> >> > requirement). Clustering is mostly for protecting against system
> >> > failure
> >> > (e.g. motherboard), and mirroring is mainly for storage hardware and
> >> > certain
> >> > database corruption as well as site failure. Though you can use
> >> > mirroring
> >> > for
> >> > protecting against system failure, clustering gives you faster and
> >> > safer
> >> > recovery with lower overhead. Any time your failover involves data,
> >> > it's
> >> > always risky. Cluster failure doesn't involve data in the database.
> >> >
> >> > We use both, though we use storage-level hardware remote mirroring
> >> > instead
> >> > of database mirroring. However, this doesn't mean that database
> >> > mirroring
> >> > is
> >> > no good. If you have thousands of databases, you probably don't want
to
> >> > maintain thousands of database mirroring setups.
> >> >
> >> > Linchi
> >> >
> >> > "Hassan" wrote:
> >> >
> >> >> I am breaking my head trying to figure what should we be using as a
> >> >> standard
> >> >> for our HA/DR planning.
> >> >>
> >> >> Can you help me figure out how to make the choice ? Cost is not a
> >> >> factor
> >> >> here. Is Mirroing a better choice than Clustering ?
> >> >>
> >> >> Help me make a wise choice for the organization...
> >> >>
> >> >>
> >> >>
> >>
> >>
> >>
>
>

"Linchi Shea" <LinchiShea@discussions.microsoft.com> wrote in message
news:6336EC94-A2EE-4173-AB73-21C2D8720ED8@microsoft.com...
>> Any issues w/ synchronous ? Would love to do so, but we are concerned
>> about
>> the latency .
>
> There is a latency, and the latency may be significant depending on how
> sentitive your app is. But the latency is manageable in most cases.
>
>> do you have to kinda
>> quiesce the source at times when the destination may get out of sync,etc
>> ?
>> What happens when the destination either falls of the n/w or is slow ?
>> Do
>> you apps timeout ?
>
> If there is a problem on the destination that may prevent synchronous
> replication, in most cases you should obviously scrifice the replication
> to
> protect the primary. But if you have to do this often, there is a bigger
> problem with the capacity or architecture.
>
> Linchi
>
> "Hassan" wrote:
>
>> Any issues w/ synchronous ? Would love to do so, but we are concerned
>> about
>> the latency .
>> I dont know if you saw my other thread on BCVs, but do you have to kinda
>> quiesce the source at times when the destination may get out of sync,etc
>> ?
>> What happens when the destination either falls of the n/w or is slow ?
>> Do
>> you apps timeout ?
>>
>> "Linchi Shea" <LinchiShea@discussions.microsoft.com> wrote in message
>> news:DC1C9AE1-F8D8-4771-96EA-2AD7A74041A9@microsoft.com...
>> > EMC plus synchronous replication for zero data loss and manual site
>> > failover
>> > in most cases.
>> >
>> > Linchi
>> >
>> > "Hassan" wrote:
>> >
>> >> Linchi,
>> >>
>> >> Do you use EMC or HDS for your storage level mirroing ? And do you
>> >> use
>> >> synchronous or asynchronous ? Do you use manaul failover for your
>> >> remote
>> >> storage level mirroing as well ?
>> >>
>> >>
>> >> "Linchi Shea" <LinchiShea@discussions.microsoft.com> wrote in message
>> >> news:19711415-07C2-4028-8AAC-761F6BC80696@microsoft.com...
>> >> >> Cost is not a factor here.
>> >> >
>> >> > If cost is not a factor, you may consider both because they protect
>> >> > different aspects of your database systems (if there is a business
>> >> > requirement). Clustering is mostly for protecting against system
>> >> > failure
>> >> > (e.g. motherboard), and mirroring is mainly for storage hardware and
>> >> > certain
>> >> > database corruption as well as site failure. Though you can use
>> >> > mirroring
>> >> > for
>> >> > protecting against system failure, clustering gives you faster and
>> >> > safer
>> >> > recovery with lower overhead. Any time your failover involves data,
>> >> > it's
>> >> > always risky. Cluster failure doesn't involve data in the database.
>> >> >
>> >> > We use both, though we use storage-level hardware remote mirroring
>> >> > instead
>> >> > of database mirroring. However, this doesn't mean that database
>> >> > mirroring
>> >> > is
>> >> > no good. If you have thousands of databases, you probably don't want
>> >> > to
>> >> > maintain thousands of database mirroring setups.
>> >> >
>> >> > Linchi
>> >> >
>> >> > "Hassan" wrote:
>> >> >
>> >> >> I am breaking my head trying to figure what should we be using as a
>> >> >> standard
>> >> >> for our HA/DR planning.
>> >> >>
>> >> >> Can you help me figure out how to make the choice ? Cost is not a
>> >> >> factor
>> >> >> here. Is Mirroing a better choice than Clustering ?
>> >> >>
>> >> >> Help me make a wise choice for the organization...
>> >> >>
>> >> >>
>> >> >>
>> >>
>> >>
>> >>
>>
>>
>>

"Linchi Shea" wrote:

> > Cost is not a factor here.
>
> If cost is not a factor, you may consider both because they protect
> different aspects of your database systems (if there is a business
> requirement). Clustering is mostly for protecting against system failure
> (e.g. motherboard), and mirroring is mainly for storage hardware and certain
> database corruption as well as site failure. Though you can use mirroring for
> protecting against system failure, clustering gives you faster and safer
> recovery with lower overhead. Any time your failover involves data, it's
> always risky. Cluster failure doesn't involve data in the database.
>
> We use both, though we use storage-level hardware remote mirroring instead
> of database mirroring. However, this doesn't mean that database mirroring is
> no good. If you have thousands of databases, you probably don't want to
> maintain thousands of database mirroring setups.
>
> Linchi
>
> "Hassan" wrote:
>
> > I am breaking my head trying to figure what should we be using as a standard
> > for our HA/DR planning.
> >
> > Can you help me figure out how to make the choice ? Cost is not a factor
> > here. Is Mirroing a better choice than Clustering ?
> >
> > Help me make a wise choice for the organization...
> >
> >