Database mirroring and automatic failover

Hi,
I need some advice on database mirroring. We have our databases
mirrored successfully in a test envirnment using High Availiblity and
Automatic Failover. I have alerts configured for transaction delays
and mirroring changes etc, but we are not sure how to best handle the
production reboots. For example, we schedule reboots on the principal
server weekly - I assume this will cause automatic failover as long as
it exceeds my timeout period. Is the suspend, resume command the best
way to handle this or do I have other options? I was thinking about
scheduling a job to run these commands during the reboot. I don't want
to set the timeout for more than 30 seconds.
ALTER DATABASE DB SET PARTNER SUSPEND
--after suspension, resume
ALTER
DATABASE DB SET PARTNER RESUME
Is this the proper way to proceed? Anyone with experience with this
issue out there or something similar?
Thanks,
KristinaHi Kristina
Why do you need to schedule or to do weekly reboots on a system configured
for high availability?
Ben Nevarez
"Kristina" wrote:
> Hi,
> I need some advice on database mirroring. We have our databases
> mirrored successfully in a test envirnment using High Availiblity and
> Automatic Failover. I have alerts configured for transaction delays
> and mirroring changes etc, but we are not sure how to best handle the
> production reboots. For example, we schedule reboots on the principal
> server weekly - I assume this will cause automatic failover as long as
> it exceeds my timeout period. Is the suspend, resume command the best
> way to handle this or do I have other options? I was thinking about
> scheduling a job to run these commands during the reboot. I don't want
> to set the timeout for more than 30 seconds.
>
> ALTER DATABASE DB SET PARTNER SUSPEND
> --after suspension, resume
> ALTER
> DATABASE DB SET PARTNER RESUME
> Is this the proper way to proceed? Anyone with experience with this
> issue out there or something similar?
> Thanks,
> Kristina
>|||On Jan 8, 8:59=A0pm, Ben Nevarez <BenNeva...@.discussions.microsoft.com>
wrote:
> Hi Kristina
> Why do you need to schedule or to do weekly reboots on a system configured=
> for high availability?
> Ben Nevarez
>
> "Kristina" wrote:
> > Hi,
> > I need some advice on database mirroring. We have our databases
> > mirrored successfully in a test envirnment using High Availiblity and
> > Automatic Failover. I have alerts configured for transaction delays
> > and mirroring changes etc, but we are not sure how to best handle the
> > production reboots. For example, we schedule reboots on the principal
> > server weekly - I assume this will cause automatic failover as long as
> > it exceeds my timeout period. Is the suspend, resume command the best
> > way to handle this or do I have other options? I was thinking about
> > scheduling a job to run these commands during the reboot. I don't want
> > to set the timeout for more than 30 seconds.
> > ALTER DATABASE DB SET PARTNER SUSPEND
> > --after suspension, resume
> > ALTER
> > DATABASE DB SET PARTNER RESUME
> > Is this the proper way to proceed? Anyone with experience with this
> > issue out there or something similar?
> > Thanks,
> > Kristina- Hide quoted text -
> - Show quoted text -
Ben,
Basically these are scheduled updates from Microsoft.
Do you have any ideas around this?
Kristina|||If you are configuring a database server with high availability you should
not have weekly reboots for Windows updates. Perhaps you can apply those
updates every one or two months, and just pay special attention to security
updates.
When you reboot the principal server the database will fail over to the
mirror, which becomes the new principal. When the server is back online it
will become the new mirror. All of this is automatic with no DBA
intervention. So I do not understand why you want to change the timeout or
the status to SUSPENDED.
Ben Nevarez
"Kristina" wrote:
> On Jan 8, 8:59 pm, Ben Nevarez <BenNeva...@.discussions.microsoft.com>
> wrote:
> > Hi Kristina
> >
> > Why do you need to schedule or to do weekly reboots on a system configured
> > for high availability?
> >
> > Ben Nevarez
> >
> >
> >
> > "Kristina" wrote:
> > > Hi,
> >
> > > I need some advice on database mirroring. We have our databases
> > > mirrored successfully in a test envirnment using High Availiblity and
> > > Automatic Failover. I have alerts configured for transaction delays
> > > and mirroring changes etc, but we are not sure how to best handle the
> > > production reboots. For example, we schedule reboots on the principal
> > > server weekly - I assume this will cause automatic failover as long as
> > > it exceeds my timeout period. Is the suspend, resume command the best
> > > way to handle this or do I have other options? I was thinking about
> > > scheduling a job to run these commands during the reboot. I don't want
> > > to set the timeout for more than 30 seconds.
> >
> > > ALTER DATABASE DB SET PARTNER SUSPEND
> > > --after suspension, resume
> > > ALTER
> > > DATABASE DB SET PARTNER RESUME
> >
> > > Is this the proper way to proceed? Anyone with experience with this
> > > issue out there or something similar?
> >
> > > Thanks,
> > > Kristina- Hide quoted text -
> >
> > - Show quoted text -
> Ben,
> Basically these are scheduled updates from Microsoft.
> Do you have any ideas around this?
> Kristina
>|||"Kristina" <KristinaDBA@.gmail.com> wrote in message
news:73b02d00-0119-455c-8384-3b7e9f80b104@.f47g2000hsd.googlegroups.com...
On Jan 8, 8:59 pm, Ben Nevarez <BenNeva...@.discussions.microsoft.com>
wrote:
>Ben,
>Basically these are scheduled updates from Microsoft.
>Do you have any ideas around this?
>Kristina
Yes, don't apply them.
Heresy you say? Sort of.
For a production high availability DB I would schedule these infrequently
and only after OTHER people have tested them on their servers. If it's a
supercritical patch, I might rush it on, but typically I'm very conservative
about applying patches, especially to production systems that are already
well controlled by other methods.
For example, simply not permitting any outside access to port 1433 will
greatly increase your safety. If the outside world can't get to it, doesn't
matter if say a buffer overflow exists (i.e. Code Red).
Greg Moore
SQL Server DBA Consulting Remote and Onsite available!
Email: sql (at) greenms.com http://www.greenms.com/sqlserver.html|||On Jan 9, 10:53=A0pm, "Greg D. Moore \(Strider\)"
<mooregr_deletet...@.greenms.com> wrote:
> "Kristina" <Kristina...@.gmail.com> wrote in message
> news:73b02d00-0119-455c-8384-3b7e9f80b104@.f47g2000hsd.googlegroups.com...
> On Jan 8, 8:59 pm, Ben Nevarez <BenNeva...@.discussions.microsoft.com>
> wrote:
> >Ben,
> >Basically these are scheduled updates from Microsoft.
> >Do you have any ideas around this?
> >Kristina
> Yes, don't apply them.
> Heresy you say? =A0Sort of.
> For a production high availability DB I would schedule these infrequently
> and only after OTHER people have tested them on their servers. =A0If it's =a
> supercritical patch, I might rush it on, but typically I'm very conservati=ve
> about applying patches, especially to production systems that are already
> well controlled by other methods.
> For example, simply not permitting any outside access to port 1433 will
> greatly increase your safety. =A0If the outside world can't get to it, doe=sn't
> matter if say a buffer overflow exists (i.e. Code Red).
> --
> Greg Moore
> SQL Server DBA Consulting =A0 =A0 =A0 =A0 =A0 Remote and Onsite available!=
> Email: sql =A0(at) =A0greenms.com =A0 =A0 =A0 =A0 =A0http://www.greenms.co=
m/sqlserver.html
Thanks for all the advice guys. I am not in charge of the servers and
the patches in my current envirnment. I will be using this information
to convice the networking team to not schedule the reboots as
frequently. If they do, I will just create a job to suspend and then
re-initiaize the mirroring during the period that the server is down. I