high availability on SQL Server 2000

We have a failover cluster that gives our SQL Server environment really great
availability - each upgrade we do on Windows or SQL Server requires a SQL
restart or a Windows reboot which only affects the uptime for a moment. A
couple times we had a problem with a server and with the failover the
customers didn't even notice the interruption in service. Despite this
Manegement is telling us that we are not "high availiability". Other "highly
available" environments (DB2 and Oracle) in our office take long outages on a
regular basis and have about 30-50 hours of downtime a year (planned and
unplanned) but we have less than 5-10. Management hasn't really told us what
we need to do to be considered highly available. What do the other DB
platforms do differently that makes them highly available.
ThanksPossibly the equivalent of the future "data mirroring" feature that will be available in SQL Server
2005 somewhere H1 2006.
--
Tibor Karaszi, SQL Server MVP
http://www.karaszi.com/sqlserver/default.asp
http://www.solidqualitylearning.com/
"Bobsie" <Bobsie@.discussions.microsoft.com> wrote in message
news:665DF2F5-E10B-4AA6-BF85-36AD8392BD65@.microsoft.com...
> We have a failover cluster that gives our SQL Server environment really great
> availability - each upgrade we do on Windows or SQL Server requires a SQL
> restart or a Windows reboot which only affects the uptime for a moment. A
> couple times we had a problem with a server and with the failover the
> customers didn't even notice the interruption in service. Despite this
> Manegement is telling us that we are not "high availiability". Other "highly
> available" environments (DB2 and Oracle) in our office take long outages on a
> regular basis and have about 30-50 hours of downtime a year (planned and
> unplanned) but we have less than 5-10. Management hasn't really told us what
> we need to do to be considered highly available. What do the other DB
> platforms do differently that makes them highly available.
> Thanks|||IMHO 'High Avalability' is more a marketing term than a specific designation.
That is, it can depend on who you are talking about or what they are talking
about.
From your desciption you are using shared disk in a cluster
(Active/Passive). That is the equivalent, in many ways, to Oracle's RAC (not
quite because only 1 server is active where in Oracle there could be 2,
but......). It is certainly the equivalent to Oracle's Physical Standby
Data Guard configuration. Both are variations on Hight Availability. As
would be replication, which is a third part.
Basically, I believe there are two types of failure 'recovery'. Share
Everything, where there is not data loss (but can be a big performance hit,
depending on the numbe of servers) and Share SomeThings (which is what High
Availability Is.) The timing for things like upgrades, backups, etc. have
more to do with aplication and environmental design, and how often you do
upgrades.
Bottom line: You are in high availability mode. The hour comparison is not
really valid. (And I would not mix planned and unplanned downtime. Only
unplanned downtime counts when discussing High Availability, IMHO, although
with Grid the times may be changing.)
--
Joseph R.P. Maloney, CSP,CCP,CDP
"Bobsie" wrote:
> We have a failover cluster that gives our SQL Server environment really great
> availability - each upgrade we do on Windows or SQL Server requires a SQL
> restart or a Windows reboot which only affects the uptime for a moment. A
> couple times we had a problem with a server and with the failover the
> customers didn't even notice the interruption in service. Despite this
> Manegement is telling us that we are not "high availiability". Other "highly
> available" environments (DB2 and Oracle) in our office take long outages on a
> regular basis and have about 30-50 hours of downtime a year (planned and
> unplanned) but we have less than 5-10. Management hasn't really told us what
> we need to do to be considered highly available. What do the other DB
> platforms do differently that makes them highly available.
> Thanks|||Failover using shared storage can provide continued service in the case of
server failure but you still go down in the case of storage failure. I have
come across this debate before and managers like the idea of a hot standby
for storage failure. I don't know what SQL Server does to deal with failure
of the storage device?
"Bobsie" wrote:
> We have a failover cluster that gives our SQL Server environment really great
> availability - each upgrade we do on Windows or SQL Server requires a SQL
> restart or a Windows reboot which only affects the uptime for a moment. A
> couple times we had a problem with a server and with the failover the
> customers didn't even notice the interruption in service. Despite this
> Manegement is telling us that we are not "high availiability". Other "highly
> available" environments (DB2 and Oracle) in our office take long outages on a
> regular basis and have about 30-50 hours of downtime a year (planned and
> unplanned) but we have less than 5-10. Management hasn't really told us what
> we need to do to be considered highly available. What do the other DB
> platforms do differently that makes them highly available.
> Thanks|||Hi
Storage failure is not SQL Server's problem. It is a hardware and OS
problem.SQL server uses what is given to it.
If they are worried about storage failure, are they running multiple EMC's
Symmetrix with SRDF?
Regards
--
Mike Epprecht, Microsoft SQL Server MVP
Zurich, Switzerland
IM: mike@.epprecht.net
MVP Program: http://www.microsoft.com/mvp
Blog: http://www.msmvps.com/epprecht/
"Robert Kinesta" <RobertKinesta@.discussions.microsoft.com> wrote in message
news:AB8401B6-D1C8-4E0D-9564-ABB3CB936752@.microsoft.com...
> Failover using shared storage can provide continued service in the case of
> server failure but you still go down in the case of storage failure. I
> have
> come across this debate before and managers like the idea of a hot standby
> for storage failure. I don't know what SQL Server does to deal with
> failure
> of the storage device?
> "Bobsie" wrote:
>> We have a failover cluster that gives our SQL Server environment really
>> great
>> availability - each upgrade we do on Windows or SQL Server requires a SQL
>> restart or a Windows reboot which only affects the uptime for a moment. A
>> couple times we had a problem with a server and with the failover the
>> customers didn't even notice the interruption in service. Despite this
>> Manegement is telling us that we are not "high availiability". Other
>> "highly
>> available" environments (DB2 and Oracle) in our office take long outages
>> on a
>> regular basis and have about 30-50 hours of downtime a year (planned and
>> unplanned) but we have less than 5-10. Management hasn't really told us
>> what
>> we need to do to be considered highly available. What do the other DB
>> platforms do differently that makes them highly available.
>> Thanks