cluster and tempdb on c:\

#41

Quote:

Plus, failover is supposed to be transparent, and if you rebuild tempdb
every time, then users will lose data in temp tables, possibly table
variables, and even any sorting or hashing that is occurring at the time
of failover.

Wait a minute! Why would any data in tempdb be lost during a failover even
if tempdb is allowed to locate on a local drive? When SQL Server restarts,
nothing in tempdb is retained anyway. You won't be able to keep any table
variables or sorting/hashing intermediate data.

If tempdb is on a shared drive, why would anything from tempdb be wiped out?
Technically, you're not "restarting" SQL Server.

Anyway, you may be right, I haven't explicitly checked this, it is just an
assumption based on casual observation. We have some high volume systems
that have failed over during busy times and did not hear a single complaint
from our users, and our test harnesses (which certainly call stored
procedures continuously, that make heavy use of tempdb) did not register a
blip either. Maybe just coincidence that nothing was ever noticed.

#42

To a client app, a failover is exactly the same as a SQL instance restart on
the same server. Nothing is different. Your test harness or users don't
complain for exactly the same reason they would not complain with the app
running against a standalone SQL instance: the stored procedure would create
whatever temp tables it may use when it is executed again.

Linchi

"Aaron Bertrand [SQL Server MVP]" wrote:

Quote:

Plus, failover is supposed to be transparent, and if you rebuild tempdb
every time, then users will lose data in temp tables, possibly table
variables, and even any sorting or hashing that is occurring at the time
of failover.

Wait a minute! Why would any data in tempdb be lost during a failover even
if tempdb is allowed to locate on a local drive? When SQL Server restarts,
nothing in tempdb is retained anyway. You won't be able to keep any table
variables or sorting/hashing intermediate data.

If tempdb is on a shared drive, why would anything from tempdb be wiped out?
Technically, you're not "restarting" SQL Server.

Anyway, you may be right, I haven't explicitly checked this, it is just an
assumption based on casual observation. We have some high volume systems
that have failed over during busy times and did not hear a single complaint
from our users, and our test harnesses (which certainly call stored
procedures continuously, that make heavy use of tempdb) did not register a
blip either. Maybe just coincidence that nothing was ever noticed.

#43

To a client app, a failover is exactly the same as a SQL instance restart on
the same server. Nothing is different. Your test harness or users don't
complain for exactly the same reason they would not complain with the app
running against a standalone SQL instance: the stored procedure would create
whatever temp tables it may use when it is executed again.

Linchi

"Aaron Bertrand [SQL Server MVP]" wrote:

Quote:

Plus, failover is supposed to be transparent, and if you rebuild tempdb
every time, then users will lose data in temp tables, possibly table
variables, and even any sorting or hashing that is occurring at the time
of failover.

Wait a minute! Why would any data in tempdb be lost during a failover even
if tempdb is allowed to locate on a local drive? When SQL Server restarts,
nothing in tempdb is retained anyway. You won't be able to keep any table
variables or sorting/hashing intermediate data.

If tempdb is on a shared drive, why would anything from tempdb be wiped out?
Technically, you're not "restarting" SQL Server.

Anyway, you may be right, I haven't explicitly checked this, it is just an
assumption based on casual observation. We have some high volume systems
that have failed over during busy times and did not hear a single complaint
from our users, and our test harnesses (which certainly call stored
procedures continuously, that make heavy use of tempdb) did not register a
blip either. Maybe just coincidence that nothing was ever noticed.

#44

To a client app, a failover is exactly the same as a SQL instance restart on
the same server. Nothing is different. Your test harness or users don't
complain for exactly the same reason they would not complain with the app
running against a standalone SQL instance: the stored procedure would create
whatever temp tables it may use when it is executed again.

Linchi

"Aaron Bertrand [SQL Server MVP]" wrote:

Quote:

Plus, failover is supposed to be transparent, and if you rebuild tempdb
every time, then users will lose data in temp tables, possibly table
variables, and even any sorting or hashing that is occurring at the time
of failover.

Wait a minute! Why would any data in tempdb be lost during a failover even
if tempdb is allowed to locate on a local drive? When SQL Server restarts,
nothing in tempdb is retained anyway. You won't be able to keep any table
variables or sorting/hashing intermediate data.

If tempdb is on a shared drive, why would anything from tempdb be wiped out?
Technically, you're not "restarting" SQL Server.

Anyway, you may be right, I haven't explicitly checked this, it is just an
assumption based on casual observation. We have some high volume systems
that have failed over during busy times and did not hear a single complaint
from our users, and our test harnesses (which certainly call stored
procedures continuously, that make heavy use of tempdb) did not register a
blip either. Maybe just coincidence that nothing was ever noticed.

#45

To a client app, a failover is exactly the same as a SQL instance restart on
the same server. Nothing is different. Your test harness or users don't
complain for exactly the same reason they would not complain with the app
running against a standalone SQL instance: the stored procedure would create
whatever temp tables it may use when it is executed again.

Linchi

"Aaron Bertrand [SQL Server MVP]" wrote:

Quote:

Plus, failover is supposed to be transparent, and if you rebuild tempdb
every time, then users will lose data in temp tables, possibly table
variables, and even any sorting or hashing that is occurring at the time
of failover.

Wait a minute! Why would any data in tempdb be lost during a failover even
if tempdb is allowed to locate on a local drive? When SQL Server restarts,
nothing in tempdb is retained anyway. You won't be able to keep any table
variables or sorting/hashing intermediate data.

If tempdb is on a shared drive, why would anything from tempdb be wiped out?
Technically, you're not "restarting" SQL Server.

Anyway, you may be right, I haven't explicitly checked this, it is just an
assumption based on casual observation. We have some high volume systems
that have failed over during busy times and did not hear a single complaint
from our users, and our test harnesses (which certainly call stored
procedures continuously, that make heavy use of tempdb) did not register a
blip either. Maybe just coincidence that nothing was ever noticed.

#46

To a client app, a failover is exactly the same as a SQL instance restart on
the same server. Nothing is different. Your test harness or users don't
complain for exactly the same reason they would not complain with the app
running against a standalone SQL instance: the stored procedure would create
whatever temp tables it may use when it is executed again.

Linchi

"Aaron Bertrand [SQL Server MVP]" wrote:

Quote:

Plus, failover is supposed to be transparent, and if you rebuild tempdb
every time, then users will lose data in temp tables, possibly table
variables, and even any sorting or hashing that is occurring at the time
of failover.

Wait a minute! Why would any data in tempdb be lost during a failover even
if tempdb is allowed to locate on a local drive? When SQL Server restarts,
nothing in tempdb is retained anyway. You won't be able to keep any table
variables or sorting/hashing intermediate data.

If tempdb is on a shared drive, why would anything from tempdb be wiped out?
Technically, you're not "restarting" SQL Server.

Anyway, you may be right, I haven't explicitly checked this, it is just an
assumption based on casual observation. We have some high volume systems
that have failed over during busy times and did not hear a single complaint
from our users, and our test harnesses (which certainly call stored
procedures continuously, that make heavy use of tempdb) did not register a
blip either. Maybe just coincidence that nothing was ever noticed.

#47

To a client app, a failover is exactly the same as a SQL instance restart on
the same server. Nothing is different. Your test harness or users don't
complain for exactly the same reason they would not complain with the app
running against a standalone SQL instance: the stored procedure would create
whatever temp tables it may use when it is executed again.

Linchi

"Aaron Bertrand [SQL Server MVP]" wrote:

Quote:

Plus, failover is supposed to be transparent, and if you rebuild tempdb
every time, then users will lose data in temp tables, possibly table
variables, and even any sorting or hashing that is occurring at the time
of failover.

Wait a minute! Why would any data in tempdb be lost during a failover even
if tempdb is allowed to locate on a local drive? When SQL Server restarts,
nothing in tempdb is retained anyway. You won't be able to keep any table
variables or sorting/hashing intermediate data.

If tempdb is on a shared drive, why would anything from tempdb be wiped out?
Technically, you're not "restarting" SQL Server.

Anyway, you may be right, I haven't explicitly checked this, it is just an
assumption based on casual observation. We have some high volume systems
that have failed over during busy times and did not hear a single complaint
from our users, and our test harnesses (which certainly call stored
procedures continuously, that make heavy use of tempdb) did not register a
blip either. Maybe just coincidence that nothing was ever noticed.

#48

To a client app, a failover is exactly the same as a SQL instance restart on
the same server. Nothing is different. Your test harness or users don't
complain for exactly the same reason they would not complain with the app
running against a standalone SQL instance: the stored procedure would create
whatever temp tables it may use when it is executed again.

Linchi

"Aaron Bertrand [SQL Server MVP]" wrote:

Quote:

Plus, failover is supposed to be transparent, and if you rebuild tempdb
every time, then users will lose data in temp tables, possibly table
variables, and even any sorting or hashing that is occurring at the time
of failover.

Wait a minute! Why would any data in tempdb be lost during a failover even
if tempdb is allowed to locate on a local drive? When SQL Server restarts,
nothing in tempdb is retained anyway. You won't be able to keep any table
variables or sorting/hashing intermediate data.

If tempdb is on a shared drive, why would anything from tempdb be wiped out?
Technically, you're not "restarting" SQL Server.

Anyway, you may be right, I haven't explicitly checked this, it is just an
assumption based on casual observation. We have some high volume systems
that have failed over during busy times and did not hear a single complaint
from our users, and our test harnesses (which certainly call stored
procedures continuously, that make heavy use of tempdb) did not register a
blip either. Maybe just coincidence that nothing was ever noticed.

#49

To a client app, a failover is exactly the same as a SQL instance restart on
the same server. Nothing is different. Your test harness or users don't
complain for exactly the same reason they would not complain with the app
running against a standalone SQL instance: the stored procedure would create
whatever temp tables it may use when it is executed again.

Linchi

"Aaron Bertrand [SQL Server MVP]" wrote:

Quote:

Plus, failover is supposed to be transparent, and if you rebuild tempdb
every time, then users will lose data in temp tables, possibly table
variables, and even any sorting or hashing that is occurring at the time
of failover.

Wait a minute! Why would any data in tempdb be lost during a failover even
if tempdb is allowed to locate on a local drive? When SQL Server restarts,
nothing in tempdb is retained anyway. You won't be able to keep any table
variables or sorting/hashing intermediate data.

If tempdb is on a shared drive, why would anything from tempdb be wiped out?
Technically, you're not "restarting" SQL Server.

Anyway, you may be right, I haven't explicitly checked this, it is just an
assumption based on casual observation. We have some high volume systems
that have failed over during busy times and did not hear a single complaint
from our users, and our test harnesses (which certainly call stored
procedures continuously, that make heavy use of tempdb) did not register a
blip either. Maybe just coincidence that nothing was ever noticed.

#50

To a client app, a failover is exactly the same as a SQL instance restart on
the same server. Nothing is different. Your test harness or users don't
complain for exactly the same reason they would not complain with the app
running against a standalone SQL instance: the stored procedure would create
whatever temp tables it may use when it is executed again.

Linchi

"Aaron Bertrand [SQL Server MVP]" wrote:

Quote:

Plus, failover is supposed to be transparent, and if you rebuild tempdb
every time, then users will lose data in temp tables, possibly table
variables, and even any sorting or hashing that is occurring at the time
of failover.

Wait a minute! Why would any data in tempdb be lost during a failover even
if tempdb is allowed to locate on a local drive? When SQL Server restarts,
nothing in tempdb is retained anyway. You won't be able to keep any table
variables or sorting/hashing intermediate data.

If tempdb is on a shared drive, why would anything from tempdb be wiped out?
Technically, you're not "restarting" SQL Server.

Anyway, you may be right, I haven't explicitly checked this, it is just an
assumption based on casual observation. We have some high volume systems
that have failed over during busy times and did not hear a single complaint
from our users, and our test harnesses (which certainly call stored
procedures continuously, that make heavy use of tempdb) did not register a
blip either. Maybe just coincidence that nothing was ever noticed.