I have created the database maintenance plan for SQL Server for full backup,
differential backup and transactional backup and scheduled. If I see the job
status as completed successfully and if I can see the physically the file
being created, am I on the safe side of thinking that I have the recoverable
backup. I normally check the restore activity once in 6 months. Do I have to
check any thing else to make sure my back up is successful? Does any one has
any experience of not being able to restore the SQL Server instances from
the backup copies.
I am making sure that I am perfect in terms of backup and restore
activities.
SQL2K.
Thank you,
AllenAllenHubatka (AllenHubtka_67@.hotmail.com) writes:
> I have created the database maintenance plan for SQL Server for full
> backup, differential backup and transactional backup and scheduled. If I
> see the job status as completed successfully and if I can see the
> physically the file being created, am I on the safe side of thinking
> that I have the recoverable backup. I normally check the restore
> activity once in 6 months. Do I have to check any thing else to make
> sure my back up is successful? Does any one has any experience of not
> being able to restore the SQL Server instances from the backup copies.
There is really only one way to check if a backup is good, and that
is to restore it and run DBCC CHECKDB on it. Whether you always do this,
or only at random occassions depends on how paranoid you are. Personally,
I'd say twice a year is a tad seldom.
Not that backups fail that often, but then again maybe you have some
creeping corruption. I have experienced that DBCC on a restored file
revealed errors that did not exhibit in the original database. This
was in SQL 6.0, so this may not happen today. Then again, it's all a
matter of how paranoid you are...
Erland Sommarskog, SQL Server MVP, esquel@.sommarskog.se
Books Online for SQL Server SP3 at
http://www.microsoft.com/sql/techinfo/productdoc/2000/books.asp
Showing posts with label plan. Show all posts
Showing posts with label plan. Show all posts
Friday, March 30, 2012
Friday, March 9, 2012
Identity in replication
We are going to move Access 2002 database tables to SQL Server 2000. The
current Access tables are replicated. I plan to "unreplicate" them before
converting (otherwise all the repl system fields go over). However, many of
the tables have AutoNumber fields that we want to convert to Identity fields
in SQL. I read an article about Merge Replication (which is our plan) and
they said to set the Identity fields to NOT FOR REPLICATION. Is that true?
It seems the wording is confusing. Thanks.
David
David,
I've seen this setting for so long now that I'd forgotten how unintuitive it
seemed at first
. You're right for merge - the correct setting is Yes, Not
for Replication. This means it should function as a normal identity column,
apart from when rows originate from the replication synchronization process,
when effectively identity inserts are allowed. This setting isn't needed for
transactional or snapshot - just Identity - yes, because subscribers can't
change the data. If they are updatable, then the same applies as in merge.
Rgds,
Paul Ibison SQL Server MVP, www.replicationanswers.com
(recommended sql server 2000 replication book:
http://www.nwsu.com/0974973602p.html)
|||Paul,
My concern is with laptop users that will be synching on occasion and
will be creating records in tables with Identity columns. Access
handles this with a random long integer. Does SQL Server do the same?
This is important because I prefer not to have to assign different
unique ID's for each laptop user. Access merely creates a random
negative or positive number. Thanks.
David
*** Sent via Developersdex http://www.codecomments.com ***
Don't just participate in USENET...get rewarded for it!
|||David,
SQL Server's identity column has a seed and increment - typically 1 and 1.
There's a description in BOL of its behaviour. With this in mind, after you
convert to SQL Server, use DBCC CHECKIDENT to reseed the identity value at
the publisher to be the highest value selected. After that you can use
automatic range management and have ranges allocated to each subscriber.
Rgds,
Paul Ibison SQL Server MVP, www.replicationanswers.com
(recommended sql server 2000 replication book:
http://www.nwsu.com/0974973602p.html)
current Access tables are replicated. I plan to "unreplicate" them before
converting (otherwise all the repl system fields go over). However, many of
the tables have AutoNumber fields that we want to convert to Identity fields
in SQL. I read an article about Merge Replication (which is our plan) and
they said to set the Identity fields to NOT FOR REPLICATION. Is that true?
It seems the wording is confusing. Thanks.
David
David,
I've seen this setting for so long now that I'd forgotten how unintuitive it
seemed at first
for Replication. This means it should function as a normal identity column,
apart from when rows originate from the replication synchronization process,
when effectively identity inserts are allowed. This setting isn't needed for
transactional or snapshot - just Identity - yes, because subscribers can't
change the data. If they are updatable, then the same applies as in merge.
Rgds,
Paul Ibison SQL Server MVP, www.replicationanswers.com
(recommended sql server 2000 replication book:
http://www.nwsu.com/0974973602p.html)
|||Paul,
My concern is with laptop users that will be synching on occasion and
will be creating records in tables with Identity columns. Access
handles this with a random long integer. Does SQL Server do the same?
This is important because I prefer not to have to assign different
unique ID's for each laptop user. Access merely creates a random
negative or positive number. Thanks.
David
*** Sent via Developersdex http://www.codecomments.com ***
Don't just participate in USENET...get rewarded for it!
|||David,
SQL Server's identity column has a seed and increment - typically 1 and 1.
There's a description in BOL of its behaviour. With this in mind, after you
convert to SQL Server, use DBCC CHECKIDENT to reseed the identity value at
the publisher to be the highest value selected. After that you can use
automatic range management and have ranges allocated to each subscriber.
Rgds,
Paul Ibison SQL Server MVP, www.replicationanswers.com
(recommended sql server 2000 replication book:
http://www.nwsu.com/0974973602p.html)
Labels:
access,
database,
identity,
microsoft,
mysql,
oracle,
plan,
replicated,
replication,
server,
sql,
tables,
thecurrent,
unreplicate
Subscribe to:
Posts (Atom)