In the office network domain I've already setup replication between two SQL
Servers( SQL11Server & SQL22Server). Now I am extending replication with SQL
servers in other non-trusted domains/workgroups (SQL33Server & SQL44Server).
SQL33Server & SQL44Server are in non-trusted domains connecting to the
office via VPN. SQL33Server can ping SQL11Server.mydomain.local fine,
but it cannot ping just SQL11Server. Is this going to be a problem if it
pings the full DNS address?
When I started MAKEPIPE on SQL11Server, from SQL33Server I have to type
READPIPE /Ssql11server.mydomain.local for it to work. If I type READPIPE
/Ssql11Server it doesn't work..
Do I have a NETBIOS resolution problem ' Or what is the problem and how can
I successful established the connections through VPN? What the steps I
must take to established a successful VPN replication?use a host file for this.
For instance your host file should contain this entry
123.345.567.678 SQL11Server
replacing the numbers with your IP address for SQL11Server
If a netbios name has a . in it, the resolution will be using DNS. It should
like your name resolution service (probably WINS) can figure out where
SQL111Server is.
"Joe Mine" <huytuanattpgdotcomdotau> wrote in message
news:utbkWlm%23DHA.2664@.TK2MSFTNGP09.phx.gbl...
> In the office network domain I've already setup replication between two
SQL
> Servers( SQL11Server & SQL22Server). Now I am extending replication with
SQL
> servers in other non-trusted domains/workgroups (SQL33Server &
SQL44Server).
> SQL33Server & SQL44Server are in non-trusted domains connecting to the
> office via VPN. SQL33Server can ping SQL11Server.mydomain.local fine,
> but it cannot ping just SQL11Server. Is this going to be a problem if it
> pings the full DNS address?
> When I started MAKEPIPE on SQL11Server, from SQL33Server I have to type
> READPIPE /Ssql11server.mydomain.local for it to work. If I type
READPIPE
> /Ssql11Server it doesn't work..
> Do I have a NETBIOS resolution problem ' Or what is the problem and how
can
> I successful established the connections through VPN? What the steps I
> must take to established a successful VPN replication?
>
>
Showing posts with label office. Show all posts
Showing posts with label office. Show all posts
Friday, March 30, 2012
Friday, March 23, 2012
replication question
My situation:
I have a DB hosted by a 3rd party ISP. I want to replicate that DB locally
with in my office on our local server, mainly for backup purposes. I then
want to replicate that backup DB to a development DB for testing purposes.
Basically I'll have a LIVE DB hosted at the ISP. A backup version of the
LIVE
DB on our local server (peace of mind really) Then a development DB for
testing
purposes.
The Publisher will be the DB at the ISP.
Question 1. Is this a doable or a reasonable solution. I figured
transaction replication
from LIVE to BACKUP. Then Snapshot from BACKUP to
DEVELOPMENT.
It isn't absolutely necessary to have the most recent
data on the development db
since I will only be using it for testing purposes. But
would be nice to draw from
real data, almost live situations.
Question 2. Can a DB or server be both a Publisher and a Subscriber in this
case the BACKUP DB
would be both?
Thanks,
George
George,
this is possible - it is known as republishing. In your case my preference
would be to use merge between the live system and the Backup system, as it
is ideal for this sort of high latency situation. This would allow you to
have the Backup system as the publisher (with merge, the
publisher/subscriber metaphor doesn't especially hold), as you can then
control the configuration settings locally, and I'd have the live system as
a push subscriber. The other system is used as a testing system and the data
could become incompatible with replication from the Backup system, so I
wouldn't use replication to transfer to this box. Instead, I'd use a backup
and restore of the whole database - Backup to Test.
HTH,
Paul Ibison
|||I'd use transactional as
1) it offers better performance
2) it doesn't add a GUID key to every table.
The con of transactional replication is that you need a PK on every table you wish to replicate, and with republishing transactional is more sensitive to schema changes. By more sensitive I mean you have to drop your publications and subscriptions, make c
hanges and re-build again.
Looking for a SQL Server replication book?
http://www.nwsu.com/0974973602.html
sql
I have a DB hosted by a 3rd party ISP. I want to replicate that DB locally
with in my office on our local server, mainly for backup purposes. I then
want to replicate that backup DB to a development DB for testing purposes.
Basically I'll have a LIVE DB hosted at the ISP. A backup version of the
LIVE
DB on our local server (peace of mind really) Then a development DB for
testing
purposes.
The Publisher will be the DB at the ISP.
Question 1. Is this a doable or a reasonable solution. I figured
transaction replication
from LIVE to BACKUP. Then Snapshot from BACKUP to
DEVELOPMENT.
It isn't absolutely necessary to have the most recent
data on the development db
since I will only be using it for testing purposes. But
would be nice to draw from
real data, almost live situations.
Question 2. Can a DB or server be both a Publisher and a Subscriber in this
case the BACKUP DB
would be both?
Thanks,
George
George,
this is possible - it is known as republishing. In your case my preference
would be to use merge between the live system and the Backup system, as it
is ideal for this sort of high latency situation. This would allow you to
have the Backup system as the publisher (with merge, the
publisher/subscriber metaphor doesn't especially hold), as you can then
control the configuration settings locally, and I'd have the live system as
a push subscriber. The other system is used as a testing system and the data
could become incompatible with replication from the Backup system, so I
wouldn't use replication to transfer to this box. Instead, I'd use a backup
and restore of the whole database - Backup to Test.
HTH,
Paul Ibison
|||I'd use transactional as
1) it offers better performance
2) it doesn't add a GUID key to every table.
The con of transactional replication is that you need a PK on every table you wish to replicate, and with republishing transactional is more sensitive to schema changes. By more sensitive I mean you have to drop your publications and subscriptions, make c
hanges and re-build again.
Looking for a SQL Server replication book?
http://www.nwsu.com/0974973602.html
sql
Wednesday, March 21, 2012
Replication Problem
We have same database at our Head office and at factory. we use merge replication over internet to synchronize both databases.
The problem is that the internet connection is not good and keep on breaking. Is there some other way to replicate database.Originally posted by saleem
We have same database at our Head office and at factory. we use merge replication over internet to synchronize both databases.
The problem is that the internet connection is not good and keep on breaking. Is there some other way to replicate database.
Have a look at Agent profiles, for the merge agent. Perhaps change the profile to a "Slow Link Agent Profile."
Tim|||Dear colleague
You can solve (partially) the problem by augmenting the timeout factor in the merge-agent.
Simply open SQL manager, and look into the database you need, under the folder replication manager, and go to the Agents folder. Open the Merge agents subfolder and look for the one which replicates the table(s) you need.
Right click on the merge agent of your choice and select properties.
Go to the Tab Steps and select the step Run Agent. click on Edit and you'll see the merge instruction appear. At the end of this instruction you can insert a timeout factor: -QueryTimeOut 120 (120 seconds).
Save the edited entry and try it out. Try also several timeout factors ( in seconds) for optimal performance.
I hope this helps you!
Greetings,
Vincent
Originally posted by saleem
We have same database at our Head office and at factory. we use merge replication over internet to synchronize both databases.
The problem is that the internet connection is not good and keep on breaking. Is there some other way to replicate database.
The problem is that the internet connection is not good and keep on breaking. Is there some other way to replicate database.Originally posted by saleem
We have same database at our Head office and at factory. we use merge replication over internet to synchronize both databases.
The problem is that the internet connection is not good and keep on breaking. Is there some other way to replicate database.
Have a look at Agent profiles, for the merge agent. Perhaps change the profile to a "Slow Link Agent Profile."
Tim|||Dear colleague
You can solve (partially) the problem by augmenting the timeout factor in the merge-agent.
Simply open SQL manager, and look into the database you need, under the folder replication manager, and go to the Agents folder. Open the Merge agents subfolder and look for the one which replicates the table(s) you need.
Right click on the merge agent of your choice and select properties.
Go to the Tab Steps and select the step Run Agent. click on Edit and you'll see the merge instruction appear. At the end of this instruction you can insert a timeout factor: -QueryTimeOut 120 (120 seconds).
Save the edited entry and try it out. Try also several timeout factors ( in seconds) for optimal performance.
I hope this helps you!
Greetings,
Vincent
Originally posted by saleem
We have same database at our Head office and at factory. we use merge replication over internet to synchronize both databases.
The problem is that the internet connection is not good and keep on breaking. Is there some other way to replicate database.
Tuesday, March 20, 2012
Replication problem
Hi all.
I work for a chain of stores and I want to ask some things.
We have two programs:
The first one is for head office, we suppose that his name is Head Office
The second one is for every store the same and we suppose that his name is Store
Till now, two programs were made in MS Access. Every day are made communications between the stores and Head Office.
How?
The Store has built-in a process of exporting to Head, with which it puts all the information added in one day (basically the selling, buys and transfers) in a database also Access that is compressed in zip and is sent to Head Office across the e-mail. This operation it does of every shop after closing in the night. In the following day in the morning, the Head office imports the .mdb from each shop separatelly (this before opening the shops), in a process of import of shop. As soon as all the imports were done, in Head Office there is a process that re-calculates the stocks and updates the principal database. After, the Head office has another process of exporting towards the shops, the new stock and some of the changes done in the previous day in Head Office, changes that are sent to the stores to make the computing of the inventory and other things (ex-. if there are new providers, or new families, etc.) This proces also generate one database Access, which is sent to stores across the e-mail. The whole process of recalculation and sending it takes a few minutes. And finally, the stores import the mail, which is the same for all, in a built-in process of importing from Head, where some tables are independent from store and are imported completely en each store, and other tables have the information that belong to more than one store and each store import only the data that belong to that store. The import in stores does before opening the stores.
This is for now.
From almost two months, I work to change the two programs to adp projects (FE), using MS SQL Server in head office and MSDE in shops (BE). Almost I have finished with the translation. But I don't know how to do the communications, I have read enough on the replication and really don't know what type of replication is better for my case.
Some suggestions?
Greetings to all, :)
koci.
PS. Sorry for my poor english, I'm not a native.Hi koci, and don't warry for your english, my english is realy baddd.
now, let me know your problem.
you have a system when your db is MS Access, and the "replication" is make by your program in a bach process, and you need know what replication type use in SQL Server?
ok, the sql server have three type of sync.
1. snapshot (in your case NO)
2. transactional (maybe, but, do you analize the connection, transaccion for day, etc, or you db log size increment much)
3. merge (OK)
in yor case: MERGE
I have a system very "parecido", and I use the merge, this type of replication is very good for system when not 100% of time was connected, you can run then replication from the Windows Synchronize and/or schedule the sync.
Espero te sirva, y que entiendas lo que quise poner.
Sludos
Abel.|||Gracias, achiola, voy a probar este tipo de replicacion, I will try this type of replication.
Saludos,
koci.
I work for a chain of stores and I want to ask some things.
We have two programs:
The first one is for head office, we suppose that his name is Head Office
The second one is for every store the same and we suppose that his name is Store
Till now, two programs were made in MS Access. Every day are made communications between the stores and Head Office.
How?
The Store has built-in a process of exporting to Head, with which it puts all the information added in one day (basically the selling, buys and transfers) in a database also Access that is compressed in zip and is sent to Head Office across the e-mail. This operation it does of every shop after closing in the night. In the following day in the morning, the Head office imports the .mdb from each shop separatelly (this before opening the shops), in a process of import of shop. As soon as all the imports were done, in Head Office there is a process that re-calculates the stocks and updates the principal database. After, the Head office has another process of exporting towards the shops, the new stock and some of the changes done in the previous day in Head Office, changes that are sent to the stores to make the computing of the inventory and other things (ex-. if there are new providers, or new families, etc.) This proces also generate one database Access, which is sent to stores across the e-mail. The whole process of recalculation and sending it takes a few minutes. And finally, the stores import the mail, which is the same for all, in a built-in process of importing from Head, where some tables are independent from store and are imported completely en each store, and other tables have the information that belong to more than one store and each store import only the data that belong to that store. The import in stores does before opening the stores.
This is for now.
From almost two months, I work to change the two programs to adp projects (FE), using MS SQL Server in head office and MSDE in shops (BE). Almost I have finished with the translation. But I don't know how to do the communications, I have read enough on the replication and really don't know what type of replication is better for my case.
Some suggestions?
Greetings to all, :)
koci.
PS. Sorry for my poor english, I'm not a native.Hi koci, and don't warry for your english, my english is realy baddd.
now, let me know your problem.
you have a system when your db is MS Access, and the "replication" is make by your program in a bach process, and you need know what replication type use in SQL Server?
ok, the sql server have three type of sync.
1. snapshot (in your case NO)
2. transactional (maybe, but, do you analize the connection, transaccion for day, etc, or you db log size increment much)
3. merge (OK)
in yor case: MERGE
I have a system very "parecido", and I use the merge, this type of replication is very good for system when not 100% of time was connected, you can run then replication from the Windows Synchronize and/or schedule the sync.
Espero te sirva, y que entiendas lo que quise poner.
Sludos
Abel.|||Gracias, achiola, voy a probar este tipo de replicacion, I will try this type of replication.
Saludos,
koci.
Monday, March 12, 2012
Replication over satelite connection.
Hi everybody, we are designing an application for ships, and I want to replicate data between the ships and a home office. TCP/IP communication is not possible, but we have a e-mail based system for transferring files.
What is the best solution? Is it possible to have the publisher/subscriber to leave a file, and then let the e-mail system transfer the file and then later the subscriber/publisher can pick up the file and do the merge?
Any ideas?
Thanx, Kjetil
Kjetil,
this is not directly possible in the replication setup. You might want to
use DTS to import the file into a database then in the DTS package. Setup
replication between this staging database and your central system and you
can initiate replication locally.
HTH,
Paul Ibison
|||Thanks, this is what I tried to say to my boss. But how can this be implemented? Should I use merge replication or transactional replication.
In my home office, I will need a 'copy' of the ship database, in the first stage 9 ships.
Is it possible to block the database in the ships for updates while the replication process is running?
Or should I use transactional replication, will I then need a 'proxy database' at each ship and one for each ship in the home office. Is it possible to implement conflict resolution methods that guarantee the same result at all sites always?
The amount of data over the satellite link (64k) is an issue, so it is important to send only the changes.
Do you think I'm on track? Is it possible to implement.
Kjetil
|||Kjso,
as you need inserts updates and deletes to be replicated over the internet,
you are essentially coding what MS developers may well be creating in SQL
2005 :-). This is not straightforward at all (to say the least). You might
want to use triggers on each table which write an audit of changes to your
audit table(s). These audit table changes are then sent over the link to a
central server where they are applied to the central database. In my opinion
this is only going to be feasible at the central server if you have
partitioned data; dealing with conflicts will make this too difficult. Then
sending only the changes required back to the other subscribers is not
simple. You say that you want only changes, but a 'snapshot' would be more
straightforward.
HTH,
Paul Ibison
What is the best solution? Is it possible to have the publisher/subscriber to leave a file, and then let the e-mail system transfer the file and then later the subscriber/publisher can pick up the file and do the merge?
Any ideas?
Thanx, Kjetil
Kjetil,
this is not directly possible in the replication setup. You might want to
use DTS to import the file into a database then in the DTS package. Setup
replication between this staging database and your central system and you
can initiate replication locally.
HTH,
Paul Ibison
|||Thanks, this is what I tried to say to my boss. But how can this be implemented? Should I use merge replication or transactional replication.
In my home office, I will need a 'copy' of the ship database, in the first stage 9 ships.
Is it possible to block the database in the ships for updates while the replication process is running?
Or should I use transactional replication, will I then need a 'proxy database' at each ship and one for each ship in the home office. Is it possible to implement conflict resolution methods that guarantee the same result at all sites always?
The amount of data over the satellite link (64k) is an issue, so it is important to send only the changes.
Do you think I'm on track? Is it possible to implement.
Kjetil
|||Kjso,
as you need inserts updates and deletes to be replicated over the internet,
you are essentially coding what MS developers may well be creating in SQL
2005 :-). This is not straightforward at all (to say the least). You might
want to use triggers on each table which write an audit of changes to your
audit table(s). These audit table changes are then sent over the link to a
central server where they are applied to the central database. In my opinion
this is only going to be feasible at the central server if you have
partitioned data; dealing with conflicts will make this too difficult. Then
sending only the changes required back to the other subscribers is not
simple. You say that you want only changes, but a 'snapshot' would be more
straightforward.
HTH,
Paul Ibison
Subscribe to:
Posts (Atom)