replication with roaming laptops

[Peter King]

I am trying to figure out the best way to handle a system with lots of
roaming laptops wanting to take the data with them.

I have the following requirement:

Web server (possibly several) which holds main database and everyone
connects to (PHP web site with updates)
Lots of laptops (up to 50) take copies of the master data with them to work
on locally (Same PHP web site)
Laptops now connect back online after making local updates and want to
synchronise their data.

Now the last item is the hard bit.

Should I write some clever PHP code to do the synchronisation ?
Or should I try and use the MySQL synchronisation functions?

I have tried the MySQL 2 way synchronisation before but not to this level.

As you can guess this also involves having a web server and MySQL server on
the laptops (any pointers to better ways here, please advise).

Any help or personal experience much appreciated

Regards Pete

[Virgil Green]

Peter King wrote:

I am trying to figure out the best way to handle a system with lots of
roaming laptops wanting to take the data with them.

I have the following requirement:

Web server (possibly several) which holds main database and everyone
connects to (PHP web site with updates)
Lots of laptops (up to 50) take copies of the master data with them
to work on locally (Same PHP web site)
Laptops now connect back online after making local updates and want to
synchronise their data.

Now the last item is the hard bit.

Should I write some clever PHP code to do the synchronisation ?
Or should I try and use the MySQL synchronisation functions?

I have tried the MySQL 2 way synchronisation before but not to this
level.

As you can guess this also involves having a web server and MySQL
server on the laptops (any pointers to better ways here, please
advise).

Any help or personal experience much appreciated

Regards Pete

Are they all taking distinct parts of the Master data or do they all have a
full copy of the Master data? Synchronizing the data between 50 independent
databases when they are all working from the same original set of data will
be quite a challenge.

Better to have them "check out" the data they are taking with them and then
"check it in" when they return. I'd do this programatically .

- Virgil

[Peter King]

They will be taking all of the data with them but only updating small
amounts of it. It is a CRM system.
Problem is there are lots of dependant tables that require cascading updates
so increases the chance of errors. We have thought to use a timestamp on
updates and a table of checkouts of data, then do the work with the PHP. Not
sure if that is easier or not.

Regards Pete

-- "Virgil Green" <vj*@DESPAMobsy dian.com> wrote in message
news:Wd******** *******@newssvr 12.news.prodigy .com...

Peter King wrote:

I am trying to figure out the best way to handle a system with lots of
roaming laptops wanting to take the data with them.

I have the following requirement:

Web server (possibly several) which holds main database and everyone
connects to (PHP web site with updates)
Lots of laptops (up to 50) take copies of the master data with them
to work on locally (Same PHP web site)
Laptops now connect back online after making local updates and want to
synchronise their data.

Now the last item is the hard bit.

Should I write some clever PHP code to do the synchronisation ?
Or should I try and use the MySQL synchronisation functions?

I have tried the MySQL 2 way synchronisation before but not to this
level.

As you can guess this also involves having a web server and MySQL
server on the laptops (any pointers to better ways here, please
advise).

Any help or personal experience much appreciated

Regards Pete
Are they all taking distinct parts of the Master data or do they all have

a full copy of the Master data? Synchronizing the data between 50 independent databases when they are all working from the same original set of data will be quite a challenge.

Better to have them "check out" the data they are taking with them and then "check it in" when they return. I'd do this programatically .

- Virgil

[Rich R]

"Virgil Green" <vj*@DESPAMobsy dian.com> wrote in message
news:Wd******** *******@newssvr 12.news.prodigy .com...

Peter King wrote:

I am trying to figure out the best way to handle a system with lots of
roaming laptops wanting to take the data with them.

I have the following requirement:

Web server (possibly several) which holds main database and everyone
connects to (PHP web site with updates)
Lots of laptops (up to 50) take copies of the master data with them
to work on locally (Same PHP web site)
Laptops now connect back online after making local updates and want to
synchronise their data.

Now the last item is the hard bit.

Should I write some clever PHP code to do the synchronisation ?
Or should I try and use the MySQL synchronisation functions?

I have tried the MySQL 2 way synchronisation before but not to this
level.

As you can guess this also involves having a web server and MySQL
server on the laptops (any pointers to better ways here, please
advise).

Any help or personal experience much appreciated

Regards Pete
Are they all taking distinct parts of the Master data or do they all have

a full copy of the Master data? Synchronizing the data between 50 independent databases when they are all working from the same original set of data will be quite a challenge.

Better to have them "check out" the data they are taking with them and then "check it in" when they return. I'd do this programatically .

- Virgil

The 'easy way' of doing roaming laptops is to keep a small database on the
laptop. Everytime the laptop does an INSERT, DELETE, or UPDATE, capture the
actual SQL in a local table on the laptop. When the laptop connects to the
server, simply replay the SQL in the table. There are some details to this
scheme, but I have used it, and it works.

Regards,

Rich

[dlcarp]

Rich R wrote:

"Virgil Green" <vj*@DESPAMobsy dian.com> wrote in message
news:Wd******** *******@newssvr 12.news.prodigy .com...

Peter King wrote:

I am trying to figure out the best way to handle a system with lots of roaming laptops wanting to take the data with them.

I have the following requirement:

Web server (possibly several) which holds main database and everyone connects to (PHP web site with updates)
Lots of laptops (up to 50) take copies of the master data with them to work on locally (Same PHP web site)
Laptops now connect back online after making local updates and want to synchronise their data.

Now the last item is the hard bit.

Should I write some clever PHP code to do the synchronisation ?
Or should I try and use the MySQL synchronisation functions?

I have tried the MySQL 2 way synchronisation before but not to this level.

As you can guess this also involves having a web server and MySQL
server on the laptops (any pointers to better ways here, please
advise).

Any help or personal experience much appreciated

Regards Pete
Are they all taking distinct parts of the Master data or do they

all have a

full copy of the Master data? Synchronizing the data between 50 independent

databases when they are all working from the same original set of

data will

be quite a challenge.

Better to have them "check out" the data they are taking with them

and then

"check it in" when they return. I'd do this programatically .

- Virgil
The 'easy way' of doing roaming laptops is to keep a small database

on the laptop. Everytime the laptop does an INSERT, DELETE, or UPDATE, capture the actual SQL in a local table on the laptop. When the laptop connects to the server, simply replay the SQL in the table. There are some details to this scheme, but I have used it, and it works.

Regards,

Rich

[David C]

Rich,

We are designing an application and have a similar requirement to
Peter's. However, the user will only want to copy his information from
the central db to his laptop, make changes, and then update the central
db at some point. I like Virgil's 'check-in, check-out' scheme.

It would be nice if the MySQL version supporting 'triggers' was
available now to be able to capture the changes.

Would you consider sharing code examples of how you're accomplishing
what you are doing?

David

[Rich R]

"David C" <dl****@gmail.c om> wrote in message
news:11******** **************@ f14g2000cwb.goo glegroups.com.. .

Rich,

We are designing an application and have a similar requirement to
Peter's. However, the user will only want to copy his information from
the central db to his laptop, make changes, and then update the central
db at some point. I like Virgil's 'check-in, check-out' scheme.

It would be nice if the MySQL version supporting 'triggers' was
available now to be able to capture the changes.

Would you consider sharing code examples of how you're accomplishing
what you are doing?

David

Yes,David, I will put out some code tomorrow. Too much vino tonight. What do
you like about check-in,check-out? Also, even if you had triggers (update
and delete) it still would not work. In a nutshell, in all your inserts,
deletes, and updates, you need to keep a file (text or a table) of actual
statements. Just like the bin log does. And then you have to decide who wins
(just like many source code managers).

Regards,
Rich

[David C]

Thanks for the feedback Rich. Hope it was good vino!

The application is a job estimating app. The estimators want to be able
to take their quotes with them on their laptops when they leave town
and work on them while disconnected from the corporate server, then
dump them back when they get back in the office. I don't know whether
we'll allow them to update any quotes other than their own or not. I
thought the "check-in" "check-out" terminology would be good to use in
the system.
Will the code you're sending be attached here?

Thanks again,
David