Mushfique Khan
asked on
Database Sync
Environment: 50+ databases running on 11gR2 (mix of RAC & standalone), OS: Solaris 10
Task: Need to sync almost all of these dbs on a regular basis from PROD, according to different releases, in order for the developers to develop/test & verify their codes and then later QA can do regression & performance testing too.
Questions:
a. what are our best & fastest options to achieve this task, currently using data pump export/import, but would like to know, is there any better option/practice that is going on out there?
b. due to the size of PROD, have to do a lot of work manually to reduce these new replica db sizes, because don't want & not needed too, that much storage @dev & QA, obviously, do share if you haven any better/expert idea to achieve this task, in a better manner.
c. also got to encrypt some data, and this is also done manually, that's the reason these sync ups took a lot of time.
On my last assignment, we used to do syncing of our dbs using LVM Snapshot/Virt-Manager, but it was Linux and using virtual machines and all was file system snap/copy & replication process and it was very fast too.
Is there any thing like LVM Snapshot on Solaris OS ... any idea/guidance will highly be appreciated?
Please advice/guide, what are our best & appropriate options.
Thanks a lot in advance, in sharing your expert thoughts, suggestions and advices.
Task: Need to sync almost all of these dbs on a regular basis from PROD, according to different releases, in order for the developers to develop/test & verify their codes and then later QA can do regression & performance testing too.
Questions:
a. what are our best & fastest options to achieve this task, currently using data pump export/import, but would like to know, is there any better option/practice that is going on out there?
b. due to the size of PROD, have to do a lot of work manually to reduce these new replica db sizes, because don't want & not needed too, that much storage @dev & QA, obviously, do share if you haven any better/expert idea to achieve this task, in a better manner.
c. also got to encrypt some data, and this is also done manually, that's the reason these sync ups took a lot of time.
On my last assignment, we used to do syncing of our dbs using LVM Snapshot/Virt-Manager, but it was Linux and using virtual machines and all was file system snap/copy & replication process and it was very fast too.
Is there any thing like LVM Snapshot on Solaris OS ... any idea/guidance will highly be appreciated?
Please advice/guide, what are our best & appropriate options.
Thanks a lot in advance, in sharing your expert thoughts, suggestions and advices.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
![Avatar of Alex [***Alex140181***]](https://filedb.experts-exchange.com/files/public/2022/11/13/8257e912-ee3c-4a93-9df0-fe9a8a1515cf.jpeg){kind=small}
Take a look at this product: http://www.dbvisit.com/products/dbvisit_replicate_real_time_oracle_database_replication/
We've tested the trial recently and it's far less complicated than i.e. GoldenGate. IMHO, it is very powerful (and much cheaper)...
Kind regards,
Alex
We've tested the trial recently and it's far less complicated than i.e. GoldenGate. IMHO, it is very powerful (and much cheaper)...
Kind regards,
Alex
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
It was a fairly involved process to begin with and took a fair bit of scripting. But then we produced a series of "known outcomes" as part of a scripted test process and user testing encompassing "using it to work", "breaking it", "robustness", "integrity", "historical relationships" etc etc...
Now because that also reflected the live environment, we ended up using it for DEV by being able to copy across at any stage. The "master" for QA was pretty much locked and only available to a few.
If it passed QA, then we would go into user acceptance testing which was a group of individuals given the added responsibility for their departments to get involved in testing on a backup of live, then upgraded according to release, and used the same way as if it were live. The inputs were retrospective samples over the passed weeks / month end/ periodic demands and compared to the outcomes in live (and by the QA team back to the expected / known outcomes).
But it isn't for the feint hearted.
It did take quite a bit of work in the early stages, and did require quite some thought and commitment to make it happen and keeping it current. Pretty much everything was scripted so if we ever need to we could do the unthinkable and start from scratch again. But the main reason is because of the payroll subsystem and we had to mask reality in such a way that we can still test a-la real life without exposing highly sensitive data.
Advantages were huge in terms or productivity amongst staff (dev, QA and user) and always a relatively easy task to recover / emulate any environment, or to seed the dev environment. The release processes into live were always dependable and only a rare exception if something went bump (normally a people problem more than a process problem).
Might sound extreme, but it was very efficient in the end.