Prev: Single column indexes supporting Foreign keys
Next: 341594 Information portals for computer and Web technology 23
From: Jason K on 22 Mar 2010 14:40
We are in the process of migrating data from db2 9.5 from Solaris
(SPARC) to Linux (x64). A big to little endian conversion has to take
place because of the architectural differences between to two
platforms. What steps can I run to verify that the data in the source
and target databases are the same. Yes, I know we should be able to
trust the export/import, but to just to be safe, how can I compare the
data to make sure it was migrated correctly?

Thanks in advance for any advice.

- Jason
From: Jason K on 23 Mar 2010 11:47
On Mar 23, 1:28 am, Ian <ianb...(a)mobileaudio.com> wrote:
> On 3/22/10 11:40 AM, Jason K wrote:
>
> > We are in the process of migrating data from db2 9.5 from Solaris
> > (SPARC) to Linux (x64). A big to little endian conversion has to take
> > place because of the architectural differences between to two
> > platforms. What steps can I run to verify that the data in the source
> > and target databases are the same. Yes, I know we should be able to
> > trust the export/import, but to just to be safe, how can I compare the
> > data to make sure it was migrated correctly?
>
> I think that you can probably do spot checks, i.e. sum(column),
> count(column), etc as basic checks.
>
> To be more precise, though, I think that you would have to have
> sorted data on your export (from SPARC system).  Then, after loading
> the data on the Linux box, you could perform the same (sorted)
> exports and then use diff or perhaps even md5sum on the the file(s)
> to make sure you get the same data out of both systems.
>
> I would suggest you validate this on a single table, though.
> Proving that DB2 loads data properly on a single table should be
> proof enough.

Thanks a lot for the suggestion, I will give that a try...
 | 
Pages: 1
Prev: Single column indexes supporting Foreign keys
Next: 341594 Information portals for computer and Web technology 23