On Sat, Sep 14, 2013 at 2:32 PM, Moshe Jacobson wrote:
> How do I migrate my 9.1 directory to a new file system with the least
> downtime possible?
>
Create a new tablespace on a the new filesystem and move everything over.
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
On Sat, Sep 14, 2013 at 6:32 AM, Moshe Jacobson wrote:
> How do I migrate my 9.1 directory to a new file system with the least
> downtime possible?
>
> I don't know if this makes any difference, but my pg_xlog directory is on
> its own volume as well, so I would have to unmount it and remount it a
On Sat, Sep 14, 2013 at 11:26 AM, CS DBA wrote:
> 3) before the swap do another rsync to bring the db as close to sync'ed as
> possible
>
> 4) shut down the primary db
>
> 5) do a final rsync (should have little to nothing to sync)
>
> 6) start the secondary db
>
Thanks. That's how I normally do
We've done this across servers, maybe it would work for you:
1) rsync the entire db to the second file system
2) do another rsync each day until you are ready to swap (of course only
changed files will be moved)
3) before the swap do another rsync to bring the db as close to sync'ed
as possi
On Sat, Sep 14, 2013 at 9:40 AM, Michael Nolan wrote:
> Have you considered setting up a synchronously replicated slave database
> on the new file system (using a port other than 5432), getting it in sync,
> then shutting both databases down (master first), switching the slave over
> to become th
Have you considered setting up a synchronously replicated slave database on
the new file system (using a port other than 5432), getting it in sync,
then shutting both databases down (master first), switching the slave over
to become the master and restarting just that database on port 5432?
--
Mike
How do I migrate my 9.1 directory to a new file system with the least
downtime possible?
I don't know if this makes any difference, but my pg_xlog directory is on
its own volume as well, so I would have to unmount it and remount it as
well, but I would not have to copy over my xlogs.
I figure the