Re: renaming a chunk
Posted in 2007
Topics: Storage & Space Management, Connectivity: ESQL/C, 4GL & Embedded SQL, Server Administration, Platform-Specific Issues
Thanks to both for your response.
I have started to test this on a test instance but will abandon that now. It makes sense I guess, that just like a new filesystem being created, when I create the new chunk and give it the character device for a path ( through the link ), that it may wipe out part of the data while initializing. I would have hoped that the intitialization would use the exact same part of the device that the first initialiaztion did, and therefore, the data would remain in tact.
Thanks again.
floyd
-----Original Message-----
From: Marco Greco [mailto:marco@4glworks.com]
Sent: Monday, June 4, 2007 10:28 AM
To: floyd@fwellers.com
Cc: informix-list@iiug.org
Subject: Re: renaming a chunk
floyd@fwellers.com wrote:
> Hi,
> We are using IDS10.0FC5
> We would like to rename a chunk to satisfy a link checking script that
> we have running on startup.
> It seems that the recommended ( apparently only available ) way is to
> do a rename during restore. That is not an option for us.
>
> I was wondering if it would work to :
>
> Bring the database to single user mode
> drop the chunk with onspaces,
> drop the link that points it to the device
> recreate the link, pointing to the exact same device, but with the new
> chunk name,
> add the chunk with onspaces.
>
> I'm assuming that since we're pointing to the same raw device, the data
> will still be intact and the database won't care.
You assume incorrectly - the data will be there and the engine will duly zap
parts of it upon adding it back.
Also, it's unlikely (there are caveats) that you will be able to drop the
chunk while it is not empty, and this is not necessarily a bad thing. (you
should also have enough hints on how to modify your strategy)
>
> Any thoughts ?
>
> Thanks,
> Floyd
>
--
Ciao,
Marco
______________________________________________________________________________
Marco Greco /UK /IBM Standard disclaimers apply!
Structured Query Scripting Language http://www.4glworks.com/sqsl.htm
4glworks http://www.4glworks.com
Informix on Linux http://www.4glworks.com/ifmxlinux.htm
Have Marco and Keith answered your question accurately? I'm not sure.
You won't be able to "drop the chunk with onspaces" unless the chunk is
empty. In which case, of course you may re-add it with a new name and
without consequence, as it will be empty when you drop it and will be empty
when you re-add it!
--
Neil Truby t:01932 724027
Director m:07798 811708
Ardenta Limited e:neil.truby@ardenta.com
<floyd@fwellers.com> wrote in message
news:mailman.102.1180967551.13675.informix-list@iiug.org...
Thanks to both for your response.
I have started to test this on a test instance but will abandon that now. It
makes sense I guess, that just like a new filesystem being created, when I
create the new chunk and give it the character device for a path ( through
the link ), that it may wipe out part of the data while initializing. I
would have hoped that the intitialization would use the exact same part of
the device that the first initialiaztion did, and therefore, the data would
remain in tact.
Thanks again.
floyd
-----Original Message-----
From: Marco Greco [mailto:marco@4glworks.com]
Sent: Monday, June 4, 2007 10:28 AM
To: floyd@fwellers.com
Cc: informix-list@iiug.org
Subject: Re: renaming a chunk
floyd@fwellers.com wrote:
> Hi,
> We are using IDS10.0FC5
> We would like to rename a chunk to satisfy a link checking script that
> we have running on startup.
> It seems that the recommended ( apparently only available ) way is to
> do a rename during restore. That is not an option for us.
>
> I was wondering if it would work to :
>
> Bring the database to single user mode
> drop the chunk with onspaces,
> drop the link that points it to the device
> recreate the link, pointing to the exact same device, but with the new
> chunk name,
> add the chunk with onspaces.
>
> I'm assuming that since we're pointing to the same raw device, the data
> will still be intact and the database won't care.
You assume incorrectly - the data will be there and the engine will duly zap
parts of it upon adding it back.
Also, it's unlikely (there are caveats) that you will be able to drop the
chunk while it is not empty, and this is not necessarily a bad thing. (you
should also have enough hints on how to modify your strategy)
>
> Any thoughts ?
>
> Thanks,
> Floyd
>
--
Ciao,
Marco
______________________________________________________________________________
Marco Greco /UK /IBM Standard disclaimers apply!
Structured Query Scripting Language http://www.4glworks.com/sqsl.htm
4glworks http://www.4glworks.com
Informix on Linux http://www.4glworks.com/ifmxlinux.htm