Backup from RSS?

Is there still, really and truly, no way to take a backup from an RSS, without breaking (and then having to re-initialize) the pair? This is such a major handicap to us. It is unfathomable to us why (assuming it is still the case-- and I would love to be wrong) IBM does not permit backups from a secondary server. Using dbexport from the RSS, and then dbimport on another server works, but takes over 8 hours, where as ontape backup and restore takes less than 20 minutes. I note that there are at least two RFEs for this capability. But, they are several years old, and IBM has not chosen to implement. DG

Original post: Is there still, really and truly, no way to take a backup from an RSS, without breaking (and then having to re-initialize) the pair? This is such a major handicap to us. It is unfathomable to us why (assuming it is still the case-- and I would love to be wrong) IBM does not permit backups from a secondary server. Using dbexport from the RSS, and then dbimport on another server works, but takes over 8 hours, where as ontape backup and restore takes less than 20 minutes. I note that there are at least two RFEs for this capability. But, they are several years old, and IBM has not chosen to implement. DG Response: Does external backup's count? Check this link https://www.ibm.com/support/knowledgecenter/en/SSGU8G_12.1.0/com.ibm.bar.doc/ids _bar_503.htm Jacques Renaut IBM Informix Advanced Support

When people take backups they want to be 100% safe and hate to hear that, "oops" that backup only works in specific conditions is awful to hear because the condition were not correct and you lost data during a restore. RSS only contains up to date logged data. If you have a non-logged database or non-logged table you can miss this data. While you might read the manual and this condition is Ok in your situation, it will not work for all and believe it or not several of those will not read the manual and end up with a backups from RSS containing less data than they desire. John F. Miller III miller3@us.ibm.com 503-747-1366 ids-bounces@iiug.org wrote on 04/11/2017 10:25:09 AM: > From: "DAVID GROVE" <david.grove@alaska.gov> > To: ids@iiug.org > Date: 04/11/2017 10:27 AM > Subject: Backup from RSS? [38913] > Sent by: ids-bounces@iiug.org > > Is there still, really and truly, no way to take a backup from an > RSS, without > breaking (and then having to re-initialize) the pair? > > This is such a major handicap to us. It is unfathomable to us why > (assuming it > is still the case-- and I would love to be wrong) IBM does not permit backups > from a secondary server. > > Using dbexport from the RSS, and then dbimport on another server works, but > takes over 8 hours, where as ontape backup and restore takes less than 20 > minutes. > > I note that there are at least two RFEs for this capability. But, they are > several years old, and IBM has not chosen to implement. > > DG > > > ***************************************************************************= **** > Forum Note: Use "Reply" to post a response in the discussion forum. >

I already do this here , few times at AIX using JFS2 snapshots.. For tests only...and works very well. At linux, you can do this using LVM snapshot, either if your chunks are raw devices... All my chunks are into JFS2 file system and I have free space to create JFS2 external snaphots... very short and simple steps... * stop the RSS * create the snapshot * restart the instance and change it to standard. (block any user access if have) * execute the backup (I'm using ontape) * stop instance * restore the snapshot * restart the instance and wait for resync ... 2017-04-11 14:25 GMT-03:00 DAVID GROVE <david.grove@alaska.gov>: > Is there still, really and truly, no way to take a backup from an RSS, > without > breaking (and then having to re-initialize) the pair? > > This is such a major handicap to us. It is unfathomable to us why > (assuming it > is still the case-- and I would love to be wrong) IBM does not permit > backups > from a secondary server. > > Using dbexport from the RSS, and then dbimport on another server works, but > takes over 8 hours, where as ontape backup and restore takes less than 20 > minutes. > > I note that there are at least two RFEs for this capability. But, they are > several years old, and IBM has not chosen to implement. > > DG > > > ************************************************************ > ******************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --94eb2c061accbedc58054cea0992

Over the years (maybe it comes with the age), I've become more "critic" of the options made... But I always try (this is arguable of course) to base my opinions on technical details. And there was a very good reason why a backup on the secondary wouldn't work. I don't mean what John mentioned (as it happened many times I disagree with John), because essentially non logged objects (databases, slobs or tables) are simply not supported at all in a cluster.... so the fact they wouldn't be on a backup would be a minor issue in my opinion). What I mean was that a backup would require writing to the database (we update the timestamps on "very old pages" - which triggered a very famous bug sequence several years ago - ). Apart from the time stamp we also update the backup information in the dbspaces / rootdbs. But relatively recently something changed... we introduced a feature for helping on storage manager's de-duplication technologies. Basically depending on a variable, we don't update the timestamps anymore. The drawback is that we can only do L0 backups. Now... my point is that with this in mind, the only change we do is the last backup information.... and that sounds like something a secondary could pass to the primary for updating... or eventually change in it's own pages (or both). Again, it would have it's limitations, but it would be much better than what we do now... and from my ignorance, sounds feasible.... Regards On Tue, Apr 11, 2017 at 6:25 PM, DAVID GROVE <david.grove@alaska.gov> wrote: > Is there still, really and truly, no way to take a backup from an RSS, > without > breaking (and then having to re-initialize) the pair? > > This is such a major handicap to us. It is unfathomable to us why > (assuming it > is still the case-- and I would love to be wrong) IBM does not permit > backups > from a secondary server. > > Using dbexport from the RSS, and then dbimport on another server works, but > takes over 8 hours, where as ontape backup and restore takes less than 20 > minutes. > > I note that there are at least two RFEs for this capability. But, they are > several years old, and IBM has not chosen to implement. > > DG > > > ************************************************************ > ******************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > -- Fernando Nunes Portugal http://informix-technology.blogspot.com My email works... but I don't check it frequently... --001a1144d1ba5b5230054cead03f