Re: IDS Performance Problem
Peter Diaz de Leon asked whether storage backups run on the secondary in HDR on IDS 10.00.FC5. Art Kagel and Madison Pruet said backups must be taken on the primary, as the backup writes a checkpoint, dummy page updates and reserved pages. Kagel agreed it is prudent to configure the storage manager on the secondary too. Answered.
Auto-generated by Claude from the posts below — may be imperfect; read the full thread.
I am going to be setting up replication (HADR) between to Linux boxes running RH Enterprise 4.0 and IDS 10.00.FC5. Do storage backups have to run on the slave server when in slave mode? Or only when in master mode? Thanks Peter
Only on the primary. Art S. Kagel ----- Original Message ----- From: Peter J Diaz de Leon <ids@iiug.org> At: 2/08 15:45:56 I am going to be setting up replication (HADR) between to Linux boxes running RH Enterprise 4.0 and IDS 10.00.FC5. Do storage backups have to run on the slave server when in slave mode? Or only when in master mode? Thanks Peter ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum.
Strange as it might seem, there are several places during the backup wh= ere we write to disk. 1) The backup Checkpoint 2) Very old pages have a dummy update performed on them 3) Reserved pages are updated to reflect that the backup has been taken= . So this means that the backup needs to be taken on the primary node. = "Peter J Diaz de = Leon" = <pdiazdeleon@infi = To nityhealthcare.co ids@iiug.org = m> = cc Sent by: = ids-bounces@iiug. Subj= ect org Re: IDS Performance Problem [83= 85] = = 02/08/2007 02:45 = PM = = = Please respond to = ids@iiug.org = = = I am going to be setting up replication (HADR) between to Linux boxes running RH Enterprise 4.0 and IDS 10.00.FC5. Do storage backups have to run on the slave server when in slave mode? Or only when in master mode? Thanks Peter ***********************************************************************= ******** Forum Note: Use "Reply" to post a response in the discussion forum. =
So during normal conditions backups are done on the primary node, even though logs are being sent to the secondary node. I am assume that a storage manager (lets say ISM) till must be configured and operational on the slave in the event it becomes the primary node? Yes? Thanks Peter Madison Pruet wrote: > Strange as it might seem, there are several places during the backup where > we write to disk. > > 1) The backup Checkpoint > 2) Very old pages have a dummy update performed on them > 3) Reserved pages are updated to reflect that the backup has been taken > .. > > So this means that the backup needs to be taken on the primary node. > >
Yes, that would be prudent. Art S. Kagel ----- Original Message ----- From: Peter J Diaz de Leon <ids@iiug.org> At: 2/09 7:54:09 So during normal conditions backups are done on the primary node, even though logs are being sent to the secondary node. I am assume that a storage manager (lets say ISM) till must be configured and operational on the slave in the event it becomes the primary node? Yes? Thanks Peter Madison Pruet wrote: > Strange as it might seem, there are several places during the backup where > we write to disk. > > 1) The backup Checkpoint > 2) Very old pages have a dummy update performed on them > 3) Reserved pages are updated to reflect that the backup has been taken > .. > > So this means that the backup needs to be taken on the primary node. > > ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum.