DDRBLOCK in IDS 9.4

Hi everyone. I have a chronic problem with an IDS 9.4 FCX21 instance that occassional hangs in a CHKPOINT needed state for unknown reasons. Informix says we need to upgrade because the version we are using has a bug that is causing this situation. The instance uses One-Way replication (it is the primary server in a PRIMARY-REPLICATE pair). There have been warnings in the online.log since the instance was created 5 years ago that because CDR is implemented we may need to add additional logical logs to avoid a log wrap-around condition or DDRBLOCK situation. My question is: Could this be the cause of the chkpoint needed state that hangs. My understanding was that if you encountered a log-wraparound or DDRBLOCK that all that would be adversely affected would be the REPLICATE server (IE: the server that RECEIVES the replicated table updates) not the PRIMARY server. In other words what could happen is the PRIMARY server MIGHT wrap-around with the logical logs and overwrite changes that HAVE NOT YET BEEN SENT TO THE REPLICATE SERVER. But that the PRIMARY server would still continue NORMALLY. Is this true? Or is it possible the warnings about DDRBLOCK on the PRIMARY server might be the cause of the CHKPOINT HANG STATE ON THE PRIMARY SERVER? Thanks

no - DDRBLOCK is going to be impacted on the source (i.e. primary). ------------------------------------- Madison Pruet, STSM IDS Replication Architect = "WILL LANDSTROM" = <willlandstrom@ya = hoo.com> = To Sent by: ids@iiug.org = ids-bounces@iiug. = cc org = Subj= ect DDRBLOCK in IDS 9.4 [13607] = 10/06/2008 06:53 = PM = = = Please respond to = ids@iiug.org = = = Hi everyone. I have a chronic problem with an IDS 9.4 FCX21 instance th= at occassional hangs in a CHKPOINT needed state for unknown reasons. Infor= mix says we need to upgrade because the version we are using has a bug that= is causing this situation. The instance uses One-Way replication (it is th= e primary server in a PRIMARY-REPLICATE pair). There have been warnings i= n the online.log since the instance was created 5 years ago that because CDR = is implemented we may need to add additional logical logs to avoid a log wrap-around condition or DDRBLOCK situation. My question is: Could this= be the cause of the chkpoint needed state that hangs. My understanding was tha= t if you encountered a log-wraparound or DDRBLOCK that all that would be adversely affected would be the REPLICATE server (IE: the server that RECEIVES th= e replicated table updates) not the PRIMARY server. In other words what c= ould happen is the PRIMARY server MIGHT wrap-around with the logical logs an= d overwrite changes that HAVE NOT YET BEEN SENT TO THE REPLICATE SERVER. = But that the PRIMARY server would still continue NORMALLY. Is this true? Or= is it possible the warnings about DDRBLOCK on the PRIMARY server might be the= cause of the CHKPOINT HANG STATE ON THE PRIMARY SERVER? Thanks ***********************************************************************= ******** Forum Note: Use "Reply" to post a response in the discussion forum. =