Replication and HA into Database

Hi all. I need to know how many of you use CDR/HDR and in your own words what's your opinion? Also what is the recommendation for a Cluster, it should be configured by Software or by Informix? What version of Informix do i have to use ? What is the behavior of Informix into a HA architecture? I would like to hear from you personal experiences over technical references. Thanks in advance Saludos... Mario Meneses Database Administrator mailto:menesem@gedas.com.mx > ---------- > De: Pons Roca, Isidre[SMTP:ipons@dtgna.altanet.org] > Enviado el: Lunes 25 de Octubre de 1999 2:10 AM > Para: informix-list@iiug.org > Asunto: Very slow index generation. > > > Hi all. > > This weekend I have reorganized my database. > > I have changed slow disks (SCSII-2 of 2Gb) for new disks (FW- > SCSII of 8.5 Gb) > My instance of Informix has a size, of data, of around some 14Gb. > > To be able to make these operations I created an identical instance > in another machine, I reorganized the new instance and pass over > the data from it schemes it of backup toward the new one. With the > new configuration I have a dbspace for rootdbs, one for the > logicalslogs, another for the physicallog, those of data (4) and one > for the catalogs of the databases (create database in...). The load > of data has been carried out with the database in NON LOG. > > Until here all well. The problem arises in the moment to regenerate > the indexes (this operation is already carrying out it with the data > loaded). The time of regeneration of the indexes is enormous > almost consuming all the resources of the CPUs of the system. > > To regenerate the indexes I have followed the advice of Art > ,PSORT_NPROCS=6, PSORT_DBTEMP=<3 filesystems> and the > PDQPRIORITY=50, but I have not noticed any yield improvement. > > To generate the primary keys and the indexes I have used two > scripts with the following syntax: > > database perico << EOF > alter table table_a add contraint primary key (col_1,.. col_n)> constraint pk_table_a; > create index i1_table_a (col_1,.. col_n);> . > . > EOF > > > Working environment: > Platform: Sun SS1000E with 4 cpu's and 524288K mem > OS, Solaris 2.4 with last ones patches. > Informix: IDS 7.24.UC7-1 > Onconfig: > # System Configuration > > SERVERNUM 0 > DBSERVERNAME perico > DBSERVERALIASES perico0,perico1 > DEADLOCK_TIMEOUT 60 > RESIDENT 1 > MULTIPROCESSOR 1 > NUMCPUVPS 3 > SINGLE_CPU_VP 0 > NOAGE 1 > AFF_SPROC 1 > AFF_NPROCS 3> > > And now the question: > Because the generation of indexes is so slow? > > Thanks in advance. > > --------------------------------------- > Isidre PONS ROCA > BASE - Gesti' d'Ingressos Locals > (Diputacio de Tarragona) > Servei de Sistemes de Informacio > Av President Lluis Companys 12-C > 43005 - Tarragona > SPAIN > Tel # +34 977 236731 > Fax # +34 977 227302 > http://www.altanet.org > ipons@dtgna.altanet.org > --------------------------------------- >

MENESES HERNANDEZ, MARIO ALFREDO <MENESEM@gedas.com.mx> wrote in message news:7v1al6$16m$1@news.xmission.com... > > Hi all. Hi Mario ! > > I need to know how many of you use CDR/HDR and in your own words > what's your opinion? I've used HDR replication for 2,5 years - we have no problems and complains. Our environment was SCO Unix 5.02 & IDS 7.22UC2. > Also what is the recommendation for a Cluster, it should be configured > by Software or by Informix? Depends on cluster. Some days ago I saw deep inside the 'oninit.exe' called functions - some of them supports Microsoft Cluster commands. In Unix environment you must properly tune switchover scripts. One of our systems has replication (HDR) in cluster - users works with first node, second one used for reports. >What version of Informix do i have to use ? It depends on your hardware and operating system. --------------------------------------------------- With best regards, Yuri Dovgart SAP R/3, Informix technical consultant, Senior System Consultant System Architecture and High Availability Systems, 'Telecominvest' company Email y_dovgart@tci.ukrtel.net ICQ 39284285 > What is the behavior of Informix into a HA architecture? > I would like to hear from you personal experiences over technical > references. > > Thanks in advance > > Saludos... > Mario Meneses > Database Administrator > mailto:menesem@gedas.com.mx > > > ---------- > > De: Pons Roca, Isidre[SMTP:ipons@dtgna.altanet.org] > > Enviado el: Lunes 25 de Octubre de 1999 2:10 AM > > Para: informix-list@iiug.org > > Asunto: Very slow index generation. > > > > > > Hi all. > > > > This weekend I have reorganized my database. > > > > I have changed slow disks (SCSII-2 of 2Gb) for new disks (FW- > > SCSII of 8.5 Gb) > > My instance of Informix has a size, of data, of around some 14Gb. > > > > To be able to make these operations I created an identical instance > > in another machine, I reorganized the new instance and pass over > > the data from it schemes it of backup toward the new one. With the > > new configuration I have a dbspace for rootdbs, one for the > > logicalslogs, another for the physicallog, those of data (4) and one > > for the catalogs of the databases (create database in...). The load > > of data has been carried out with the database in NON LOG. > > > > Until here all well. The problem arises in the moment to regenerate > > the indexes (this operation is already carrying out it with the data > > loaded). The time of regeneration of the indexes is enormous > > almost consuming all the resources of the CPUs of the system. > > > > To regenerate the indexes I have followed the advice of Art > > ,PSORT_NPROCS=6, PSORT_DBTEMP=<3 filesystems> and the > > PDQPRIORITY=50, but I have not noticed any yield improvement. > > > > To generate the primary keys and the indexes I have used two > > scripts with the following syntax: > > > > database perico << EOF > > alter table table_a add contraint primary key (col_1,.. col_n)> > constraint pk_table_a; > > create index i1_table_a (col_1,.. col_n);> > . > > . > > EOF > > > > > > Working environment: > > Platform: Sun SS1000E with 4 cpu's and 524288K mem > > OS, Solaris 2.4 with last ones patches. > > Informix: IDS 7.24.UC7-1 > > Onconfig: > > # System Configuration > > > > SERVERNUM 0 > > DBSERVERNAME perico > > DBSERVERALIASES perico0,perico1 > > DEADLOCK_TIMEOUT 60 > > RESIDENT 1 > > MULTIPROCESSOR 1 > > NUMCPUVPS 3 > > SINGLE_CPU_VP 0 > > NOAGE 1 > > AFF_SPROC 1 > > AFF_NPROCS 3> > > > > > And now the question: > > Because the generation of indexes is so slow? > > > > Thanks in advance. > > > > --------------------------------------- > > Isidre PONS ROCA > > BASE - Gesti' d'Ingressos Locals > > (Diputacio de Tarragona) > > Servei de Sistemes de Informacio > > Av President Lluis Companys 12-C > > 43005 - Tarragona > > SPAIN > > Tel # +34 977 236731 > > Fax # +34 977 227302 > > http://www.altanet.org > > ipons@dtgna.altanet.org > > --------------------------------------- > >