Replication slowing down database performance

Yuvapradeep

New Member
Hi All,

I am running progress 11.6.4 on Linux server.

When the replication is enabled the batch jobs are running very slow, and facing application degradation and after disabling the replication the jobs run faster and jobs get finished sooner.

1) Progress version, Service Pack and product
- OpenEdge Release 11.6.4 as of Tue Oct 3 01:15:46 EDT 2017
2) OS and version
- Linux rc51e00vm018b 4.1.12-112.16.7.el7uek.x86_64 #2 SMP Fri Apr 6 14:23:45 PDT 2018 x86_64 x86_64 x86_64 GNU/Linux

  • activity gets better when we stop the replication, as we tried this on last activity
  • There are huge number of reads and writes during activity and As per Progress knowledgebase increasing the source database startup parameter Database Service Manager - IPC Queue Size (-pica) might help. Please suggest.
  • We have increased CPU cores at server level to improve the performance, it did not help last activity.
  • I have attached the parameter list which is recommended by Progress and the parameter on our database

Can you please explain what can be the issue here and what parameters needs to be modified to resolve issue?


Thanks,
 

Attachments

  • Progress_recommend_parameters.xlsx
    10.4 KB · Views: 3

TomBascom

Curmudgeon
It is extremely important that both the source and the target have equal hardware resources, a very high quality network between them and equal startup parameter settings and configuration.

Replicating to an under-powered or misconfigured target can be painfully slow.

The attached parameters don't say anything about your target and are fairly pedestrian. The one that stands out is that -pica is apparently not being set on the source and is therefore at its default value (64) which is inadequate for everyone. You should definitely increase it as suggested.

Beyond that there may be many opportunities to improve performance but that would require a more comprehensive review of your system that should also include information about the workload and what specific activities the users feel are "slow". As well as all of the current configuration settings and the hardware environment.
 
Top