How to calculate the optimum -pica setting for OpenEdge Replication
Can anybody explain what the knowledgebase article is trying to say?
What does "the optimum -pica" mean? Is it for situation when a network between the source and target databases is broken? Then the answer is simple: the higher -pica the better. But instead of "optimum value" I would call it a "memory reserve" that would give the time to fix the broken network.
apply the following formula: -pica = (TAIW / BlockCount) * 1.25 Where: TAIW – Total AI Writes BlockCount – Block count for OpenEdge Version (Version 9 = 18.2, Release 10 – 9.16)
AI Writes are in the aiblocksize unit while the -pica is in KB but the formula does not use the -aiblocksize. How it's possible?
What is the BlockCount? What kind of blocks? And why their "count" is a decimal value?
When the article was originally written it was aimed not at the failed network situation but at helping the user appropriately size the pica queue (memory size).
This was back when the -pica was significantly smaller than the current limit of 1000000.
The article is more appropriate for smaller, lower volume sites which might reasonably calculate the general volume of total AI writes in a given period of time.
For bigger shops (or sites which have the fear of a network outage) it would be better to pick a higher value or the maximum value of -pica 1000000 (which will equate to about 993 megs of RAM) if the site has sufficient free RAM to accommodate the max pica setting.
What are the mechanisms behind the the formula?
Why the constant (BlockCount) is almost twice smaller in V10/11?
It's just a curiosity.