Hi Robert,
I think, I have fair understanding of atomic batches. By synchronously, I meant
client making a blocking atomic batch call to execute all table updates in one
go. I understand that Cassandra statements will be executed as per CL.
I want to know how people generally handle a scenario when there is just 1
transaction table but there is overhead of updating multiple manually created
index tables and reeporting tables.
Do they go with atomic batches which have some performance cost
OR
They just update the transaction table and the responsibility of maintaining
consistency of transaction table with index and reporting table lies with
clients and txn table updates are not blocking till other index/reporting
tables are updated?
OR
There are better ways to deal with the scenario where u have one txn table,3
index tables and 2 reporting tables?
I think atomic batches takes the headache of atomicity to the server and thus
data consistency can be maintained. Bothered about the cost and any
negatives,especially when 5 writes are required for every transaction table
write.
Thanks
Anuj
Sent from Yahoo Mail on Android
From:Robert Wille rwi...@fold3.com
Date:Fri, 24 Jul, 2015 at 12:03 am
Subject:Re: Best Practise for Updating Index and Reporting Tables
My guess is that you don’t understand what an atomic batch is, give that you
used the phrase “updated synchronously”. Atomic batches do not provide
isolation, and do not guarantee immediate consistency. The only thing an atomic
batch guarantees is that all of the statements in the batch will eventually be
executed. Both approaches are eventually consistent, so you have to deal with
inconsistency either way.
On Jul 23, 2015, at 11:46 AM, Anuj Wadehra anujw_2...@yahoo.co.in wrote:
We have a transaction table,3 manually created index tables and few tables for
reporting.
One option is to go for atomic batch mutations so that for each transaction
every index table and other reporting tables are updated synchronously.
Other option is to update other tables async, there may be consistency issues
if some mutations drop under load or node goes down. Logic for rolling back or
retrying idempodent updates will be at client.
We dont have a persistent queue in the system yet and even if we introduce one
so that transaction table is updated and other updates are done async via
queue, we are bothered about its throughput as we go for around 1000 tps in
large clusters. We value consistency but small delay in updating index and
reporting table is acceptable.
Which design seems more appropriate?
Thanks
Anuj
Sent from Yahoo Mail on Android