Hello!
1) Roughly yes: you need to do checkpoint, RAM does not need wal/ but only
db/, and also you need checkpoint page buffer with persistence which you
can add to memory region when going pure-RAM.
2) I don't think that per cache metrics exist.
Regards,
--
Ilya Kasnacheev
пн, 25 февр.
Yes of course.
All of my data is imported through sql. I have it in the warehouse, and then
just inserting them to ignite.
I was trying through SQL query get size in bytes for single row and then
based on that trying to predict resources needed. That's why I was confused
seeing so much more
Hello!
Did you look at https://apacheignite.readme.io/docs/capacity-planning ?
Regards,
--
Ilya Kasnacheev
пн, 25 февр. 2019 г. в 13:54, newigniter :
> Tnx ilya!
>
> Never said it is significant. Not all of my caches will be so small in
> size.
> This was only an example. I was just trying
Tnx ilya!
Never said it is significant. Not all of my caches will be so small in size.
This was only an example. I was just trying to predict (and understand) how
much resources will I need for all of my data I am planning to import.
Regards
--
Sent from:
Hello!
First of all, Ignite is a data grid designed for gigabytes of data. So 71M
is not significant.
Secondly, your cache probably has 1024 partitions configured by default,
and each of those has quite some metadata and directory structure. Try
decreasing that number if you plan on having very
created a table and inserted 9013 records in it.
The table has 17 columns and the total size of the data before import into
ignite is 282,625 bytes.
I have native persistence enabled and have 2 nodes in my cluster. Data is
replicated to both nodes.
I am using ignite 2.7.
After import into
I created a table and inserted 9013 records in it.
The table has 17 columns and the total size of the data before import into
ignite is 282,625 bytes.
I have native persistence enabled and have 2 nodes in my cluster. Data is
replicated to both nodes.
I am using ignite 2.7.
After import into