|
|
| Author |
Message |
< Advanced Erlang/OTP ~ Name Registrations and Mnesia replication limits |
| Noki4475 |
 Posted: Sun Aug 09, 2009 1:57 pm |
 |
|
|
Joined: 09 Aug 2009
Posts: 3
|
Hi,
I'm new to erlang and i just would like to know if someone would see some limitation on the usage of Name registration (global) and Database replication with Mnesia : Can i use them even in a VERY large erlang cluster (many hundreds or more nodes) ?
Thks in advance for your answers.
Noki. |
|
|
| Back to top |
 
|
| Mazen |
| Posted: Mon Aug 10, 2009 6:01 am |
|
|
|
User
Joined: 20 Jul 2006
Posts: 164
Location: London
|
No, I would never recommend using mnesia in a cluster with more then perhaps 7 nodes (depending on the amount of data you are replicating and the amount of updates on this data).
If you have high load on your system then I wouldn't recommend more then 3. Things will start getting really ugly if you use transactions as well.
Note: The numbers 7 and 3 are completely subjective.
/Mazen |
|
|
| Back to top |
|
| uwiger |
| Posted: Tue Aug 11, 2009 11:54 am |
|
|
|
User
Joined: 03 Jul 2006
Posts: 604
Location: Sweden
|
Mazen wrote: No, I would never recommend using mnesia in a cluster with more then perhaps 7 nodes (depending on the amount of data you are replicating and the amount of updates on this data).
If you have high load on your system then I wouldn't recommend more then 3. Things will start getting really ugly if you use transactions as well.
Note: The numbers 7 and 3 are completely subjective.
/Mazen
There is probably some general rule of thumb that planning database clusters is difficult.
There is a benchmark program under lib/mnesia/examples/bench/ simulating a subscriber database pattern. At one time, it was part of a comparison between proprietary and commercially available cluster database alternatives. In this comparison, mnesia fared surprisingly well, tying with the best commercially available option on raw throughput and scalability, and beating all on response time. This particular benchmar scaled practically linearly up to 50 nodes with mnesia, which was all the computers that could be mustered for the test...
The winner on raw throughput back then wasn't commercially available, but is now, as MySQL Cluster.
But if you want to have a large database cluster, you really need to think about access patterns, and consistency and durability requirements. It's easy enough to get better throughput with dirty access than with transactions, but if you really need transaction consistency, you will have to pay a price. Not all parameters are tunable in the same DBMS - sometimes, you have to switch to an entirely different product to meet your requirements. |
|
|
| Back to top |
|
| Noki4475 |
| Posted: Thu Aug 13, 2009 3:35 pm |
|
|
|
Joined: 09 Aug 2009
Posts: 3
|
Thank you very much Mazen and uwiger for your helpful replies.
Noki. |
|
|
| Back to top |
|
|
|
All times are GMT
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You cannot download files in this forum
|
|
|
Digg It
Del.icio.us
Reddit
Facebook
Stumble Upon
Technorati
