Corundum and Mongo throughput.

Hello

I’ll give this a go and give feedback. I have DO serves now and am playing with Mongo Clusters.

My many guides here: http://fearby.com

Q1) Do you know the rough mongoDB throughput (queries per minute) on MongoDB on a 2 core 2GB ram Digital Ocean server?

Q2) Will reinstalling Coronium overwrite and MongoDB shards or replication servers that I may add?

Locally (iMac) I am doing geo queries and can get 13,000/min with Node and no promises and 32,000.min with promises, but on a clusters of mongoldb Amazon web servers in Sydney called from a node process in singapore from a client in Australia I am getting only 3000 a minute (50/sec).  Not good.

The serve his in singapore because thats the the closest DO location fro me and the AWS is in Sydney the closest server location for mongoldb.com clusters.

I’d be happy to contribute to the codebase.

I’m not totally sure how to answer your question. I have not done any comparisons, so I have no data there.

During beta, Coronium would be a new instance, so you’d have to transfer what you need (for now).

I notice “geo” queries. The Lua Mongo driver does not support those properly, so there is a JS work-around being looked into. There is someone on the team that is looking at the core module though, so hopefully something can be done.

Feel free to join in the Dev Network, it sounds like you’re very knowledgeable about Mongo.

Also, some dev boxes are available, but they are pretty minimal on resources, so they may not provide good testing results.

https://github.com/coroniumcloud

Cheers.

Thanks, I am chasing the highest throughput it I can and recent mongo config changes talking to a HTTPS://cloud.mongodb.com cluster has seen almost double the throughput by querying the secondary servers over the primary (sacrificing consistency). Normally sharding gives scalability. All my custom mongo geo calls are happening in mongoose and NodeJS on Digital ocean and not in corona. I’d kill for lua code that can talk to mongo atlas clusters direct and return a query (saving 400-800ms in server hops from digital ocean in Singapore from Australia then Singapore back to Australia where mongo DB atlas is.

I am happy to test and share (pending time).

I just tried to signup to help with dev and got an error:

Temporarily Unavailable
We are currently experiencing abnormally high traffic volume and working to restore all services as quickly as possible. Thank you for your patience.

My GitHub is: SimonFearby and my email is Simon AT Fearby.com

I’m not totally sure how to answer your question. I have not done any comparisons, so I have no data there.

During beta, Coronium would be a new instance, so you’d have to transfer what you need (for now).

I notice “geo” queries. The Lua Mongo driver does not support those properly, so there is a JS work-around being looked into. There is someone on the team that is looking at the core module though, so hopefully something can be done.

Feel free to join in the Dev Network, it sounds like you’re very knowledgeable about Mongo.

Also, some dev boxes are available, but they are pretty minimal on resources, so they may not provide good testing results.

https://github.com/coroniumcloud

Cheers.

Thanks, I am chasing the highest throughput it I can and recent mongo config changes talking to a HTTPS://cloud.mongodb.com cluster has seen almost double the throughput by querying the secondary servers over the primary (sacrificing consistency). Normally sharding gives scalability. All my custom mongo geo calls are happening in mongoose and NodeJS on Digital ocean and not in corona. I’d kill for lua code that can talk to mongo atlas clusters direct and return a query (saving 400-800ms in server hops from digital ocean in Singapore from Australia then Singapore back to Australia where mongo DB atlas is.

I am happy to test and share (pending time).

I just tried to signup to help with dev and got an error:

Temporarily Unavailable
We are currently experiencing abnormally high traffic volume and working to restore all services as quickly as possible. Thank you for your patience.

My GitHub is: SimonFearby and my email is Simon AT Fearby.com