Re: [mongodb-user] Re: MongoDB on appliance

[Sam Millman]

It could be possible your getting very high fragmentation with high write rates, have you tried commands such as compact: http://docs.mongodb.org/manual/reference/command/compact/

On 9 November 2012 09:00, Jose Silva <pela...@gmail.com> wrote:

Hi, I would like to have more comments on that topic. I'm using MongoDB in our solution, and I'm pretty much happy with the solution (We have a high write rate, that would be impossible to cope with any RDBM). However we have problem with the huge files that are allocated in advance.  local.x, journaling, etc, etc.. they are huge and they keep growing.

As partial solution we configured smallfiles=true and optSize (for replica) = 10 Mb and we are monitoring them. Any other tip/idea about "how to keep those extra files managable for devices where disk space does matter" ?

thanks

On Wednesday, October 3, 2012 5:16:25 PM UTC+2, Vlad wrote:

Hi mongodb-user,

I am currently evaluating feasibility of using MongoDB in our application.
Here's the basic outline of our use case.

1. My main feature of interest at this time is flexible schema/efficient search(indexing). Our deployment model would be single-node, no replication, no sharding.
Our transaction volume is expected to be moderate and expected collection size average about 1 million items/per collection in likely less than 10 collections.
High availability currently is less critical for us, but data durability is important.

2. Our product offering is security appliance that runs in a headless, unattended fashion.
Assume there will be no human intervention to perform monitoring or execute manual recovery steps upon system crash.
Also, assume there's no inbound/outbound web traffic from 3-rd party vendors so we can't use MMS service for monitoring.
The way I see it we would need to develop kind of stripped-down, very basic version of MMS to perform basic monitoring/recovery for a single-node instance.

I am just trying to assess feasibility of such solution.
I understand the new journaling feature will probably go some way in addressing this as far as data loss is concerned(self-recovery on crash), but there may be other pitfalls that I am unaware of.

Does anyone use Mongo in a similar setting?
Would you recommend on using it?


Thanks

--
You received this message because you are subscribed to the Google
Groups "mongodb-user" group.
To post to this group, send email to mongod...@googlegroups.com
To unsubscribe from this group, send email to
mongodb-user...@googlegroups.com
See also the IRC channel -- freenode.net#mongodb

[Sam Millman]

However I just realised that isn't a good answer to what you WERE actually asking.