Re: Cassandra Node Density thresholds

[Jonathan Haddad]

50% free is unnecessary.  The only reason to keep that much free is if you
wanted to regularly run major compactions, which you shouldn't.

I'd aim for 75%.  Bootstrap new nodes in when you get close to that
number.  Ensure you don't have any sstables larger than your available
space and you'll be just fine.  Run cleanup on all nodes after you finish
bootstrapping new nodes in to reclaim disk space.

On Fri, May 19, 2017 at 9:42 AM Nitan Kainth  wrote:

> Asad,
>
> Ideally you should keep 50% free disk space. Yes, right now you have good
> option to scale horizontally or add more disk space if you can. Adding more
> nodes will give you more scalability for reads/writes and disk space will
> just give you more capacity  for current load.
>
> Sent from my iPhone
>
> On May 19, 2017, at 11:04 AM, ZAIDI, ASAD A  wrote:
>
> Hello Folks -
>
> I'm using open source apache Cassandra 2.2 .My cluster is spread over 14
> nodes in cluster in two data centers.
>
>
>
> My DC1 data center nodes are reaching 2TB of consumed volume. we don't
> have much space left on disk.
>
> I am wondering if there is guideline available that can point me to
> certain best practice that describe when we should add more nodes to the
> cluster.  should we add more storage or add more nodes. I guess we should
> scale Cassandra horizontally so adding node may be better option.. i am
> looking for a criteria that describes node density thresholds, if there are
> any.
>
> Can you guys please share your thoughts , experience. I'll much appreciate
> your reply. Thanks/Asad
>
>
>
>
>
>

Re: Cassandra Node Density thresholds

[Nitan Kainth]

Asad,

Ideally you should keep 50% free disk space. Yes, right now you have good 
option to scale horizontally or add more disk space if you can. Adding more 
nodes will give you more scalability for reads/writes and disk space will just 
give you more capacity  for current load.

Sent from my iPhone

> On May 19, 2017, at 11:04 AM, ZAIDI, ASAD A  wrote:
> 
> Hello Folks -
> 
> I'm using open source apache Cassandra 2.2 .My cluster is spread over 14 
> nodes in cluster in two data centers.
>  
> My DC1 data center nodes are reaching 2TB of consumed volume. we don't have 
> much space left on disk.
> 
> I am wondering if there is guideline available that can point me to certain 
> best practice that describe when we should add more nodes to the cluster.  
> should we add more storage or add more nodes. I guess we  should scale 
> Cassandra horizontally so adding node may be better option.. i am looking for 
> a criteria that describes node density thresholds, if there are any.
> 
> Can you guys please share your thoughts , experience. I'll much appreciate 
> your reply. Thanks/Asad
> 
>  
> 
>  

Re: Cassandra Node Density thresholds

[daemeon reiydelle]

500 nodes, 20tb of ACTIVE DATA per node in hdfs, no brainer, no problem.
But remember the cross DC traffic will get substantial.

“All men dream, but not equally. Those who dream by night in the dusty
recesses of their minds wake up in the day to find it was vanity, but the
dreamers of the day are dangerous men, for they may act their dreams with
open eyes, to make it possible.” — T.E. Lawrence

sent from my mobile
Daemeon Reiydelle
skype daemeon.c.m.reiydelle
USA 415.501.0198

On May 19, 2017 9:05 AM, "ZAIDI, ASAD A"  wrote:

> Hello Folks -
>
> I'm using open source apache Cassandra 2.2 .My cluster is spread over 14
> nodes in cluster in two data centers.
>
>
>
> My DC1 data center nodes are reaching 2TB of consumed volume. we don't
> have much space left on disk.
>
> I am wondering if there is guideline available that can point me to
> certain best practice that describe when we should add more nodes to the
> cluster.  should we add more storage or add more nodes. I guess we should
> scale Cassandra horizontally so adding node may be better option.. i am
> looking for a criteria that describes node density thresholds, if there are
> any.
>
> Can you guys please share your thoughts , experience. I'll much appreciate
> your reply. Thanks/Asad
>
>
>
>
>