Several ElasticSearchs on the same servers

Elastic Stack Elasticsearch
1

Hi,

I am building a shared elasticsearch cluster (4 nodes) that will be used by
several projects (at least 4).

Because each project need to control configuration, node restart, and
version upgrades, a single cluster with multi-tenant is not an option.

Some projects will have few documents indexed (10K docs of 1 KB per doc),
but will be querried massively.
Other project will receive logs continuously. Logs will be deleted after
some month. But documents will not be querried often.

I am looking for any feedbacks how to configure my different ES processes :

  • Is LVM a goog idea to ensure disk quota ?
  • is Xmx enough to isolate process memory or should I use memory cgrouphttps://www.kernel.org/doc/Documentation/cgroups/memory.txt
    ?
  • What is the miminum memory advised for each process ?
  • should mlockall be enabled on each ES ?
  • should I set some limitation on the network ?

Thanks for your help.

Romain

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

2

Any idea ?

Le jeudi 7 mars 2013 07:49:44 UTC+1, Filirom1 a écrit :

Hi,

I am building a shared elasticsearch cluster (4 nodes) that will be used
by several projects (at least 4).

Because each project need to control configuration, node restart, and
version upgrades, a single cluster with multi-tenant is not an option.

Some projects will have few documents indexed (10K docs of 1 KB per doc),
but will be querried massively.
Other project will receive logs continuously. Logs will be deleted after
some month. But documents will not be querried often.

I am looking for any feedbacks how to configure my different ES processes
:

  • Is LVM a goog idea to ensure disk quota ?
  • is Xmx enough to isolate process memory or should I use memory cgrouphttps://www.kernel.org/doc/Documentation/cgroups/memory.txt
    ?
  • What is the miminum memory advised for each process ?
  • should mlockall be enabled on each ES ?
  • should I set some limitation on the network ?

Thanks for your help.

Romain

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.