Optimal usage of big server (use VMs or not ...)

Elastic Stack Elasticsearch
1

Regarding the recommendations for the maximal Java heap size, what is the
best practice to utilize one server with lots of RAM (12 or 16 cores, 256
GB RAM):

  1. Create 4 VMs, each 64 GB RAM and 32 GB Java heap size per VM - more
    shards per index

  2. Use bare metal (256 GB RAM total), 128 GB Java heap size - less shards
    per index

Besides the less usable RAM due to pointer compression not taking place,
does the large heap of alternative 2 affect the garbage collection and
stability in such a bad way, that I can consider this as* really not
recommended*?

On the other hand, are there serious performance (disk I/O) or other known
implications using ES in a VM (alternative 1)?

Or is it worth to try it out?

Thanks,

Jan

--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/6737bb7c-9f4d-4b86-aaf3-7a566e5d6809%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

2

A third option has come to my mind:

How about starting several nodes on one server? Any known disadvantages for
this option compared to virtual machines?

Cheers,

Jan

--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/4839d60f-f0ca-445d-85f0-561ac166341d%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

3

We use VMs as we knew the tech, we did consider containers (eg docker) but
they were still new to us and considered a risk.

Running multiple instance on a single node can work, you just need to be
able to manage it. The Elasticsearch puppet module lets you do this though.

Regards,
Mark Walkom

Infrastructure Engineer
Campaign Monitor
email: markw@campaignmonitor.com
web: www.campaignmonitor.com

On 11 October 2014 18:00, knacktus@googlemail.com wrote:

A third option has come to my mind:

How about starting several nodes on one server? Any known disadvantages
for this option compared to virtual machines?

Cheers,

Jan

--
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.
To view this discussion on the web visit
https://groups.google.com/d/msgid/elasticsearch/4839d60f-f0ca-445d-85f0-561ac166341d%40googlegroups.com
https://groups.google.com/d/msgid/elasticsearch/4839d60f-f0ca-445d-85f0-561ac166341d%40googlegroups.com?utm_medium=email&utm_source=footer
.
For more options, visit https://groups.google.com/d/optout.

--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/CAEM624b-gB6X3e%2Bg8cgWPh%3D7%3DPp4prAy61dCbuDYY_w-8-Ucpg%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.