1

is there some optimal setup when configuring gigaspaces on 8 servers?
I was thinking of using 4 gsm-s (and lus-s) in total and 2 gsc-s per server.
I was thinking of using the following sla for the space:

<os-sla:sla cluster-schema="partitioned-sync2backup" number-of-instances="8" number-of-backups="1" max-instances-per-vm="1"/>

I want to make use of the resources for a map-reduce application which will be ran on a dedicated additional server.

Regards,
Despot

despot
  • 7,167
  • 9
  • 44
  • 63

1 Answers1

3

You do not need more than 2 GSM's (and LUS's) per cluster of this size.

Have atleast 8 instances and 1 backup in your space sla so that all available GSC's are used (16 == 2 GSC's per server * 8 servers).

Btw, optimum number of GSC's on each server is dependent on available memory and CPU cores. Depending on use case you may be able to allocate more or less GSC's on each server.

SKP
  • 135
  • 4
  • I meant "number-of-instances="8"" - I edited it. "Btw, optimum number of GSC's on each server is dependent on available memory and CPU cores. " So do you know what the rule of thumb is (best practices) regarding the relationship between number of GSC's with number of CPU cores and available memory? Is there some documentation on this I can read? – despot Oct 06 '13 at 08:24
  • 2
    It mostly depends on use case. I have seen apps with 2-8 GB GSC's per CPU core depending on whether the app has in memory processing and data (smaller heap) or caching only (larger heap). – SKP Oct 07 '13 at 03:11